deprecated class and property rendering

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

deprecated class and property rendering

Melanie Courtot-2
Hi,

When I open the file at http://purl.obolibrary.org/obo/pato.owl, there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.

Note that there are two alternative RDF syntax:
s rdf:type owl:DeprecatedClass . or
s rdf:type owl:DeprecatedProperty .
can be replaced by
s owl:deprecated "true"^^xsd:boolean .

Finally, in the attached test file:
- the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all
- the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.

Thanks,
Melanie





---
Mélanie Courtot
MSFHR/PCIRN trainee, TFL- BCCRC
675 West 10th Avenue
Vancouver, BC
V5Z 1L3, Canada








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

testDeprecated2.owl (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: deprecated class and property rendering

Timothy Redmond
On 06/15/2011 03:10 PM, Melanie Courtot wrote:
Hi,

When I open the file at http://purl.obolibrary.org/obo/pato.owl, there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.

This is very confusing and hard to work with.

There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.

Note that there are two alternative RDF syntax:
s rdf:type owl:DeprecatedClass . or 
s rdf:type owl:DeprecatedProperty .
can be replaced by
s owl:deprecated "true"^^xsd:boolean .

Finally, in the attached test file:
- the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all

Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class.

I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.  One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.

- the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.

I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).

-Timothy


Thanks,
Melanie

--- Mélanie Courtot MSFHR/PCIRN trainee, TFL- BCCRC 675 West 10th Avenue Vancouver, BC V5Z 1L3, Canada
_______________________________________________ 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

ParsingDeprecated.png (14K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: deprecated class and property rendering

Melanie Courtot-2
Hi Timothy,


>> Hi,
>>
>> When I open the file at
>> http://purl.obolibrary.org/obo/pato.owl
>> , there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
>> I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.
>>
>
> This is very confusing and hard to work with.
>
> There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.

In our "historical" files, such as OBI, we have an Obsolete Class, http://www.geneontology.org/formats/oboInOwl#ObsoleteClass. (this is meant to change and we are planning on relying on owl:deprecated in the future)
When deprecating, we assert the deprecated class under this one and remove all logical axioms (but for the subClass of ObsoleteClass)
I found that that was a nice way of rendering, as it allows me to search and visualize obsolete classes, but I can also close the whole obsolete hierarchy at once and ignore it if I want. I don't know how feasible and specification compliant something similar would be, but this may be an idea. I think (and Chris can correct me if I am wrong) in OBOEdit classes are additionally showing up in red.

>> Note that there are two alternative RDF syntax:
>> s rdf:type owl:DeprecatedClass . or
>> s rdf:type owl:DeprecatedProperty .
>> can be replaced by
>> s owl:deprecated "true"^^xsd:boolean .
>>
>> Finally, in the attached test file:
>> - the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all
>>
>
> Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class.

Thanks, I didn't know the view option; very useful!

Actually Chris had pointed me to the same spec yesterday, but I still expected the class (or something) to show up - as is, I am just "losing" my entity (though I agree, the OWLAPI followed the spec and did the right thing).


>
> I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.  

I am wondering why the specification was not written that way - I can't think of a case where that was not the intent. Maybe to be in sync with DeprecatedProperty?
I am not sure about adding an owl:Class though; when possible I prefer to stick to the spec; I am worried we'd run into other issues :)

> One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.

In this case maybe we should consider a global placeholder "deprecated entity"? Maybe in a different tab?

>
>> - the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.
>>
>
> I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).

I see - thanks for the explanation!
Melanie

>
> -Timothy
>
>
>> Thanks,
>> Melanie
>>
>>
>>
>>
>>
>>
>> ---
>> Mélanie Courtot
>> MSFHR/PCIRN trainee, TFL- BCCRC
>> 675 West 10th Avenue
>> Vancouver, BC
>> V5Z 1L3, Canada
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> p4-feedback mailing list
>>
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> <ParsingDeprecated.png>_______________________________________________
> 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: deprecated class and property rendering

Timothy Redmond

I was working on this issue and I just noticed a very funny thing about the pato ontology you provided.  Technically it has no deprecated classes!  The specification states that

An annotation with the owl:deprecated annotation property and the value equal to "true"^^xsd:boolean can be used to specify that an IRI is deprecated.

All of the annotation properties have type xsd:string and so they do not have this value.

I am not sure of this part but I went to the bottom and saw that it said:

<!-- Generated by the OWL API (version 3.0.0.1413) http://owlapi.sourceforge.net -->

I think that this might mean that it was generated by Protege 4.0 and this might explain this problem.

-Timothy


On 06/16/2011 09:29 AM, Melanie Courtot wrote:
Hi Timothy,


Hi,

When I open the file at 
http://purl.obolibrary.org/obo/pato.owl
, there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.

This is very confusing and hard to work with.

There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.
In our "historical" files, such as OBI, we have an Obsolete Class, http://www.geneontology.org/formats/oboInOwl#ObsoleteClass. (this is meant to change and we are planning on relying on owl:deprecated in the future)
When deprecating, we assert the deprecated class under this one and remove all logical axioms (but for the subClass of ObsoleteClass) 
I found that that was a nice way of rendering, as it allows me to search and visualize obsolete classes, but I can also close the whole obsolete hierarchy at once and ignore it if I want. I don't know how feasible and specification compliant something similar would be, but this may be an idea. I think (and Chris can correct me if I am wrong) in OBOEdit classes are additionally showing up in red.

Note that there are two alternative RDF syntax:
s rdf:type owl:DeprecatedClass . or 
s rdf:type owl:DeprecatedProperty .
can be replaced by
s owl:deprecated "true"^^xsd:boolean .

Finally, in the attached test file:
- the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all

Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class. 
Thanks, I didn't know the view option; very useful!

Actually Chris had pointed me to the same spec yesterday, but I still expected the class (or something) to show up - as is, I am just "losing" my entity (though I agree, the OWLAPI followed the spec and did the right thing).


I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.  
I am wondering why the specification was not written that way - I can't think of a case where that was not the intent. Maybe to be in sync with DeprecatedProperty?
I am not sure about adding an owl:Class though; when possible I prefer to stick to the spec; I am worried we'd run into other issues :)

One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.
In this case maybe we should consider a global placeholder "deprecated entity"? Maybe in a different tab?


        
- the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.

I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).
I see - thanks for the explanation!
Melanie

-Timothy


Thanks,
Melanie






---
Mélanie Courtot
MSFHR/PCIRN trainee, TFL- BCCRC
675 West 10th Avenue
Vancouver, BC
V5Z 1L3, Canada









_______________________________________________
p4-feedback mailing list

[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
<ParsingDeprecated.png>_______________________________________________
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: deprecated class and property rendering

Alan Ruttenberg-2
In reply to this post by Melanie Courtot-2
My recommendation would be a solution that covered a number of needs
at once. There is a general need for one to be able to create ad-hoc
groupings of terms for various purposes. This could be accomplished by
adding the following facility:

Allow the user to specify
1) An annotation property
2) A value or set of values (if you want to be generous leave room so
it is easy to extend this to a sparql query for terms)
3) The name of the grouping
4) The class/property or group below which the grouping should be placed
5) A checkbox indicating whether children of the term should be visible or not
6) Whether the setting should apply globally or be associated with the
active ontology (if the latter, save settings as ontology annotations)
7) A color/choice of icon for display
8) An override on what annotation property is used for rendering

The deprecated entity case is then a special case
1) owl:deprecated
2) "true"^^xsd:boolean
3) "Deprecated" (default, but changeable)
4) owl:Thing (default, but changeable)
5) n/a
6) global
7) Perhaps http://en.wikipedia.org/wiki/File:IEC5010_On_Off_Symbol.svg
8) no override.

Other cases useful for OBI would be to create groupings by curation
status, or within the deprecated to create groupings by obsolescence
reason, or to record seed terms for different modules, or to have
views that showed what part of OBI would look like to a specific
community. Since arbitrary grouping is such a useful thing, I'm sure
this facility would be of use to many other users. It also might
improve the quality of ontologies as they would not feel forced to
overload the class hierarchy in order to create user interfaces that
work for them.

-Alan

On Thu, Jun 16, 2011 at 12:29 PM, Melanie Courtot <[hidden email]> wrote:

> Hi Timothy,
>
>
>>> Hi,
>>>
>>> When I open the file at
>>> http://purl.obolibrary.org/obo/pato.owl
>>> , there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
>>> I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.
>>>
>>
>> This is very confusing and hard to work with.
>>
>> There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.
>
> In our "historical" files, such as OBI, we have an Obsolete Class, http://www.geneontology.org/formats/oboInOwl#ObsoleteClass. (this is meant to change and we are planning on relying on owl:deprecated in the future)
> When deprecating, we assert the deprecated class under this one and remove all logical axioms (but for the subClass of ObsoleteClass)
> I found that that was a nice way of rendering, as it allows me to search and visualize obsolete classes, but I can also close the whole obsolete hierarchy at once and ignore it if I want. I don't know how feasible and specification compliant something similar would be, but this may be an idea. I think (and Chris can correct me if I am wrong) in OBOEdit classes are additionally showing up in red.
>
>>> Note that there are two alternative RDF syntax:
>>> s rdf:type owl:DeprecatedClass . or
>>> s rdf:type owl:DeprecatedProperty .
>>> can be replaced by
>>> s owl:deprecated "true"^^xsd:boolean .
>>>
>>> Finally, in the attached test file:
>>> - the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all
>>>
>>
>> Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class.
>
> Thanks, I didn't know the view option; very useful!
>
> Actually Chris had pointed me to the same spec yesterday, but I still expected the class (or something) to show up - as is, I am just "losing" my entity (though I agree, the OWLAPI followed the spec and did the right thing).
>
>
>>
>> I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.
>
> I am wondering why the specification was not written that way - I can't think of a case where that was not the intent. Maybe to be in sync with DeprecatedProperty?
> I am not sure about adding an owl:Class though; when possible I prefer to stick to the spec; I am worried we'd run into other issues :)
>
>> One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.
>
> In this case maybe we should consider a global placeholder "deprecated entity"? Maybe in a different tab?
>
>>
>>> - the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.
>>>
>>
>> I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).
>
> I see - thanks for the explanation!
> Melanie
>
>>
>> -Timothy
>>
>>
>>> Thanks,
>>> Melanie
>>>
>>>
>>>
>>>
>>>
>>>
>>> ---
>>> Mélanie Courtot
>>> MSFHR/PCIRN trainee, TFL- BCCRC
>>> 675 West 10th Avenue
>>> Vancouver, BC
>>> V5Z 1L3, Canada
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> p4-feedback mailing list
>>>
>>> [hidden email]
>>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>
>> <ParsingDeprecated.png>_______________________________________________
>> 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: deprecated class and property rendering

David Osumi-Sutherland
I like this suggestion.  Filtering the class hierarchy on annotation properties can be very useful.  A similar suggestion came up in recent discussion among OBO-Edit users who are beginning to migrate to Protege.

 
On 21 Jun 2011, at 17:36, Alan Ruttenberg wrote:

> My recommendation would be a solution that covered a number of needs
> at once. There is a general need for one to be able to create ad-hoc
> groupings of terms for various purposes. This could be accomplished by
> adding the following facility:
>
> Allow the user to specify
> 1) An annotation property
> 2) A value or set of values (if you want to be generous leave room so
> it is easy to extend this to a sparql query for terms)
> 3) The name of the grouping
> 4) The class/property or group below which the grouping should be placed
> 5) A checkbox indicating whether children of the term should be visible or not
> 6) Whether the setting should apply globally or be associated with the
> active ontology (if the latter, save settings as ontology annotations)
> 7) A color/choice of icon for display
> 8) An override on what annotation property is used for rendering
>
> The deprecated entity case is then a special case
> 1) owl:deprecated
> 2) "true"^^xsd:boolean
> 3) "Deprecated" (default, but changeable)
> 4) owl:Thing (default, but changeable)
> 5) n/a
> 6) global
> 7) Perhaps http://en.wikipedia.org/wiki/File:IEC5010_On_Off_Symbol.svg
> 8) no override.
>
> Other cases useful for OBI would be to create groupings by curation
> status, or within the deprecated to create groupings by obsolescence
> reason, or to record seed terms for different modules, or to have
> views that showed what part of OBI would look like to a specific
> community. Since arbitrary grouping is such a useful thing, I'm sure
> this facility would be of use to many other users. It also might
> improve the quality of ontologies as they would not feel forced to
> overload the class hierarchy in order to create user interfaces that
> work for them.
>
> -Alan
>
> On Thu, Jun 16, 2011 at 12:29 PM, Melanie Courtot <[hidden email]> wrote:
>> Hi Timothy,
>>
>>
>>>> Hi,
>>>>
>>>> When I open the file at
>>>> http://purl.obolibrary.org/obo/pato.owl
>>>> , there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
>>>> I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property > deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.
>>>>
>>>
>>> This is very confusing and hard to work with.
>>>
>>> There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.
>>
>> In our "historical" files, such as OBI, we have an Obsolete Class, http://www.geneontology.org/formats/oboInOwl#ObsoleteClass. (this is meant to change and we are planning on relying on owl:deprecated in the future)
>> When deprecating, we assert the deprecated class under this one and remove all logical axioms (but for the subClass of ObsoleteClass)
>> I found that that was a nice way of rendering, as it allows me to search and visualize obsolete classes, but I can also close the whole obsolete hierarchy at once and ignore it if I want. I don't know how feasible and specification compliant something similar would be, but this may be an idea. I think (and Chris can correct me if I am wrong) in OBOEdit classes are additionally showing up in red.
>>
>>>> Note that there are two alternative RDF syntax:
>>>> s rdf:type owl:DeprecatedClass . or
>>>> s rdf:type owl:DeprecatedProperty .
>>>> can be replaced by
>>>> s owl:deprecated "true"^^xsd:boolean .
>>>>
>>>> Finally, in the attached test file:
>>>> - the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all
>>>>
>>>
>>> Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class.
>>
>> Thanks, I didn't know the view option; very useful!
>>
>> Actually Chris had pointed me to the same spec yesterday, but I still expected the class (or something) to show up - as is, I am just "losing" my entity (though I agree, the OWLAPI followed the spec and did the right thing).
>>
>>
>>>
>>> I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.
>>
>> I am wondering why the specification was not written that way - I can't think of a case where that was not the intent. Maybe to be in sync with DeprecatedProperty?
>> I am not sure about adding an owl:Class though; when possible I prefer to stick to the spec; I am worried we'd run into other issues :)
>>
>>> One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.
>>
>> In this case maybe we should consider a global placeholder "deprecated entity"? Maybe in a different tab?
>>
>>>
>>>> - the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.
>>>>
>>>
>>> I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).
>>
>> I see - thanks for the explanation!
>> Melanie
>>
>>>
>>> -Timothy
>>>
>>>
>>>> Thanks,
>>>> Melanie
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> ---
>>>> Mélanie Courtot
>>>> MSFHR/PCIRN trainee, TFL- BCCRC
>>>> 675 West 10th Avenue
>>>> Vancouver, BC
>>>> V5Z 1L3, Canada
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> p4-feedback mailing list
>>>>
>>>> [hidden email]
>>>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>>
>>> <ParsingDeprecated.png>_______________________________________________
>>> 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

David Osumi-Sutherland, PhD
Ontologist
Virtual Fly Brain / FlyBase
Department of Genetics
University of Cambridge
Downing Street
Cambridge, CB2 3EH, UK
+44 (0)1223 333 963
http://www.virtualflybrain.org




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

Re: deprecated class and property rendering

Timothy Redmond
In reply to this post by Alan Ruttenberg-2

Sounds good.  My one comment on this is that this sounds like a good
plugin.  This type of functionality probably doesn't belong in the
core.  Another advantage of this as a plugin is that it becomes a
possible student project.

-Timothy


On 06/21/2011 09:36 AM, Alan Ruttenberg wrote:

> My recommendation would be a solution that covered a number of needs
> at once. There is a general need for one to be able to create ad-hoc
> groupings of terms for various purposes. This could be accomplished by
> adding the following facility:
>
> Allow the user to specify
> 1) An annotation property
> 2) A value or set of values (if you want to be generous leave room so
> it is easy to extend this to a sparql query for terms)
> 3) The name of the grouping
> 4) The class/property or group below which the grouping should be placed
> 5) A checkbox indicating whether children of the term should be visible or not
> 6) Whether the setting should apply globally or be associated with the
> active ontology (if the latter, save settings as ontology annotations)
> 7) A color/choice of icon for display
> 8) An override on what annotation property is used for rendering
>
> The deprecated entity case is then a special case
> 1) owl:deprecated
> 2) "true"^^xsd:boolean
> 3) "Deprecated" (default, but changeable)
> 4) owl:Thing (default, but changeable)
> 5) n/a
> 6) global
> 7) Perhaps http://en.wikipedia.org/wiki/File:IEC5010_On_Off_Symbol.svg
> 8) no override.
>
> Other cases useful for OBI would be to create groupings by curation
> status, or within the deprecated to create groupings by obsolescence
> reason, or to record seed terms for different modules, or to have
> views that showed what part of OBI would look like to a specific
> community. Since arbitrary grouping is such a useful thing, I'm sure
> this facility would be of use to many other users. It also might
> improve the quality of ontologies as they would not feel forced to
> overload the class hierarchy in order to create user interfaces that
> work for them.
>
> -Alan
>
> On Thu, Jun 16, 2011 at 12:29 PM, Melanie Courtot<[hidden email]>  wrote:
>> Hi Timothy,
>>
>>
>>>> Hi,
>>>>
>>>> When I open the file at
>>>> http://purl.obolibrary.org/obo/pato.owl
>>>> , there is no rendering difference between the (long) list of deprecated classes and the non deprecated root of the hierarchy, which makes it confusing when browsing.
>>>> I was wondering if it would be possible to have a rendering option to that effect? Or is there something already that I am not aware of? I tried "render by annotation property>  deprecated" but this just gives me either "true" or the numerical ID of the class, which is not very helpful either.
>>>>
>>> This is very confusing and hard to work with.
>>>
>>> There is a GForge for this item already (http://goo.gl/ODKS9).  It is actually two separate gforges in one which I don't really like.  It is suggested there that the the deprecated items can be simply hidden from view.  But even if this option is there it probably isn't enough.  I will look at this problem and see what I can do.
>> In our "historical" files, such as OBI, we have an Obsolete Class, http://www.geneontology.org/formats/oboInOwl#ObsoleteClass. (this is meant to change and we are planning on relying on owl:deprecated in the future)
>> When deprecating, we assert the deprecated class under this one and remove all logical axioms (but for the subClass of ObsoleteClass)
>> I found that that was a nice way of rendering, as it allows me to search and visualize obsolete classes, but I can also close the whole obsolete hierarchy at once and ignore it if I want. I don't know how feasible and specification compliant something similar would be, but this may be an idea. I think (and Chris can correct me if I am wrong) in OBOEdit classes are additionally showing up in red.
>>
>>>> Note that there are two alternative RDF syntax:
>>>> s rdf:type owl:DeprecatedClass . or
>>>> s rdf:type owl:DeprecatedProperty .
>>>> can be replaced by
>>>> s owl:deprecated "true"^^xsd:boolean .
>>>>
>>>> Finally, in the attached test file:
>>>> - the rendering of the 2 classes is different, with the second class using the syntax owl:DeprecatedClass not showing up at all
>>>>
>>> Actually I think that the OWL api might be doing the right thing here.  (I don't know that I like it though.)  If you look at at the parser specification (http://goo.gl/sMdeZ) it says that the owl:DeprecatedClass declaration be converted into an annotation property.  (I found the relevant line by searching forward for deprecated).  I attached a screenshot of the relevant specification.  This is what the OWL api did.  If you look at the functional syntax view (Window->Views->Ontology Views->OWL Functional Syntax) for your example, you will see that the annotation property is there.  But there simply is no declaration that the IRI is the IRI for a class.
>> Thanks, I didn't know the view option; very useful!
>>
>> Actually Chris had pointed me to the same spec yesterday, but I still expected the class (or something) to show up - as is, I am just "losing" my entity (though I agree, the OWLAPI followed the spec and did the right thing).
>>
>>
>>> I am wondering if this is a case where the OWL api should ignore the specification and simply do what was clearly intended - add a owl:Class type.
>> I am wondering why the specification was not written that way - I can't think of a case where that was not the intent. Maybe to be in sync with DeprecatedProperty?
>> I am not sure about adding an owl:Class though; when possible I prefer to stick to the spec; I am worried we'd run into other issues :)
>>
>>> One reason not to do this is that in the owl:DeprecatedProperty case the OWL api has no such fix; it is not known what type of property is involved.
>> In this case maybe we should consider a global placeholder "deprecated entity"? Maybe in a different tab?
>>
>>>> - the class which has label set to "value" shows highlighted in pink. Is "value" some sort of reserved vocabulary word? I can imagine it is, but I didn't find any documentation to that effect.
>>>>
>>> I knew about this bug and thought that it had a GForge ticket.  "value" is highlighted because it happens to be a keyword in the Manchester OWL syntax.  Probably if I figure out the deprecated rendering bug above I will know how to fix this one.  I couldn't find the GForge ticket so I added it (http://goo.gl/TfrBi).
>> I see - thanks for the explanation!
>> Melanie
>>
>>> -Timothy
>>>
>>>
>>>> Thanks,
>>>> Melanie
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> ---
>>>> Mélanie Courtot
>>>> MSFHR/PCIRN trainee, TFL- BCCRC
>>>> 675 West 10th Avenue
>>>> Vancouver, BC
>>>> V5Z 1L3, Canada
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> p4-feedback mailing list
>>>>
>>>> [hidden email]
>>>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>>> <ParsingDeprecated.png>_______________________________________________
>>> 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