What to do with the pizza ontology after creating it?

classic Classic list List threaded Threaded
23 messages Options
12
Reply | Threaded
Open this post in threaded view
|

What to do with the pizza ontology after creating it?

Sylvia Breau

I’m struggling to understand how to use the pizza ontology after creating it. Let’s say I had a group of people (non-ontologists) who want to search for pizza. Not sure how to ask the question…does the ontology get published somewhere, and then there is a search interface on a web site?

 

I’m looking into Protégé because my company is exploring the possibility of using ontology as part of our eCommerce solution. I’m having trouble making the leap from the ontology development at the back end and how to deliver it to end users.

 

Any help with filling in this gap would be greatly appreciated!

 

DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Samuel Croset
Hi Sylvia,

In your case I guess you should consider the ontology as something similar to a database but from which you could hopefully extract more information out of. The ontology would ideally present more granular product choices to your consumers (in the context of a business).

Example: Let's say you run a pizza shop and that the pizza ontology describes what you have to sell.
You could either use the ontology in the back-end as a model just like you do it with a database in a web application. The developer would write the queries to get the results to be displayed.

You could also let the users interact with the ontology via an interface. Assuming that users are going to be capable of creating OWL queries is certainly a big leap of faith, but you could simplify the interaction.
For instance you want to give users the possibility to filter pizzas based on the country of origin, which corresponds to the following OWL expression: hasCountryOfOrigin value ?
You could just let the user pick the ? and then you run the query over the ontology. Here the user has to possibility to filter a product based on particular features. You could think of more complicated expressions in order to come up with expressive filters.

Just like for relational databases, users don't really create SQL requests themselves, but mostly via an interface. I guess same goes for ontologies. One alternative to let users directly formulate query would be to use rdfs:label values to construct OWL class expressions. The query before could then easily be transformed in something like: 'has country of origin' value 'America'.

Hope it answers somehow your question. I'm not mentioning other use-cases such as standardised vocabulary: You could think of extending your ontology pizzas re-using schema.org concept (such as Restaurant) and aim to display rich snippets about your shop in search engines, etc. 

Ontologies should help you to enhance the interaction between your consumers (finding the product they want), search engines (they know and understand what you sell) and the back-end (formulation of expressive queries).

Cheers,

Samuel 


On Thu, Feb 14, 2013 at 9:44 PM, Sylvia Breau <[hidden email]> wrote:

I’m struggling to understand how to use the pizza ontology after creating it. Let’s say I had a group of people (non-ontologists) who want to search for pizza. Not sure how to ask the question…does the ontology get published somewhere, and then there is a search interface on a web site?

 

I’m looking into Protégé because my company is exploring the possibility of using ontology as part of our eCommerce solution. I’m having trouble making the leap from the ontology development at the back end and how to deliver it to end users.

 

Any help with filling in this gap would be greatly appreciated!

 

DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

James Naish
In reply to this post by Sylvia Breau
Hi Sylvia,

I'm not an expert ontologist but am a software engineer by trade who,
after a good few years hearing about the wonders that ontologies have
to offer, finally decided to take the plunge.

My understanding of ontology is as a form of data modelling that is
significantly richer than relational data modelling. OWL can express a
much wider range of relationships than a relational model can, and so
this allows us to ask much richer queries.

It's really up to you how you exploit this capability - I personally
view it as a technology which can complement traditional database
technologies by allowing me to make connections between entities that
I couldn't make using standard SQL queries.

That being the case, I view it as a back-end technology around which I
can wrap user interfaces that support different kinds of query over my
data model. A search engine would be one way of doing this, but the
application I've built recently -a tool to support software reuse -
doesn't allow users to enter queries manually and instead provides the
user with interfaces which hide the complexity of the underlying data
model.

There are several ways you might exploit the powerful reasoning
capabilities which OWL offers in an eCommerce solution: recommending
products to users; more accurate product searches based on semantic,
rather than syntactic, queries;  these are two obvious cases, but I'm
sure your imagination can take you further.

I guess my point really is that there is no "right" or "wrong" way to
use ontology technology. It's a powerful data modelling tool which you
can use to describe the data over which your e-commerce solution would
operate. You can use it as a tool to support search, but it can also
be used as an exclusively back-end technology which provides more
powerful ways for choosing what data to present to users and when.

Hope this helps,

James.

On 2/14/13, Sylvia Breau <[hidden email]> wrote:

> I'm struggling to understand how to use the pizza ontology after creating
> it. Let's say I had a group of people (non-ontologists) who want to search
> for pizza. Not sure how to ask the question...does the ontology get
> published somewhere, and then there is a search interface on a web site?
>
>
>
> I'm looking into Protégé because my company is exploring the possibility of
> using ontology as part of our eCommerce solution. I'm having trouble making
> the leap from the ontology development at the back end and how to deliver it
> to end users.
>
>
>
> Any help with filling in this gap would be greatly appreciated!
>
>
>
>
>
> DISCLAIMER:
> This e-mail is intended for the use of the addressee(s) only and may contain
> privileged, confidential, or proprietary information that is
> exempt from disclosure under law. If you are not the intended recipient,
> please do not read, copy, use or disclose the contents of this
> communication to others. Please notify the sender that you have received
> this e-mail in error by replying to the e-mail. Please then
> delete the e-mail and destroy any copies of it. Thank you.
>
>
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Tom Cloyd
I want to say that this whole brief thread is of real value to me, as I
stand on the brink of diving into Protege. It is not entirely clear to
me what can be done with this. Coming from a decent understanding of
relational databases, and having played just a little with the pizza
ontology, I got the same impression which you offer here, James. Your
background is much richer than mine, relative to data modeling, so I
really appreciate your thoughts.

Tom
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd / [hidden email] / (435) 272-3332

On 02/15/2013 03:35 AM, James Naish wrote:

> Hi Sylvia,
>
> I'm not an expert ontologist but am a software engineer by trade who,
> after a good few years hearing about the wonders that ontologies have
> to offer, finally decided to take the plunge.
>
> My understanding of ontology is as a form of data modelling that is
> significantly richer than relational data modelling. OWL can express a
> much wider range of relationships than a relational model can, and so
> this allows us to ask much richer queries.
>
> It's really up to you how you exploit this capability - I personally
> view it as a technology which can complement traditional database
> technologies by allowing me to make connections between entities that
> I couldn't make using standard SQL queries.
>
> That being the case, I view it as a back-end technology around which I
> can wrap user interfaces that support different kinds of query over my
> data model. A search engine would be one way of doing this, but the
> application I've built recently -a tool to support software reuse -
> doesn't allow users to enter queries manually and instead provides the
> user with interfaces which hide the complexity of the underlying data
> model.
>
> There are several ways you might exploit the powerful reasoning
> capabilities which OWL offers in an eCommerce solution: recommending
> products to users; more accurate product searches based on semantic,
> rather than syntactic, queries;  these are two obvious cases, but I'm
> sure your imagination can take you further.
>
> I guess my point really is that there is no "right" or "wrong" way to
> use ontology technology. It's a powerful data modelling tool which you
> can use to describe the data over which your e-commerce solution would
> operate. You can use it as a tool to support search, but it can also
> be used as an exclusively back-end technology which provides more
> powerful ways for choosing what data to present to users and when.
>
> Hope this helps,
>
> James.
>
> On 2/14/13, Sylvia Breau <[hidden email]> wrote:
>> I'm struggling to understand how to use the pizza ontology after creating
>> it. Let's say I had a group of people (non-ontologists) who want to search
>> for pizza. Not sure how to ask the question...does the ontology get
>> published somewhere, and then there is a search interface on a web site?
>>
>>
>>
>> I'm looking into Protégé because my company is exploring the possibility of
>> using ontology as part of our eCommerce solution. I'm having trouble making
>> the leap from the ontology development at the back end and how to deliver it
>> to end users.
>>
>>
>>
>> Any help with filling in this gap would be greatly appreciated!
>>
>>
>>
>>
>>
>> DISCLAIMER:
>> This e-mail is intended for the use of the addressee(s) only and may contain
>> privileged, confidential, or proprietary information that is
>> exempt from disclosure under law. If you are not the intended recipient,
>> please do not read, copy, use or disclose the contents of this
>> communication to others. Please notify the sender that you have received
>> this e-mail in error by replying to the e-mail. Please then
>> delete the e-mail and destroy any copies of it. Thank you.
>>
>>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>


--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: (435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

James Naish
Tom, I'm not an expert in OWL as I say, but I have applied it now to a single significant project and learnt a lot from it. My understanding is that one of the criticisms that has been launched against ontology-based development and, more broadly, the semantic web vision is that there has yet to be a really good "killer app" launched which demonstrates the real value of ontology-based development over more traditional relational or syntactic data modelling. Others with a deeper research background in this area than mine can probably tell you a lot more! Maybe a really useful thing for the community to do as a whole would be to set up some kind of ontology app showcase site where people can post about the applications they're building using OWL ontologies so that newcomers can get a better understanding of what can be down with the technology as this seems to be a major hurdle to adopting the technology!
 
I'm based at the University of Manchester so I hear a lot about it, but I've also never been sure of the value of OWL over other data modelling techniques. I wanted to find out more for myself so I dived in at the deep end and used it to build an implementation of my PhD work - a reuse-based engine for model transformation. At the other end of this project, while the pizza ontology example is useful for getting to grips with the basics of protege, it doesn't do OWL or ontology-based development justice as it isn't sufficient to illustrate what can really be down with the technology. There are some neat features though which I've found useful in my own project.
 
The main purpose, I guess, of OWL is to provide a language for data modelling which allows for the discovery of facts over a data model by making implicit relationships and properties explicit. One example from my own application as to how this capability can be really quite useful is in modelling sub-typing relationships. In my approach, software models are expressed as a set of OWL individuals (Facts and Properties) which instantiate a set of OWL classes (FactTypes and PropertyTypes). I wanted to model the subtype relationship, so I created an OWL:ObjectProperty called "subTypeOf" whose domain and range is FactType. Now the question was how to process the relationship to take inheritance and upcasting into account? There are two basic solutions to this. One is to handle it programmatically by writing a routine to check whether a Fact is of a type which inherits from another type wherever this is relevant. In OWL, however, I have an alternative option:
  • To create a rule to handle upcasting: Type(?f, Fact), PropertyValue(?f, instantiatesFactType, ?ftSub), PropertyValue(?ftSub, subTypeOf, ?ftSuper) -> PropertyValue(?f, instantiatesFactType, ?ftSuper)
  • To create a rule to handle inheritance: Type(?ftSub, FactType), Type(?ftSuper, FactType), PropertyValue(?ftSub, subTypeOf, ?ftSuper), PropertyValue(?ftSuper, hasPropertyOfType, ?p) -> PropertyValue(?ftSub, hasPropertyOfType, ?p)
This is much more straightforward because whenever I process my ontologies using OWL API and a suitable reasoner (Pellet or Hermit), the reasoner will apply these rules to automatically realise the inheritance and upcasting effects I was after. It's one example but I'm sure there are plenty of other cases that others could give.
 
My key conclusions of my own experiences are that there are pros, cons and some definite quirks versus relational modelling when it comes to using OWL.
 
Pros:
  • Very rich and flexible data modelling approach which allows for relationships and concepts to be expressed which cannot be expressed using more traditional data modelling techniques;
  • Powerful reasoning technology which supports the inference of relationships that would need to be programmatically inferred or explicitly stated under less powerful technologies;
  • Easy portability of OWL models which supports nifty application development in a word where online, web-based data processign and sharing is ever more important.
Quirks (not a complete list but some possibly useful pointers):
  • Relationships do not have to be expressed explicitly to be inferred;
  • OWL is not type-safe in the traditional sense. Asserting axioms about a particular object which do not conform to the explicit rules about that object will not cause any error. Instead, it will cause OWL reasoners to infer that the object is of a type for which those axioms are type safe. For instance, if, in my example above, I asserted that a Fact, f1, was a subTypeOf another Fact, f2, despite the fact that the Domain and Range of the subTypeOf relationship are both "FactType", not "Fact", then OWL would not complain but instead would infer that f1 and f2 were FactTypes as well as Facts - a logically correct, but empirically incorrect, inference;
  • Cardinality relationships are not so easily expressed in OWL, which is about open world reasoning.
Cons:
  • The quirks above can appear to be drawbacks in certain cases, but you learn to love them (or at least appreciate the logic behind them) as you get used to them;
  • The major con at this stage is the performance of OWL reasoners. Complex reasoning can be very slow over large ontologies (upwards of 45 minutes in some cases which I came across), so you need to understand how these reasoners work and be prepared to optimise rules and queries accordingly. That said, I've found Pellet seems to perform a lot better when used from within Protege than within a standalone Java app, so that may not always be an issue.
Anyway, just a few more thoughts there so I hope that's helpful.
 
James.
 
 
On Sun, Feb 17, 2013 at 10:09 PM, Tom Cloyd <[hidden email]> wrote:
I want to say that this whole brief thread is of real value to me, as I stand on the brink of diving into Protege. It is not entirely clear to me what can be done with this. Coming from a decent understanding of relational databases, and having played just a little with the pizza ontology, I got the same impression which you offer here, James. Your background is much richer than mine, relative to data modeling, so I really appreciate your thoughts.

Tom
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332" target="_blank" value="+14352723332">(435) 272-3332


On 02/15/2013 03:35 AM, James Naish wrote:
Hi Sylvia,

I'm not an expert ontologist but am a software engineer by trade who,
after a good few years hearing about the wonders that ontologies have
to offer, finally decided to take the plunge.

My understanding of ontology is as a form of data modelling that is
significantly richer than relational data modelling. OWL can express a
much wider range of relationships than a relational model can, and so
this allows us to ask much richer queries.

It's really up to you how you exploit this capability - I personally
view it as a technology which can complement traditional database
technologies by allowing me to make connections between entities that
I couldn't make using standard SQL queries.

That being the case, I view it as a back-end technology around which I
can wrap user interfaces that support different kinds of query over my
data model. A search engine would be one way of doing this, but the
application I've built recently -a tool to support software reuse -
doesn't allow users to enter queries manually and instead provides the
user with interfaces which hide the complexity of the underlying data
model.

There are several ways you might exploit the powerful reasoning
capabilities which OWL offers in an eCommerce solution: recommending
products to users; more accurate product searches based on semantic,
rather than syntactic, queries;  these are two obvious cases, but I'm
sure your imagination can take you further.

I guess my point really is that there is no "right" or "wrong" way to
use ontology technology. It's a powerful data modelling tool which you
can use to describe the data over which your e-commerce solution would
operate. You can use it as a tool to support search, but it can also
be used as an exclusively back-end technology which provides more
powerful ways for choosing what data to present to users and when.

Hope this helps,

James.

On 2/14/13, Sylvia Breau <[hidden email]> wrote:
I'm struggling to understand how to use the pizza ontology after creating
it. Let's say I had a group of people (non-ontologists) who want to search
for pizza. Not sure how to ask the question...does the ontology get
published somewhere, and then there is a search interface on a web site?



I'm looking into Protégé because my company is exploring the possibility of
using ontology as part of our eCommerce solution. I'm having trouble making
the leap from the ontology development at the back end and how to deliver it
to end users.



Any help with filling in this gap would be greatly appreciated!





DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain
privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient,
please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received
this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: <a href="tel:%28435%29%20272-3332" target="_blank" value="+14352723332">(435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Tom Cloyd
James,

Please see my comments, below...

On 02/17/2013 05:40 PM, James Naish wrote:
Tom, I'm not an expert in OWL as I say, but I have applied it now to a single significant project and learnt a lot from it. My understanding is that one of the criticisms that has been launched against ontology-based development and, more broadly, the semantic web vision is that there has yet to be a really good "killer app" launched which demonstrates the real value of ontology-based development over more traditional relational or syntactic data modelling. Others with a deeper research background in this area than mine can probably tell you a lot more! Maybe a really useful thing for the community to do as a whole would be to set up some kind of ontology app showcase site where people can post about the applications they're building using OWL ontologies so that newcomers can get a better understanding of what can be down with the technology as this seems to be a major hurdle to adopting the technology!
I absolutely agree. As I have intermittently observed this community, through this list, it's clear that there are folks here who are professionally involved, and very much in the deep end of the pool, AND there are those of us who are exploring or thinking of exploring this tool, because we have problems to solve for which we haven't yet found a solution. We're hope that Protege, and more generally an ontology built using it, might be our answer.

I have looked at some pretty complex ontologies, and simply been unable to see what is actually being done with them. That's not helpful to me. Too many trees; not enough forest, to my eyes (and I imagine that this is my problem, not the ontology's).

Basically I see two sorts of uses:

1. Data modeling - to use your terms. This, in my vocabulary, is "knowledge capture". It involves data records embodying basically 3 fields - entity 1 > relationship > entity 2. If we elaborate "entity" and "relationship" (which is really a type of entity) so that they can have qualifiers - properties, then all sorts of interesting possibilities emerge. I know all this has been thought out by others, and better than I could do it, but my point is that this sort of construction is an accessible starting point for many of us, I expect. The simple reason for this is that I have just abstracted a simple assertion, which I take to be the form all knowledge (if we set aside the realization that merely being able to label some part of the world reliably - to determine that an entity exists - is itself a "knowing"). So if I wish to be able to capture the core "knowledge" in, say, a corpus or research papers, what I will be doing is framing the essential assertions made by this literature. That, indeed, is my primary interest in using Protege, at this point.

2. Reasoning - This seems to be the main interest of most of the members of this discussion list. When one sets up assertions and classes to which they belong, there are implications. Reason software can tease these out. There are also classification error (assertion conflicts), and software can point these out too. This is an enticing prospect. Most people in my field (in fact everyone who I have knowledge of) has had NO contact with this sort of operation whatsoever. I see people doing biomedical ontologies; I see absolutely nothing in psychology, much less psychotherapy research.
 
I'm based at the University of Manchester so I hear a lot about it, but I've also never been sure of the value of OWL over other data modelling techniques. I wanted to find out more for myself so I dived in at the deep end and used it to build an implementation of my PhD work - a reuse-based engine for model transformation. At the other end of this project, while the pizza ontology example is useful for getting to grips with the basics of protege, it doesn't do OWL or ontology-based development justice as it isn't sufficient to illustrate what can really be down with the technology. There are some neat features though which I've found useful in my own project.
I do not, sadly, grasp the purport of your PhD work, but that you attempted to implement some aspect of it with OWL seems reasonable and intriguing. That's basically the kind of thing I want to do also. I agree that the pizza ontology seem to stop short of where I should thing this technology could go, and where I hope it does go. However, it does have one virtue: as far as it goes, it is not difficult to grasp. It is an illustration. We just need more elaborated ones, and ones that are perhaps less tutorials and more use-case descriptions.

That would be relevant IF anyone seeks to engage more people in trying out this technology. I don't know if that interest exists or not.
 
The main purpose, I guess, of OWL is to provide a language for data modelling which allows for the discovery of facts over a data model by making implicit relationships and properties explicit.
Well, that's the SECOND of the two uses I've identified above, and the one of least interest to me. I see great value merely in the first. I am, of course, interested to see where the second might take me, but that would require my accomplishing the first to begin with, and I have not yet attempted to do that.
One example from my own application as to how this capability can be really quite useful is in modelling sub-typing relationships. In my approach, software models are expressed as a set of OWL individuals (Facts and Properties) which instantiate a set of OWL classes (FactTypes and PropertyTypes). I wanted to model the subtype relationship, so I created an OWL:ObjectProperty called "subTypeOf" whose domain and range is FactType. Now the question was how to process the relationship to take inheritance and upcasting into account? There are two basic solutions to this. One is to handle it programmatically by writing a routine to check whether a Fact is of a type which inherits from another type wherever this is relevant. In OWL, however, I have an alternative option:
  • To create a rule to handle upcasting: Type(?f, Fact), PropertyValue(?f, instantiatesFactType, ?ftSub), PropertyValue(?ftSub, subTypeOf, ?ftSuper) -> PropertyValue(?f, instantiatesFactType, ?ftSuper)
  • To create a rule to handle inheritance: Type(?ftSub, FactType), Type(?ftSuper, FactType), PropertyValue(?ftSub, subTypeOf, ?ftSuper), PropertyValue(?ftSuper, hasPropertyOfType, ?p) -> PropertyValue(?ftSub, hasPropertyOfType, ?p)
This is much more straightforward because whenever I process my ontologies using OWL API and a suitable reasoner (Pellet or Hermit), the reasoner will apply these rules to automatically realise the inheritance and upcasting effects I was after. It's one example but I'm sure there are plenty of other cases that others could give.
If I understand you, what you've discovered is that OWL allows a more efficient exploration of aspects of this modeling problem than would otherwise be available to you. That sounds like a win for you, and certainly for those who designed Protege/OWL to be more than a theoretical exercise.
 
My key conclusions of my own experiences are that there are pros, cons and some definite quirks versus relational modelling when it comes to using OWL.
Immediate reaction - just as with is likely to be true  with any other technology you might have used.
 
Pros:
  • Very rich and flexible data modelling approach which allows for relationships and concepts to be expressed which cannot be expressed using more traditional data modelling techniques;
  • Powerful reasoning technology which supports the inference of relationships that would need to be programmatically inferred or explicitly stated under less powerful technologies;
  • Easy portability of OWL models which supports nifty application development in a word where online, web-based data processign and sharing is ever more important.
Quirks (not a complete list but some possibly useful pointers):
  • Relationships do not have to be expressed explicitly to be inferred;
  • OWL is not type-safe in the traditional sense. Asserting axioms about a particular object which do not conform to the explicit rules about that object will not cause any error. Instead, it will cause OWL reasoners to infer that the object is of a type for which those axioms are type safe. For instance, if, in my example above, I asserted that a Fact, f1, was a subTypeOf another Fact, f2, despite the fact that the Domain and Range of the subTypeOf relationship are both "FactType", not "Fact", then OWL would not complain but instead would infer that f1 and f2 were FactTypes as well as Facts - a logically correct, but empirically incorrect, inference;
  • Cardinality relationships are not so easily expressed in OWL, which is about open world reasoning.
Cons:
  • The quirks above can appear to be drawbacks in certain cases, but you learn to love them (or at least appreciate the logic behind them) as you get used to them;
  • The major con at this stage is the performance of OWL reasoners. Complex reasoning can be very slow over large ontologies (upwards of 45 minutes in some cases which I came across), so you need to understand how these reasoners work and be prepared to optimise rules and queries accordingly. That said, I've found Pellet seems to perform a lot better when used from within Protege than within a standalone Java app, so that may not always be an issue.
Anyway, just a few more thoughts there so I hope that's helpful.
 
James.
That's all a lot more than I hoped for. A fair amount of it is likely to be especially useful to me after I've invested more time in formalizing my assertions. I find your report of your experience, especially as a newcomer to this all, particularly interesting, and basically rather reassuring. Thanks for taking the time to share this, and I suspect I'm not the only one who will benefit from it.

I feel encouraged to go ahead with my own personal experiment, and do expect it will be worth the time I give it.

Thanks, and I hope you'll continue to report your observations, if you continue working with your ontology, and with this technology.

Which leads to a final obvious question: Do you plan to continue working with this technology? What do you see it doing for you?

Tom
 
 
On Sun, Feb 17, 2013 at 10:09 PM, Tom Cloyd <[hidden email]> wrote:
I want to say that this whole brief thread is of real value to me, as I stand on the brink of diving into Protege. It is not entirely clear to me what can be done with this. Coming from a decent understanding of relational databases, and having played just a little with the pizza ontology, I got the same impression which you offer here, James. Your background is much richer than mine, relative to data modeling, so I really appreciate your thoughts.

Tom
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd / [hidden email] / <a moz-do-not-send="true" href="tel:%28435%29%20272-3332" target="_blank" value="+14352723332">(435) 272-3332


On 02/15/2013 03:35 AM, James Naish wrote:
Hi Sylvia,

I'm not an expert ontologist but am a software engineer by trade who,
after a good few years hearing about the wonders that ontologies have
to offer, finally decided to take the plunge.

My understanding of ontology is as a form of data modelling that is
significantly richer than relational data modelling. OWL can express a
much wider range of relationships than a relational model can, and so
this allows us to ask much richer queries.

It's really up to you how you exploit this capability - I personally
view it as a technology which can complement traditional database
technologies by allowing me to make connections between entities that
I couldn't make using standard SQL queries.

That being the case, I view it as a back-end technology around which I
can wrap user interfaces that support different kinds of query over my
data model. A search engine would be one way of doing this, but the
application I've built recently -a tool to support software reuse -
doesn't allow users to enter queries manually and instead provides the
user with interfaces which hide the complexity of the underlying data
model.

There are several ways you might exploit the powerful reasoning
capabilities which OWL offers in an eCommerce solution: recommending
products to users; more accurate product searches based on semantic,
rather than syntactic, queries;  these are two obvious cases, but I'm
sure your imagination can take you further.

I guess my point really is that there is no "right" or "wrong" way to
use ontology technology. It's a powerful data modelling tool which you
can use to describe the data over which your e-commerce solution would
operate. You can use it as a tool to support search, but it can also
be used as an exclusively back-end technology which provides more
powerful ways for choosing what data to present to users and when.

Hope this helps,

James.

On 2/14/13, Sylvia Breau <[hidden email]> wrote:
I'm struggling to understand how to use the pizza ontology after creating
it. Let's say I had a group of people (non-ontologists) who want to search
for pizza. Not sure how to ask the question...does the ontology get
published somewhere, and then there is a search interface on a web site?



I'm looking into Protégé because my company is exploring the possibility of
using ontology as part of our eCommerce solution. I'm having trouble making
the leap from the ontology development at the back end and how to deliver it
to end users.



Any help with filling in this gap would be greatly appreciated!





DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain
privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient,
please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received
this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



--

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: <a moz-do-not-send="true" href="tel:%28435%29%20272-3332" target="_blank" value="+14352723332">(435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



-- 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: (435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Phillip Lord
In reply to this post by Tom Cloyd


The pizza ontology is an exemplar, of course. It's not really that
useful, in and off it's own right.

I think that the two main uses for ontologies, at from my experience,
are providing a navigational structure. So, as you say, you could find a
pizza by a number of different routes; and make it easier to maintain
those routes. So, vegetarian pizza and cheese pizza get populated for
you. Not so big an advantage for pizzas, but if you have a large parts
catalogue, then much more use.

The second main use is to annotate a large number of individuals for
retrieval or statistical analysis. Think Dewey Decimal and a library.
You can ask, which fields of knowledge have increased in popularity in
the last 50 years.

And the third use (cue spanish inquisition etc...) is to provide a
transferable data model. So things like FOAF could be represented in
relational data model, but how publish and transfer this knowledge?
Strictly this is more a property of OWL rather than ontologies per se,
but it's worth mentioning.

Phil



Tom Cloyd <[hidden email]> writes:

> I want to say that this whole brief thread is of real value to me, as I stand
> on the brink of diving into Protege. It is not entirely clear to me what can
> be done with this. Coming from a decent understanding of relational databases,
> and having played just a little with the pizza ontology, I got the same
> impression which you offer here, James. Your background is much richer than
> mine, relative to data modeling, so I really appreciate your thoughts.
>
> Tom
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Tom Cloyd / [hidden email] / (435) 272-3332
>
> On 02/15/2013 03:35 AM, James Naish wrote:
>> Hi Sylvia,
>>
>> I'm not an expert ontologist but am a software engineer by trade who,
>> after a good few years hearing about the wonders that ontologies have
>> to offer, finally decided to take the plunge.
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

James Naish
No one expected the Spanish Reference!

On Mon, Feb 18, 2013 at 12:53 PM, Phillip Lord <[hidden email]> wrote:


The pizza ontology is an exemplar, of course. It's not really that
useful, in and off it's own right.

I think that the two main uses for ontologies, at from my experience,
are providing a navigational structure. So, as you say, you could find a
pizza by a number of different routes; and make it easier to maintain
those routes. So, vegetarian pizza and cheese pizza get populated for
you. Not so big an advantage for pizzas, but if you have a large parts
catalogue, then much more use.

The second main use is to annotate a large number of individuals for
retrieval or statistical analysis. Think Dewey Decimal and a library.
You can ask, which fields of knowledge have increased in popularity in
the last 50 years.

And the third use (cue spanish inquisition etc...) is to provide a
transferable data model. So things like FOAF could be represented in
relational data model, but how publish and transfer this knowledge?
Strictly this is more a property of OWL rather than ontologies per se,
but it's worth mentioning.

Phil



Tom Cloyd <[hidden email]> writes:

> I want to say that this whole brief thread is of real value to me, as I stand
> on the brink of diving into Protege. It is not entirely clear to me what can
> be done with this. Coming from a decent understanding of relational databases,
> and having played just a little with the pizza ontology, I got the same
> impression which you offer here, James. Your background is much richer than
> mine, relative to data modeling, so I really appreciate your thoughts.
>
> Tom
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332" value="+14352723332">(435) 272-3332
>
> On 02/15/2013 03:35 AM, James Naish wrote:
>> Hi Sylvia,
>>
>> I'm not an expert ontologist but am a software engineer by trade who,
>> after a good few years hearing about the wonders that ontologies have
>> to offer, finally decided to take the plunge.
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Julian Vincent
In reply to this post by Phillip Lord
Many thanks for this brief discussion.  I've been learning about ontologies for about 4 years now, mostly in glorious isolation.  My mission is to build a design interface between biology and engineering, which I am doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/  Please visit, dissect and correct!).  A database was abysmally inadequate for the purpose whereas the ontology can introduce all the hierarchical relationships of biology very easily and cope with my changing visions of what I want, which will eventually be an AI system for robotics.  Unfortunately I don't think I've got much beyond the pizza level of understanding, despite grappling with books on the semantic web, bio-ontologies, etc.  I still can't understand the reasoning behind some of the more complex constructions used by Matthew Horridge in his "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put together in order to demonstrate some ideas.  But, Matthew, WHAT were
 the ideas???).  Is the Protege community ready for a higher level tutorial text on ontologies, their structures, and the reasons for preferring one or the other?  I'd love to have one.  I don't want to see yet another paper saying that we connected these boxes like this . . . or this . . .   What was the coding you used and why?  The pizza tutorial is excellent, but it's just the first course.  More!  More!!

Julian Vincent


On 18 Feb 2013, at 12:53, Phillip Lord wrote:

>
>
> The pizza ontology is an exemplar, of course. It's not really that
> useful, in and off it's own right.
>
> I think that the two main uses for ontologies, at from my experience,
> are providing a navigational structure. So, as you say, you could find a
> pizza by a number of different routes; and make it easier to maintain
> those routes. So, vegetarian pizza and cheese pizza get populated for
> you. Not so big an advantage for pizzas, but if you have a large parts
> catalogue, then much more use.
>
> The second main use is to annotate a large number of individuals for
> retrieval or statistical analysis. Think Dewey Decimal and a library.
> You can ask, which fields of knowledge have increased in popularity in
> the last 50 years.
>
> And the third use (cue spanish inquisition etc...) is to provide a
> transferable data model. So things like FOAF could be represented in
> relational data model, but how publish and transfer this knowledge?
> Strictly this is more a property of OWL rather than ontologies per se,
> but it's worth mentioning.
>
> Phil
>
>
>
> Tom Cloyd <[hidden email]> writes:
>
>> I want to say that this whole brief thread is of real value to me, as I stand
>> on the brink of diving into Protege. It is not entirely clear to me what can
>> be done with this. Coming from a decent understanding of relational databases,
>> and having played just a little with the pizza ontology, I got the same
>> impression which you offer here, James. Your background is much richer than
>> mine, relative to data modeling, so I really appreciate your thoughts.
>>
>> Tom
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Tom Cloyd / [hidden email] / (435) 272-3332
>>
>> On 02/15/2013 03:35 AM, James Naish wrote:
>>> Hi Sylvia,
>>>
>>> I'm not an expert ontologist but am a software engineer by trade who,
>>> after a good few years hearing about the wonders that ontologies have
>>> to offer, finally decided to take the plunge.
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

James Naish
This discusssion has got me thinking. I'm sure there's plenty of great conferences out there on OWL, ontologies, DLs, etc for those who are in the know, but it seems to me there are plenty of us out there whose interests are predominantly practical rather than research-oriented. That is, even if we're involved in research, our contributions apply OWL but may not contribute to the OWL/DL/Ontology literature itself. Maybe there are enough of us about to get some form of hacking symposium together on a semi-regular (annual, bi-annual?) basis. I'm thinking something along the lines of the hacking clubs of the 70s in which people would turn up with practical problems, share solutions, and discuss visions for future pragmatic applications of new technologies, but with a focus on ontology-/OWL-based development.
 
What do people reckon? Maybe there's something similar out there already, but this kind of practically-oriented group for newbies might well be a useful structure.
 
James.

On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]> wrote:
Many thanks for this brief discussion.  I've been learning about ontologies for about 4 years now, mostly in glorious isolation.  My mission is to build a design interface between biology and engineering, which I am doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/  Please visit, dissect and correct!).  A database was abysmally inadequate for the purpose whereas the ontology can introduce all the hierarchical relationships of biology very easily and cope with my changing visions of what I want, which will eventually be an AI system for robotics.  Unfortunately I don't think I've got much beyond the pizza level of understanding, despite grappling with books on the semantic web, bio-ontologies, etc.  I still can't understand the reasoning behind some of the more complex constructions used by Matthew Horridge in his "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put together in order to demonstrate some ideas.  But, Matthew, WHAT were
 the ideas???).  Is the Protege community ready for a higher level tutorial text on ontologies, their structures, and the reasons for preferring one or the other?  I'd love to have one.  I don't want to see yet another paper saying that we connected these boxes like this . . . or this . . .   What was the coding you used and why?  The pizza tutorial is excellent, but it's just the first course.  More!  More!!

Julian Vincent


On 18 Feb 2013, at 12:53, Phillip Lord wrote:

>
>
> The pizza ontology is an exemplar, of course. It's not really that
> useful, in and off it's own right.
>
> I think that the two main uses for ontologies, at from my experience,
> are providing a navigational structure. So, as you say, you could find a
> pizza by a number of different routes; and make it easier to maintain
> those routes. So, vegetarian pizza and cheese pizza get populated for
> you. Not so big an advantage for pizzas, but if you have a large parts
> catalogue, then much more use.
>
> The second main use is to annotate a large number of individuals for
> retrieval or statistical analysis. Think Dewey Decimal and a library.
> You can ask, which fields of knowledge have increased in popularity in
> the last 50 years.
>
> And the third use (cue spanish inquisition etc...) is to provide a
> transferable data model. So things like FOAF could be represented in
> relational data model, but how publish and transfer this knowledge?
> Strictly this is more a property of OWL rather than ontologies per se,
> but it's worth mentioning.
>
> Phil
>
>
>
> Tom Cloyd <[hidden email]> writes:
>
>> I want to say that this whole brief thread is of real value to me, as I stand
>> on the brink of diving into Protege. It is not entirely clear to me what can
>> be done with this. Coming from a decent understanding of relational databases,
>> and having played just a little with the pizza ontology, I got the same
>> impression which you offer here, James. Your background is much richer than
>> mine, relative to data modeling, so I really appreciate your thoughts.
>>
>> Tom
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332" value="+14352723332">(435) 272-3332
>>
>> On 02/15/2013 03:35 AM, James Naish wrote:
>>> Hi Sylvia,
>>>
>>> I'm not an expert ontologist but am a software engineer by trade who,
>>> after a good few years hearing about the wonders that ontologies have
>>> to offer, finally decided to take the plunge.
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Julian Vincent
I'd certainly welcome something along these lines.  I've spent more time rewriting my ontology than expanding it into the tool I want it to be.   I've spent ages experimenting with different ways of putting the information together and discovering the difficulties, trying to understand why some ontologies seem to say the same thing so many times and so (apparently) self-referentially and why others seem to say nothing at all!

Julian



On 18 Feb 2013, at 13:47, James Naish wrote:

> This discusssion has got me thinking. I'm sure there's plenty of great conferences out there on OWL, ontologies, DLs, etc for those who are in the know, but it seems to me there are plenty of us out there whose interests are predominantly practical rather than research-oriented. That is, even if we're involved in research, our contributions apply OWL but may not contribute to the OWL/DL/Ontology literature itself. Maybe there are enough of us about to get some form of hacking symposium together on a semi-regular (annual, bi-annual?) basis. I'm thinking something along the lines of the hacking clubs of the 70s in which people would turn up with practical problems, share solutions, and discuss visions for future pragmatic applications of new technologies, but with a focus on ontology-/OWL-based development.
>  
> What do people reckon? Maybe there's something similar out there already, but this kind of practically-oriented group for newbies might well be a useful structure.
>  
> James.
>
> On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]> wrote:
> Many thanks for this brief discussion.  I've been learning about ontologies for about 4 years now, mostly in glorious isolation.  My mission is to build a design interface between biology and engineering, which I am doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/  Please visit, dissect and correct!).  A database was abysmally inadequate for the purpose whereas the ontology can introduce all the hierarchical relationships of biology very easily and cope with my changing visions of what I want, which will eventually be an AI system for robotics.  Unfortunately I don't think I've got much beyond the pizza level of understanding, despite grappling with books on the semantic web, bio-ontologies, etc.  I still can't understand the reasoning behind some of the more complex constructions used by Matthew Horridge in his "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put together in order to demonstrate some ideas.  But, Matthew, WHAT wer
 e

>  the ideas???).  Is the Protege community ready for a higher level tutorial text on ontologies, their structures, and the reasons for preferring one or the other?  I'd love to have one.  I don't want to see yet another paper saying that we connected these boxes like this . . . or this . . .   What was the coding you used and why?  The pizza tutorial is excellent, but it's just the first course.  More!  More!!
>
> Julian Vincent
>
>
> On 18 Feb 2013, at 12:53, Phillip Lord wrote:
>
> >
> >
> > The pizza ontology is an exemplar, of course. It's not really that
> > useful, in and off it's own right.
> >
> > I think that the two main uses for ontologies, at from my experience,
> > are providing a navigational structure. So, as you say, you could find a
> > pizza by a number of different routes; and make it easier to maintain
> > those routes. So, vegetarian pizza and cheese pizza get populated for
> > you. Not so big an advantage for pizzas, but if you have a large parts
> > catalogue, then much more use.
> >
> > The second main use is to annotate a large number of individuals for
> > retrieval or statistical analysis. Think Dewey Decimal and a library.
> > You can ask, which fields of knowledge have increased in popularity in
> > the last 50 years.
> >
> > And the third use (cue spanish inquisition etc...) is to provide a
> > transferable data model. So things like FOAF could be represented in
> > relational data model, but how publish and transfer this knowledge?
> > Strictly this is more a property of OWL rather than ontologies per se,
> > but it's worth mentioning.
> >
> > Phil
> >
> >
> >
> > Tom Cloyd <[hidden email]> writes:
> >
> >> I want to say that this whole brief thread is of real value to me, as I stand
> >> on the brink of diving into Protege. It is not entirely clear to me what can
> >> be done with this. Coming from a decent understanding of relational databases,
> >> and having played just a little with the pizza ontology, I got the same
> >> impression which you offer here, James. Your background is much richer than
> >> mine, relative to data modeling, so I really appreciate your thoughts.
> >>
> >> Tom
> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >> Tom Cloyd / [hidden email] / (435) 272-3332
> >>
> >> On 02/15/2013 03:35 AM, James Naish wrote:
> >>> Hi Sylvia,
> >>>
> >>> I'm not an expert ontologist but am a software engineer by trade who,
> >>> after a good few years hearing about the wonders that ontologies have
> >>> to offer, finally decided to take the plunge.
> > _______________________________________________
> > p4-feedback mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Phillip Lord
In reply to this post by Julian Vincent

We have some stuff like this at ontogenesis.

http://ontogenesis.knowledgeblog.org/table-of-contents

There are articles on disjointness, higher order knowledge, covering
axioms, and my favourite, whether cyclists pay tax (by far the best
article on the site, to my mind).

Phil

Julian Vincent <[hidden email]> writes:

> Many thanks for this brief discussion. I've been learning about ontologies for
> about 4 years now, mostly in glorious isolation. My mission is to build a
> design interface between biology and engineering, which I am doing with
> Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/ Please visit, dissect
> and correct!). A database was abysmally inadequate for the purpose whereas the
> ontology can introduce all the hierarchical relationships of biology very
> easily and cope with my changing visions of what I want, which will eventually
> be an AI system for robotics. Unfortunately I don't think I've got much beyond
> the pizza level of understanding, despite grappling with books on the semantic
> web, bio-ontologies, etc. I still can't understand the reasoning behind some
> of the more complex constructions used by Matthew Horridge in his
> "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put
> together in order to demonstrate some ideas. But, Matthew, WHAT were
>  the ideas???). Is the Protege community ready for a higher level tutorial
> text on ontologies, their structures, and the reasons for preferring one or
> the other? I'd love to have one. I don't want to see yet another paper saying
> that we connected these boxes like this . . . or this . . . What was the
> coding you used and why? The pizza tutorial is excellent, but it's just the
> first course. More! More!!
>
> Julian Vincent
>
>
> On 18 Feb 2013, at 12:53, Phillip Lord wrote:
>
>>
>>
>> The pizza ontology is an exemplar, of course. It's not really that
>> useful, in and off it's own right.
>>
>> I think that the two main uses for ontologies, at from my experience,
>> are providing a navigational structure. So, as you say, you could find a
>> pizza by a number of different routes; and make it easier to maintain
>> those routes. So, vegetarian pizza and cheese pizza get populated for
>> you. Not so big an advantage for pizzas, but if you have a large parts
>> catalogue, then much more use.
>>
>> The second main use is to annotate a large number of individuals for
>> retrieval or statistical analysis. Think Dewey Decimal and a library.
>> You can ask, which fields of knowledge have increased in popularity in
>> the last 50 years.
>>
>> And the third use (cue spanish inquisition etc...) is to provide a
>> transferable data model. So things like FOAF could be represented in
>> relational data model, but how publish and transfer this knowledge?
>> Strictly this is more a property of OWL rather than ontologies per se,
>> but it's worth mentioning.
>>
>> Phil
>>
>>
>>
>> Tom Cloyd <[hidden email]> writes:
>>
>>> I want to say that this whole brief thread is of real value to me, as I stand
>>> on the brink of diving into Protege. It is not entirely clear to me what can
>>> be done with this. Coming from a decent understanding of relational
>>> databases,
>>> and having played just a little with the pizza ontology, I got the same
>>> impression which you offer here, James. Your background is much richer than
>>> mine, relative to data modeling, so I really appreciate your thoughts.
>>>
>>> Tom
>>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> Tom Cloyd / [hidden email] / (435) 272-3332
>>>
>>> On 02/15/2013 03:35 AM, James Naish wrote:
>>>> Hi Sylvia,
>>>>
>>>> I'm not an expert ontologist but am a software engineer by trade who,
>>>> after a good few years hearing about the wonders that ontologies have
>>>> to offer, finally decided to take the plunge.
>> _______________________________________________
>> p4-feedback mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
>

--
Phillip Lord,                           Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU                                
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Phillip Lord
In reply to this post by James Naish

The UK Ontology Network is an attempt along those lines

http://dream.inf.ed.ac.uk/events/ukont-13/

I've also thought about doing something similiar but virtual; the idea
would be that people write a short "position paper" (this is what I am
doing and why) and submit it; then we publish these all on the web one
after the next, and run a short-duration mailing list to discuss them.
Like a conference, without the travelling.

If there is interest, we can organise this.

Phil

James Naish <[hidden email]> writes:

> This discusssion has got me thinking. I'm sure there's plenty of great
> conferences out there on OWL, ontologies, DLs, etc for those who are in the
> know, but it seems to me there are plenty of us out there whose interests
> are predominantly practical rather than research-oriented. That is, even if
> we're involved in research, our contributions apply OWL but may not
> contribute to the OWL/DL/Ontology literature itself. Maybe there are enough
> of us about to get some form of hacking symposium together on a
> semi-regular (annual, bi-annual?) basis. I'm thinking something along the
> lines of the hacking clubs of the 70s in which people would turn up with
> practical problems, share solutions, and discuss visions for future
> pragmatic applications of new technologies, but with a focus on
> ontology-/OWL-based development.
>
> What do people reckon? Maybe there's something similar out there already,
> but this kind of practically-oriented group for newbies might well be a
> useful structure.
>
> James.
>
> On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]>wrote:
>
>> Many thanks for this brief discussion.  I've been learning about
>> ontologies for about 4 years now, mostly in glorious isolation.  My mission
>> is to build a design interface between biology and engineering, which I am
>> doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/ Please
>> visit, dissect and correct!). A database was abysmally inadequate
>> for the purpose whereas the ontology can introduce all the hierarchical
>> relationships of biology very easily and cope with my changing visions of
>> what I want, which will eventually be an AI system for robotics.
>>  Unfortunately I don't think I've got much beyond the pizza level of
>> understanding, despite grappling with books on the semantic web,
>> bio-ontologies, etc.  I still can't understand the reasoning behind some of
>> the more complex constructions used by Matthew Horridge in his
>> "Photography" ontology (which, he annoyingly wrote somewhere, he quickly
>> put together in order to demonstrate some ideas.  But, Matthew, WHAT were
>>  the ideas???).  Is the Protege community ready for a higher level
>> tutorial text on ontologies, their structures, and the reasons for
>> preferring one or the other?  I'd love to have one.  I don't want to see
>> yet another paper saying that we connected these boxes like this . . . or
>> this . . .   What was the coding you used and why?  The pizza tutorial is
>> excellent, but it's just the first course.  More!  More!!
>>
>> Julian Vincent
>>
>>
>> On 18 Feb 2013, at 12:53, Phillip Lord wrote:
>>
>> >
>> >
>> > The pizza ontology is an exemplar, of course. It's not really that
>> > useful, in and off it's own right.
>> >
>> > I think that the two main uses for ontologies, at from my experience,
>> > are providing a navigational structure. So, as you say, you could find a
>> > pizza by a number of different routes; and make it easier to maintain
>> > those routes. So, vegetarian pizza and cheese pizza get populated for
>> > you. Not so big an advantage for pizzas, but if you have a large parts
>> > catalogue, then much more use.
>> >
>> > The second main use is to annotate a large number of individuals for
>> > retrieval or statistical analysis. Think Dewey Decimal and a library.
>> > You can ask, which fields of knowledge have increased in popularity in
>> > the last 50 years.
>> >
>> > And the third use (cue spanish inquisition etc...) is to provide a
>> > transferable data model. So things like FOAF could be represented in
>> > relational data model, but how publish and transfer this knowledge?
>> > Strictly this is more a property of OWL rather than ontologies per se,
>> > but it's worth mentioning.
>> >
>> > Phil
>> >
>> >
>> >
>> > Tom Cloyd <[hidden email]> writes:
>> >
>> >> I want to say that this whole brief thread is of real value to me, as I
>> stand
>> >> on the brink of diving into Protege. It is not entirely clear to me
>> what can
>> >> be done with this. Coming from a decent understanding of relational
>> databases,
>> >> and having played just a little with the pizza ontology, I got the same
>> >> impression which you offer here, James. Your background is much richer
>> than
>> >> mine, relative to data modeling, so I really appreciate your thoughts.
>> >>
>> >> Tom
>> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >> Tom Cloyd / [hidden email] / (435) 272-3332
>> >>
>> >> On 02/15/2013 03:35 AM, James Naish wrote:
>> >>> Hi Sylvia,
>> >>>
>> >>> I'm not an expert ontologist but am a software engineer by trade who,
>> >>> after a good few years hearing about the wonders that ontologies have
>> >>> to offer, finally decided to take the plunge.
>> > _______________________________________________
>> > p4-feedback mailing list
>> > [hidden email]
>> > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>> _______________________________________________
>> p4-feedback mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

--
Phillip Lord,                           Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU                                
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creatingit?

Sylvia Breau
In reply to this post by James Naish

Hello Phillip, James, Tom, Samuel…

 

Thank you all for your enlightening answers. I think I have a better understanding now of how Protégé as a tool might fit into a search strategy.

 

I’m coming to ontology from the library and information science field. As a relative non-programmer who has tooled around with OWL, I couldn’t see how to get from the pizza ontology at the Protégé back end to a user-friendly front end that customers could interact with, the way they do now with websites that are driven by more conventional data structures, like relational databases or flat file structures.

 

If I understand correctly, there needs to be a query interface that is layered over the ontology, in order to access what’s “in” the ontology. This query interface could then interact with another layer it, which would be a web interface with search boxes and hyperlinked results, and so on. Please correct me and/or elaborate if my interpretation is off the mark!

 

I agree that a key strength of ontology is in the smart-web potential in navigation and browsing. From a search perspective, ontologies still need to rely heavily on the approaches to controlled vocabularies that have been in existence for years. However, ontology creates the possibility of resolving many, many of the limitations that are endemic to traditional taxonomies, thesauri, classification systems, and other controlled vocabularies. Ontology is powerful on both those fronts.

 

It does seem like a lot of the material on ontologies are still at the level of explaining relationship types and helping people to a conceptual understanding, but I agree with the comment that there aren’t good resources that will help laypeople what the ontology can actually do for them. While preparing my presentation about ontology as an alternative to my company’s current data architecture, I hunted for something I could use as an example to show what an ontology could look like to the end user. The choices were either things that were far too complex because the interface was the tool in which the ontology was built, or library online databases, which use the simple ontological approach embedded in Dublin Core. I ended up mocking something up, but I wasn’t sure my mockup was what would actually happen if we structure our data the way I envision.

 

This has been a fascinating process of discovery. I think the potential of ontology to create a quantum leap in findability is really here. Like many new fields that are unfolding right under our feet as we’re trying to learn them, there is confusion as we all sort of grope toward a future we can’t see yet. Happily, the Web itself is evolved enough where we can create user communities and be resources for each other. So much better than trying to learn Web creation in the early days of the Web itself, when the resources just weren’t there.

 

And, I didn’t expect the Spanish Inquisition either!

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of James Naish
Sent: Monday, February 18, 2013 8:18 AM
To: Protege 4.x support and discussion
Subject: Re: [p4-feedback] What to do with the pizza ontology after creatingit?

 

No one expected the Spanish Reference!

On Mon, Feb 18, 2013 at 12:53 PM, Phillip Lord <[hidden email]> wrote:



The pizza ontology is an exemplar, of course. It's not really that
useful, in and off it's own right.

I think that the two main uses for ontologies, at from my experience,
are providing a navigational structure. So, as you say, you could find a
pizza by a number of different routes; and make it easier to maintain
those routes. So, vegetarian pizza and cheese pizza get populated for
you. Not so big an advantage for pizzas, but if you have a large parts
catalogue, then much more use.

The second main use is to annotate a large number of individuals for
retrieval or statistical analysis. Think Dewey Decimal and a library.
You can ask, which fields of knowledge have increased in popularity in
the last 50 years.

And the third use (cue spanish inquisition etc...) is to provide a
transferable data model. So things like FOAF could be represented in
relational data model, but how publish and transfer this knowledge?
Strictly this is more a property of OWL rather than ontologies per se,
but it's worth mentioning.

Phil




Tom Cloyd <[hidden email]> writes:

> I want to say that this whole brief thread is of real value to me, as I stand
> on the brink of diving into Protege. It is not entirely clear to me what can
> be done with this. Coming from a decent understanding of relational databases,
> and having played just a little with the pizza ontology, I got the same
> impression which you offer here, James. Your background is much richer than
> mine, relative to data modeling, so I really appreciate your thoughts.
>
> Tom
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332">(435) 272-3332
>
> On 02/15/2013 03:35 AM, James Naish wrote:
>> Hi Sylvia,
>>
>> I'm not an expert ontologist but am a software engineer by trade who,
>> after a good few years hearing about the wonders that ontologies have
>> to offer, finally decided to take the plunge.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

 

DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

James Naish
In reply to this post by Phillip Lord
Phil,
 
This definitely sounds like a good idea. I wonder about the use of position papers though? Hacking clubs tend to work well because anyone could turn up and chip in in quite a free-flowing way - it can be chaotic, but fun and keeps the ideas flowing. How about finding a way to make this a bit less formal? For example, we could organise a series of web conferences (not sure what software we'd use but I'm sure there's something out there). It might work like this:
  1. People can register that they have an idea or a question that they want to raise;
  2. Time to be allocated to people in a round-robin fashion. Anyone can contribute, as long as they wait their turn and follow the other brief community rules;
  3. We split conferences into sessions: a questions session, an answers session. Question sessions would involve a brief contribution from the contributor, and more discussion amongst other participants in search of a good answer to the question. Answers sessions might involve longer presentations from users, with questions and brief contributions from the "floor".
This wouldn't rule out position papers, of course. I think building up a library of real-world case studies would actually be really useful. But it would complement this by providing an open and easy-access forum for fast, free-flowing, open debate about problems and solutions.
 
What do you think?
 
James.
 
On Mon, Feb 18, 2013 at 2:29 PM, Phillip Lord <[hidden email]> wrote:

The UK Ontology Network is an attempt along those lines

http://dream.inf.ed.ac.uk/events/ukont-13/

I've also thought about doing something similiar but virtual; the idea
would be that people write a short "position paper" (this is what I am
doing and why) and submit it; then we publish these all on the web one
after the next, and run a short-duration mailing list to discuss them.
Like a conference, without the travelling.

If there is interest, we can organise this.

Phil

James Naish <[hidden email]> writes:

> This discusssion has got me thinking. I'm sure there's plenty of great
> conferences out there on OWL, ontologies, DLs, etc for those who are in the
> know, but it seems to me there are plenty of us out there whose interests
> are predominantly practical rather than research-oriented. That is, even if
> we're involved in research, our contributions apply OWL but may not
> contribute to the OWL/DL/Ontology literature itself. Maybe there are enough
> of us about to get some form of hacking symposium together on a
> semi-regular (annual, bi-annual?) basis. I'm thinking something along the
> lines of the hacking clubs of the 70s in which people would turn up with
> practical problems, share solutions, and discuss visions for future
> pragmatic applications of new technologies, but with a focus on
> ontology-/OWL-based development.
>
> What do people reckon? Maybe there's something similar out there already,
> but this kind of practically-oriented group for newbies might well be a
> useful structure.
>
> James.
>
> On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]>wrote:
>
>> Many thanks for this brief discussion.  I've been learning about
>> ontologies for about 4 years now, mostly in glorious isolation.  My mission
>> is to build a design interface between biology and engineering, which I am
>> doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/ Please
>> visit, dissect and correct!). A database was abysmally inadequate
>> for the purpose whereas the ontology can introduce all the hierarchical
>> relationships of biology very easily and cope with my changing visions of
>> what I want, which will eventually be an AI system for robotics.
>>  Unfortunately I don't think I've got much beyond the pizza level of
>> understanding, despite grappling with books on the semantic web,
>> bio-ontologies, etc.  I still can't understand the reasoning behind some of
>> the more complex constructions used by Matthew Horridge in his
>> "Photography" ontology (which, he annoyingly wrote somewhere, he quickly
>> put together in order to demonstrate some ideas.  But, Matthew, WHAT were
>>  the ideas???).  Is the Protege community ready for a higher level
>> tutorial text on ontologies, their structures, and the reasons for
>> preferring one or the other?  I'd love to have one.  I don't want to see
>> yet another paper saying that we connected these boxes like this . . . or
>> this . . .   What was the coding you used and why?  The pizza tutorial is
>> excellent, but it's just the first course.  More!  More!!
>>
>> Julian Vincent
>>
>>
>> On 18 Feb 2013, at 12:53, Phillip Lord wrote:
>>
>> >
>> >
>> > The pizza ontology is an exemplar, of course. It's not really that
>> > useful, in and off it's own right.
>> >
>> > I think that the two main uses for ontologies, at from my experience,
>> > are providing a navigational structure. So, as you say, you could find a
>> > pizza by a number of different routes; and make it easier to maintain
>> > those routes. So, vegetarian pizza and cheese pizza get populated for
>> > you. Not so big an advantage for pizzas, but if you have a large parts
>> > catalogue, then much more use.
>> >
>> > The second main use is to annotate a large number of individuals for
>> > retrieval or statistical analysis. Think Dewey Decimal and a library.
>> > You can ask, which fields of knowledge have increased in popularity in
>> > the last 50 years.
>> >
>> > And the third use (cue spanish inquisition etc...) is to provide a
>> > transferable data model. So things like FOAF could be represented in
>> > relational data model, but how publish and transfer this knowledge?
>> > Strictly this is more a property of OWL rather than ontologies per se,
>> > but it's worth mentioning.
>> >
>> > Phil
>> >
>> >
>> >
>> > Tom Cloyd <[hidden email]> writes:
>> >
>> >> I want to say that this whole brief thread is of real value to me, as I
>> stand
>> >> on the brink of diving into Protege. It is not entirely clear to me
>> what can
>> >> be done with this. Coming from a decent understanding of relational
>> databases,
>> >> and having played just a little with the pizza ontology, I got the same
>> >> impression which you offer here, James. Your background is much richer
>> than
>> >> mine, relative to data modeling, so I really appreciate your thoughts.
>> >>
>> >> Tom
>> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >> Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332" value="+14352723332">(435) 272-3332
>> >>
>> >> On 02/15/2013 03:35 AM, James Naish wrote:
>> >>> Hi Sylvia,
>> >>>
>> >>> I'm not an expert ontologist but am a software engineer by trade who,
>> >>> after a good few years hearing about the wonders that ontologies have
>> >>> to offer, finally decided to take the plunge.
>> > _______________________________________________
>> > p4-feedback mailing list
>> > [hidden email]
>> > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>> _______________________________________________
>> p4-feedback mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

--
Phillip Lord,                           Phone: <a href="tel:%2B44%20%280%29%20191%20222%207827" value="+441912227827">+44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Phillip Lord

It's possible; however, my experience of web conferences is that they
have two flaws. First, a talking head on screen is extremely boring,
unless it's a really good stand-up comedian; ontologies are many things,
but a good source for gags is not one. Second, is that they are overly
sensitive to time zones; the NCBO Webinars sound interesting, for
instance, but at 6pm in evening?

When I say "position paper" -- I really mean a light-weight write up of
what you are doing. It doesn't have to be perfect, so long as it covers
things that you are think are interesting. But short is important. They
all go on the web, spread over a week or so. Then the discussion happens
by email; again, informal, taking as long as it takes.

I described the idea a while back on my blog; I called it NearCon
(nearly a conference).

http://www.russet.org.uk/blog/2144

The key point is that wanted to keep everything asynchronous; it should
work with people in different parts of the world and with different
abilities to commit time. Just a thought; be willing to try it, if there
is interest.

Phil

James Naish <[hidden email]> writes:

> This definitely sounds like a good idea. I wonder about the use of position
> papers though? Hacking clubs tend to work well because anyone could turn up
> and chip in in quite a free-flowing way - it can be chaotic, but fun and
> keeps the ideas flowing. How about finding a way to make this a bit less
> formal? For example, we could organise a series of web conferences (not
> sure what software we'd use but I'm sure there's something out there). It
> might work like this:
>
>    1. People can register that they have an idea or a question that they
>    want to raise;
>    2. Time to be allocated to people in a round-robin fashion. Anyone can
>    contribute, as long as they wait their turn and follow the other brief
>    community rules;
>    3. We split conferences into sessions: a questions session, an answers
>    session. Question sessions would involve a brief contribution from
>    the contributor, and more discussion amongst other participants in search
>    of a good answer to the question. Answers sessions might involve longer
>    presentations from users, with questions and brief contributions from the
>    "floor".
>
> This wouldn't rule out position papers, of course. I think building up a
> library of real-world case studies would actually be really useful. But it
> would complement this by providing an open and easy-access forum for fast,
> free-flowing, open debate about problems and solutions.
>
> What do you think?
>
> James.
>
> On Mon, Feb 18, 2013 at 2:29 PM, Phillip Lord
> <[hidden email]>wrote:
>
>>
>> The UK Ontology Network is an attempt along those lines
>>
>> http://dream.inf.ed.ac.uk/events/ukont-13/
>>
>> I've also thought about doing something similiar but virtual; the idea
>> would be that people write a short "position paper" (this is what I am
>> doing and why) and submit it; then we publish these all on the web one
>> after the next, and run a short-duration mailing list to discuss them.
>> Like a conference, without the travelling.
>>
>> If there is interest, we can organise this.
>>
>> Phil
>>
>> James Naish <[hidden email]> writes:
>>
>> > This discusssion has got me thinking. I'm sure there's plenty of great
>> > conferences out there on OWL, ontologies, DLs, etc for those who are in
>> the
>> > know, but it seems to me there are plenty of us out there whose interests
>> > are predominantly practical rather than research-oriented. That is, even
>> if
>> > we're involved in research, our contributions apply OWL but may not
>> > contribute to the OWL/DL/Ontology literature itself. Maybe there are
>> enough
>> > of us about to get some form of hacking symposium together on a
>> > semi-regular (annual, bi-annual?) basis. I'm thinking something along the
>> > lines of the hacking clubs of the 70s in which people would turn up with
>> > practical problems, share solutions, and discuss visions for future
>> > pragmatic applications of new technologies, but with a focus on
>> > ontology-/OWL-based development.
>> >
>> > What do people reckon? Maybe there's something similar out there already,
>> > but this kind of practically-oriented group for newbies might well be a
>> > useful structure.
>> >
>> > James.
>> >
>> > On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <
>> [hidden email]>wrote:
>> >
>> >> Many thanks for this brief discussion.  I've been learning about
>> >> ontologies for about 4 years now, mostly in glorious isolation.  My
>> mission
>> >> is to build a design interface between biology and engineering, which I
>> am
>> >> doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/Please
>> >> visit, dissect and correct!). A database was abysmally inadequate
>> >> for the purpose whereas the ontology can introduce all the hierarchical
>> >> relationships of biology very easily and cope with my changing visions
>> of
>> >> what I want, which will eventually be an AI system for robotics.
>> >>  Unfortunately I don't think I've got much beyond the pizza level of
>> >> understanding, despite grappling with books on the semantic web,
>> >> bio-ontologies, etc.  I still can't understand the reasoning behind
>> some of
>> >> the more complex constructions used by Matthew Horridge in his
>> >> "Photography" ontology (which, he annoyingly wrote somewhere, he quickly
>> >> put together in order to demonstrate some ideas.  But, Matthew, WHAT
>> were
>> >>  the ideas???).  Is the Protege community ready for a higher level
>> >> tutorial text on ontologies, their structures, and the reasons for
>> >> preferring one or the other?  I'd love to have one.  I don't want to see
>> >> yet another paper saying that we connected these boxes like this . . .
>> or
>> >> this . . .   What was the coding you used and why?  The pizza tutorial
>> is
>> >> excellent, but it's just the first course.  More!  More!!
>> >>
>> >> Julian Vincent
>> >>
>> >>
>> >> On 18 Feb 2013, at 12:53, Phillip Lord wrote:
>> >>
>> >> >
>> >> >
>> >> > The pizza ontology is an exemplar, of course. It's not really that
>> >> > useful, in and off it's own right.
>> >> >
>> >> > I think that the two main uses for ontologies, at from my experience,
>> >> > are providing a navigational structure. So, as you say, you could
>> find a
>> >> > pizza by a number of different routes; and make it easier to maintain
>> >> > those routes. So, vegetarian pizza and cheese pizza get populated for
>> >> > you. Not so big an advantage for pizzas, but if you have a large parts
>> >> > catalogue, then much more use.
>> >> >
>> >> > The second main use is to annotate a large number of individuals for
>> >> > retrieval or statistical analysis. Think Dewey Decimal and a library.
>> >> > You can ask, which fields of knowledge have increased in popularity in
>> >> > the last 50 years.
>> >> >
>> >> > And the third use (cue spanish inquisition etc...) is to provide a
>> >> > transferable data model. So things like FOAF could be represented in
>> >> > relational data model, but how publish and transfer this knowledge?
>> >> > Strictly this is more a property of OWL rather than ontologies per se,
>> >> > but it's worth mentioning.
>> >> >
>> >> > Phil
>> >> >
>> >> >
>> >> >
>> >> > Tom Cloyd <[hidden email]> writes:
>> >> >
>> >> >> I want to say that this whole brief thread is of real value to me,
>> as I
>> >> stand
>> >> >> on the brink of diving into Protege. It is not entirely clear to me
>> >> what can
>> >> >> be done with this. Coming from a decent understanding of relational
>> >> databases,
>> >> >> and having played just a little with the pizza ontology, I got the
>> same
>> >> >> impression which you offer here, James. Your background is much
>> richer
>> >> than
>> >> >> mine, relative to data modeling, so I really appreciate your
>> thoughts.
>> >> >>
>> >> >> Tom
>> >> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >> >> Tom Cloyd / [hidden email] / (435) 272-3332
>> >> >>
>> >> >> On 02/15/2013 03:35 AM, James Naish wrote:
>> >> >>> Hi Sylvia,
>> >> >>>
>> >> >>> I'm not an expert ontologist but am a software engineer by trade
>> who,
>> >> >>> after a good few years hearing about the wonders that ontologies
>> have
>> >> >>> to offer, finally decided to take the plunge.
>> >> > _______________________________________________
>> >> > p4-feedback mailing list
>> >> > [hidden email]
>> >> > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>> >>
>> >> _______________________________________________
>> >> p4-feedback mailing list
>> >> [hidden email]
>> >> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>> >>
>> >
>> > _______________________________________________
>> > p4-feedback mailing list
>> > [hidden email]
>> > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>> --
>> Phillip Lord,                           Phone: +44 (0) 191 222 7827
>> Lecturer in Bioinformatics,             Email:
>> [hidden email]
>> School of Computing Science,
>> http://homepages.cs.ncl.ac.uk/phillip.lord
>> Room 914 Claremont Tower,               skype: russet_apples
>> Newcastle University,                   twitter: phillord
>> NE1 7RU
>> _______________________________________________
>> p4-feedback mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

--
Phillip Lord,                           Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU                                
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creatingit?

Phillip Lord
In reply to this post by Sylvia Breau


Amigo is a nice example, over the Gene Ontology.

http://amigo.geneontology.org/cgi-bin/amigo/browse.cgi

Or neurolex.

http://neurolex.org/wiki/Category:Neuron

Sorry for the bio focus. It's what I am interested in.

I think we do need more less sciency interfaces. I don't know whether it
has an OWL underneath

http://www.bbc.co.uk/iplayer

has a very ontology navigational feel to it, which remaining easy to
use.

Phil

Sylvia Breau <[hidden email]> writes:

> It does seem like a lot of the material on ontologies are still at the level
> of explaining relationship types and helping people to a conceptual
> understanding, but I agree with the comment that there aren't good resources
> that will help laypeople what the ontology can actually do for them. While
> preparing my presentation about ontology as an alternative to my company's
> current data architecture, I hunted for something I could use as an example to
> show what an ontology could look like to the end user. The choices were either
> things that were far too complex because the interface was the tool in which
> the ontology was built, or library online databases, which use the simple
> ontological approach embedded in Dublin Core. I ended up mocking something up,
> but I wasn't sure my mockup was what would actually happen if we structure our
> data the way I envision.
>
>  
>
> This has been a fascinating process of discovery. I think the potential of
> ontology to create a quantum leap in findability is really here. Like many new
> fields that are unfolding right under our feet as we're trying to learn them,
> there is confusion as we all sort of grope toward a future we can't see yet.
> Happily, the Web itself is evolved enough where we can create user communities
> and be resources for each other. So much better than trying to learn Web
> creation in the early days of the Web itself, when the resources just weren't
> there.
>
>  
>
> And, I didn't expect the Spanish Inquisition either!
>
>  
>
> From: [hidden email] [mailto:[hidden email]] On Behalf Of James Naish
> Sent: Monday, February 18, 2013 8:18 AM
> To: Protege 4.x support and discussion
> Subject: Re: [p4-feedback] What to do with the pizza ontology after creatingit?
>
>  
>
> No one expected the Spanish Reference!
>
> On Mon, Feb 18, 2013 at 12:53 PM, Phillip Lord <[hidden email]> wrote:
>
>
>
> The pizza ontology is an exemplar, of course. It's not really that
> useful, in and off it's own right.
>
> I think that the two main uses for ontologies, at from my experience,
> are providing a navigational structure. So, as you say, you could find a
> pizza by a number of different routes; and make it easier to maintain
> those routes. So, vegetarian pizza and cheese pizza get populated for
> you. Not so big an advantage for pizzas, but if you have a large parts
> catalogue, then much more use.
>
> The second main use is to annotate a large number of individuals for
> retrieval or statistical analysis. Think Dewey Decimal and a library.
> You can ask, which fields of knowledge have increased in popularity in
> the last 50 years.
>
> And the third use (cue spanish inquisition etc...) is to provide a
> transferable data model. So things like FOAF could be represented in
> relational data model, but how publish and transfer this knowledge?
> Strictly this is more a property of OWL rather than ontologies per se,
> but it's worth mentioning.
>
> Phil
>
>
>
>
> Tom Cloyd <[hidden email]> writes:
>
>> I want to say that this whole brief thread is of real value to me, as I stand
>> on the brink of diving into Protege. It is not entirely clear to me what can
>> be done with this. Coming from a decent understanding of relational databases,
>> and having played just a little with the pizza ontology, I got the same
>> impression which you offer here, James. Your background is much richer than
>> mine, relative to data modeling, so I really appreciate your thoughts.
>>
>> Tom
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Tom Cloyd / [hidden email] / (435) 272-3332 <tel:%28435%29%20272-3332>
>>
>> On 02/15/2013 03:35 AM, James Naish wrote:
>>> Hi Sylvia,
>>>
>>> I'm not an expert ontologist but am a software engineer by trade who,
>>> after a good few years hearing about the wonders that ontologies have
>>> to offer, finally decided to take the plunge.
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
>  
>
>
>
> DISCLAIMER:
> This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
> exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
> communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
> delete the e-mail and destroy any copies of it. Thank you.
>
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

--
Phillip Lord,                           Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU                                
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Tom Cloyd
In reply to this post by James Naish
Pragmatics, not theoretics. That what many of us need. Examples of success usage which demonstrate what those of us who have not yet done it could do. THAT would be distinctly helpful.

James, your idea below is excellent. It would be a knowledge generator. We also need informal knowledge capture and preservation, and for that I merely propose a Content Management System (CMS) website. I'm developing one for an advocacy group I run, and so far the concept is looking very workable, relative to the object of informal knowledge presentation. We're going essentially to be creating a library which someone coming in to the area entirely new could use to get up to speed. There is a need for that. A wiki almost does that, but I think a CMS does it better. I'm not wed at all to this particular implementation of the idea knowledge preservation, however - not at all, but to the idea itself I am.

In this conceptualization, your proposed symposium, howsoever carried out, would be building a corpus of material of value to many more than the mere contributors/participants.

I'm very much in favor of this idea, and will definitely contribute if we can get it going.

Tom

On 02/18/2013 06:47 AM, James Naish wrote:
This discusssion has got me thinking. I'm sure there's plenty of great conferences out there on OWL, ontologies, DLs, etc for those who are in the know, but it seems to me there are plenty of us out there whose interests are predominantly practical rather than research-oriented. That is, even if we're involved in research, our contributions apply OWL but may not contribute to the OWL/DL/Ontology literature itself. Maybe there are enough of us about to get some form of hacking symposium together on a semi-regular (annual, bi-annual?) basis. I'm thinking something along the lines of the hacking clubs of the 70s in which people would turn up with practical problems, share solutions, and discuss visions for future pragmatic applications of new technologies, but with a focus on ontology-/OWL-based development.
 
What do people reckon? Maybe there's something similar out there already, but this kind of practically-oriented group for newbies might well be a useful structure.
 
James.

On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]> wrote:
Many thanks for this brief discussion.  I've been learning about ontologies for about 4 years now, mostly in glorious isolation.  My mission is to build a design interface between biology and engineering, which I am doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/  Please visit, dissect and correct!).  A database was abysmally inadequate for the purpose whereas the ontology can introduce all the hierarchical relationships of biology very easily and cope with my changing visions of what I want, which will eventually be an AI system for robotics.  Unfortunately I don't think I've got much beyond the pizza level of understanding, despite grappling with books on the semantic web, bio-ontologies, etc.  I still can't understand the reasoning behind some of the more complex constructions used by Matthew Horridge in his "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put together in order to demonstrate some ideas.  But, Matthew, WHAT were
 the ideas???).  Is the Protege community ready for a higher level tutorial text on ontologies, their structures, and the reasons for preferring one or the other?  I'd love to have one.  I don't want to see yet another paper saying that we connected these boxes like this . . . or this . . .   What was the coding you used and why?  The pizza tutorial is excellent, but it's just the first course.  More!  More!!

Julian Vincent


On 18 Feb 2013, at 12:53, Phillip Lord wrote:

>
>
> The pizza ontology is an exemplar, of course. It's not really that
> useful, in and off it's own right.
>
> I think that the two main uses for ontologies, at from my experience,
> are providing a navigational structure. So, as you say, you could find a
> pizza by a number of different routes; and make it easier to maintain
> those routes. So, vegetarian pizza and cheese pizza get populated for
> you. Not so big an advantage for pizzas, but if you have a large parts
> catalogue, then much more use.
>
> The second main use is to annotate a large number of individuals for
> retrieval or statistical analysis. Think Dewey Decimal and a library.
> You can ask, which fields of knowledge have increased in popularity in
> the last 50 years.
>
> And the third use (cue spanish inquisition etc...) is to provide a
> transferable data model. So things like FOAF could be represented in
> relational data model, but how publish and transfer this knowledge?
> Strictly this is more a property of OWL rather than ontologies per se,
> but it's worth mentioning.
>
> Phil
>
>
>
> Tom Cloyd <[hidden email]> writes:
>
>> I want to say that this whole brief thread is of real value to me, as I stand
>> on the brink of diving into Protege. It is not entirely clear to me what can
>> be done with this. Coming from a decent understanding of relational databases,
>> and having played just a little with the pizza ontology, I got the same
>> impression which you offer here, James. Your background is much richer than
>> mine, relative to data modeling, so I really appreciate your thoughts.
>>
>> Tom
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Tom Cloyd / [hidden email] / <a moz-do-not-send="true" href="tel:%28435%29%20272-3332" value="+14352723332">(435) 272-3332
>>
>> On 02/15/2013 03:35 AM, James Naish wrote:
>>> Hi Sylvia,
>>>
>>> I'm not an expert ontologist but am a software engineer by trade who,
>>> after a good few years hearing about the wonders that ontologies have
>>> to offer, finally decided to take the plunge.
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


-- 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: (435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creating it?

Sylvia Breau

Yes…pragmatics!! There is plenty of theoretical material generated from the research that is done in this field.

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Tom Cloyd
Sent: Monday, February 18, 2013 11:14 AM
To: Protege 4.x support and discussion
Subject: Re: [p4-feedback] What to do with the pizza ontology after creating it?

 

Pragmatics, not theoretics. That what many of us need. Examples of success usage which demonstrate what those of us who have not yet done it could do. THAT would be distinctly helpful.

James, your idea below is excellent. It would be a knowledge generator. We also need informal knowledge capture and preservation, and for that I merely propose a Content Management System (CMS) website. I'm developing one for an advocacy group I run, and so far the concept is looking very workable, relative to the object of informal knowledge presentation. We're going essentially to be creating a library which someone coming in to the area entirely new could use to get up to speed. There is a need for that. A wiki almost does that, but I think a CMS does it better. I'm not wed at all to this particular implementation of the idea knowledge preservation, however - not at all, but to the idea itself I am.

In this conceptualization, your proposed symposium, howsoever carried out, would be building a corpus of material of value to many more than the mere contributors/participants.

I'm very much in favor of this idea, and will definitely contribute if we can get it going.

Tom

On 02/18/2013 06:47 AM, James Naish wrote:

This discusssion has got me thinking. I'm sure there's plenty of great conferences out there on OWL, ontologies, DLs, etc for those who are in the know, but it seems to me there are plenty of us out there whose interests are predominantly practical rather than research-oriented. That is, even if we're involved in research, our contributions apply OWL but may not contribute to the OWL/DL/Ontology literature itself. Maybe there are enough of us about to get some form of hacking symposium together on a semi-regular (annual, bi-annual?) basis. I'm thinking something along the lines of the hacking clubs of the 70s in which people would turn up with practical problems, share solutions, and discuss visions for future pragmatic applications of new technologies, but with a focus on ontology-/OWL-based development.

 

What do people reckon? Maybe there's something similar out there already, but this kind of practically-oriented group for newbies might well be a useful structure.

 

James.

On Mon, Feb 18, 2013 at 1:40 PM, Julian Vincent <[hidden email]> wrote:

Many thanks for this brief discussion.  I've been learning about ontologies for about 4 years now, mostly in glorious isolation.  My mission is to build a design interface between biology and engineering, which I am doing with Protege 4 and TRIZ (https://wiki.bath.ac.uk/display/OOB/  Please visit, dissect and correct!).  A database was abysmally inadequate for the purpose whereas the ontology can introduce all the hierarchical relationships of biology very easily and cope with my changing visions of what I want, which will eventually be an AI system for robotics.  Unfortunately I don't think I've got much beyond the pizza level of understanding, despite grappling with books on the semantic web, bio-ontologies, etc.  I still can't understand the reasoning behind some of the more complex constructions used by Matthew Horridge in his "Photography" ontology (which, he annoyingly wrote somewhere, he quickly put together in order to demonstrate some ideas.  But, Matthew, WHAT were
 the ideas???).  Is the Protege community ready for a higher level tutorial text on ontologies, their structures, and the reasons for preferring one or the other?  I'd love to have one.  I don't want to see yet another paper saying that we connected these boxes like this . . . or this . . .   What was the coding you used and why?  The pizza tutorial is excellent, but it's just the first course.  More!  More!!

Julian Vincent



On 18 Feb 2013, at 12:53, Phillip Lord wrote:

>
>
> The pizza ontology is an exemplar, of course. It's not really that
> useful, in and off it's own right.
>
> I think that the two main uses for ontologies, at from my experience,
> are providing a navigational structure. So, as you say, you could find a
> pizza by a number of different routes; and make it easier to maintain
> those routes. So, vegetarian pizza and cheese pizza get populated for
> you. Not so big an advantage for pizzas, but if you have a large parts
> catalogue, then much more use.
>
> The second main use is to annotate a large number of individuals for
> retrieval or statistical analysis. Think Dewey Decimal and a library.
> You can ask, which fields of knowledge have increased in popularity in
> the last 50 years.
>
> And the third use (cue spanish inquisition etc...) is to provide a
> transferable data model. So things like FOAF could be represented in
> relational data model, but how publish and transfer this knowledge?
> Strictly this is more a property of OWL rather than ontologies per se,
> but it's worth mentioning.
>
> Phil
>
>
>
> Tom Cloyd <[hidden email]> writes:
>
>> I want to say that this whole brief thread is of real value to me, as I stand
>> on the brink of diving into Protege. It is not entirely clear to me what can
>> be done with this. Coming from a decent understanding of relational databases,
>> and having played just a little with the pizza ontology, I got the same
>> impression which you offer here, James. Your background is much richer than
>> mine, relative to data modeling, so I really appreciate your thoughts.
>>
>> Tom
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Tom Cloyd / [hidden email] / <a href="tel:%28435%29%20272-3332">(435) 272-3332
>>
>> On 02/15/2013 03:35 AM, James Naish wrote:
>>> Hi Sylvia,
>>>
>>> I'm not an expert ontologist but am a software engineer by trade who,
>>> after a good few years hearing about the wonders that ontologies have
>>> to offer, finally decided to take the plunge.
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback





_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback




-- 
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd, MS MA
Private practice Psychotherapist
Cedar City / St. George, Utah, U.S.A: (435) 272-3332
<< [hidden email] >> (email) << TomCloyd.com >> (website)
<< Sleightmind.com >> (mental health issues weblog)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

Reply | Threaded
Open this post in threaded view
|

Re: What to do with the pizza ontology after creatingit?

David Osumi-Sutherland
In reply to this post by Phillip Lord
Hi Sylvia,

Blowing my own trumpet here, but I think our site provides some interesting example of the kinds of things you can do with an ontology that would not be possible with a traditional RDBMS approach:

All of the queries on http://www.virtualflybrain.org use OWL-DL queries via the EL reasoner elk.  

The ontology we use serves as a query-able repository of information about fruit fly neuroanatomy.

For example, you can query for all the neurons that have some part in a part of the brain called the medulla:

This query uses property chains and object property hierarchy for for some basic spatial reasoning and other useful inferences over relations (object properties).  Easy to do in OWL. Very difficult (perhaps not possible) in SQL. And I could not have built the underlying ontology without using reasoners to automate classification and check consistency.

In addition, we have a growing body of OWL individuals, recording the properties of structures depicted in image data.   Here are some examples of a queries over this data.  Note that the same spatial reasoning is at work here as in the first query.


This last one shows inferred and asserted classifications using the ontology in the type column.

Hope this example helps,

Cheers,
David


On 18 Feb 2013, at 16:01, Phillip Lord wrote:



Amigo is a nice example, over the Gene Ontology.

http://amigo.geneontology.org/cgi-bin/amigo/browse.cgi

Or neurolex.

http://neurolex.org/wiki/Category:Neuron

Sorry for the bio focus. It's what I am interested in.

I think we do need more less sciency interfaces. I don't know whether it
has an OWL underneath

http://www.bbc.co.uk/iplayer

has a very ontology navigational feel to it, which remaining easy to
use.

Phil

Sylvia Breau <[hidden email]> writes:
It does seem like a lot of the material on ontologies are still at the level
of explaining relationship types and helping people to a conceptual
understanding, but I agree with the comment that there aren't good resources
that will help laypeople what the ontology can actually do for them. While
preparing my presentation about ontology as an alternative to my company's
current data architecture, I hunted for something I could use as an example to
show what an ontology could look like to the end user. The choices were either
things that were far too complex because the interface was the tool in which
the ontology was built, or library online databases, which use the simple
ontological approach embedded in Dublin Core. I ended up mocking something up,
but I wasn't sure my mockup was what would actually happen if we structure our
data the way I envision.



This has been a fascinating process of discovery. I think the potential of
ontology to create a quantum leap in findability is really here. Like many new
fields that are unfolding right under our feet as we're trying to learn them,
there is confusion as we all sort of grope toward a future we can't see yet.
Happily, the Web itself is evolved enough where we can create user communities
and be resources for each other. So much better than trying to learn Web
creation in the early days of the Web itself, when the resources just weren't
there.



And, I didn't expect the Spanish Inquisition either!



From: [hidden email] [mailto:[hidden email]] On Behalf Of James Naish
Sent: Monday, February 18, 2013 8:18 AM
To: Protege 4.x support and discussion
Subject: Re: [p4-feedback] What to do with the pizza ontology after creatingit?



No one expected the Spanish Reference!

On Mon, Feb 18, 2013 at 12:53 PM, Phillip Lord <[hidden email]> wrote:



The pizza ontology is an exemplar, of course. It's not really that
useful, in and off it's own right.

I think that the two main uses for ontologies, at from my experience,
are providing a navigational structure. So, as you say, you could find a
pizza by a number of different routes; and make it easier to maintain
those routes. So, vegetarian pizza and cheese pizza get populated for
you. Not so big an advantage for pizzas, but if you have a large parts
catalogue, then much more use.

The second main use is to annotate a large number of individuals for
retrieval or statistical analysis. Think Dewey Decimal and a library.
You can ask, which fields of knowledge have increased in popularity in
the last 50 years.

And the third use (cue spanish inquisition etc...) is to provide a
transferable data model. So things like FOAF could be represented in
relational data model, but how publish and transfer this knowledge?
Strictly this is more a property of OWL rather than ontologies per se,
but it's worth mentioning.

Phil




Tom Cloyd <[hidden email]> writes:

I want to say that this whole brief thread is of real value to me, as I stand
on the brink of diving into Protege. It is not entirely clear to me what can
be done with this. Coming from a decent understanding of relational databases,
and having played just a little with the pizza ontology, I got the same
impression which you offer here, James. Your background is much richer than
mine, relative to data modeling, so I really appreciate your thoughts.

Tom
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Cloyd / [hidden email] / (435) 272-3332 <tel:%28435%29%20272-3332>

On 02/15/2013 03:35 AM, James Naish wrote:
Hi Sylvia,

I'm not an expert ontologist but am a software engineer by trade who,
after a good few years hearing about the wonders that ontologies have
to offer, finally decided to take the plunge.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback





DISCLAIMER:
This e-mail is intended for the use of the addressee(s) only and may contain privileged, confidential, or proprietary information that is
exempt from disclosure under law. If you are not the intended recipient, please do not read, copy, use or disclose the contents of this
communication to others. Please notify the sender that you have received this e-mail in error by replying to the e-mail. Please then
delete the e-mail and destroy any copies of it. Thank you.


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback

--
Phillip Lord,                           Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics,             Email: [hidden email]
School of Computing Science,            http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower,               skype: russet_apples
Newcastle University,                   twitter: phillord
NE1 7RU                                 
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
12