[XMLPipeDB-developer] Out-of-date OBO XML schema files

[John D. N. D. <do...@lm...>]

Here's the message that I sent to the GO folks regarding the DTD/XML  
schema issues that we've been having for the past year...

John David N. Dionisio, PhD
Assistant Professor, Computer Science
Loyola Marymount University


Begin forwarded message:

> I'd like to report an issue regarding the assorted schema files  
> (DTD, Relax NG, XSD) that are currently available for your OBO XML  
> format.  For reference, I'm using the files available here, as  
> linked from your main web site:
>
> 	http://geneontology.cvs.sourceforge.net/viewvc/geneontology/go-dev/xml/dtd/
>
> I'm one of the principals of the XMLPipeDB project (http://xmlpipedb.cs.lmu.edu 
> ), one of whose components is an automated OBO XML-to-relational  
> database loader that is based on an XML schema (either DTD or XSD).   
> This loader uses the schema file to automatically generate end-to- 
> end code for reading OBO XML into a relational database.
>
> Lately, we have observed that your daily term OBO XML downloads no  
> longer correspond to any of the schema files that I could find  
> online.  That is, the OBO XML files contain tags and properties that  
> are not present in any publicly available schema (whether DTD, Relax  
> NG, or XSD --- I'm aware that the DTD is marked as "deprecated," but  
> this becomes a non-issue when every schema file I could find appears  
> to be out of date).
>
> Because our software is automated, it relies on the completeness of  
> these schema files in order to work properly.  As a workaround, I  
> have been maintaining a manual, reverse-engineered version of your  
> DTD, adding any new tags or attributes that we have encountered so  
> far:
>
> 	http://xmlpipedb.svn.sourceforge.net/viewvc/xmlpipedb/trunk/godb/xsd/go_daily-obo-xml-manual.dtd?view=log
>
> This workaround has been sufficient as long as changes to the DTD  
> remain infrequent.  Lately, however, we have noticed further  
> additions (specifically, an apparently new created_by tag; our  
> manual file has not been updated yet to accommodate this change)  
> that have prompted us to seek your help --- after all, manual edits  
> of someone else's DTD or XSD file defeat the purpose of our  
> software.  Thus, I was wondering if it would be possible for you to  
> update either your OBO XML DTD or XSD file for completeness, such  
> that they match your posted daily term DB OBO XML files.  Or, if you  
> have been maintaining updated versions of these files outside of  
> your SourceForge repository and geneontology.org web site, please  
> let me know where to find them.