[ome-devel] Experimenter OMEName in the XML Schema

[Ilya Goldberg]

Its come to our attention that we've inadvertently left OMEName as a 
required element in the OME XML Schema.  It really shouldn't be there 
at all, but at the very least it should be optional.  Specifically, we 
propose to make it optional now (preserving backward compatibility with 
existing documents), with the hope of deprecating it in a future 
version of the schema.

A bit of background:  Experimenters are not only specified in the 
schema, they contain information about actual users of OME software.  
The OMEName is the username they use for logging into OME.  Obviously 
its not something we ever want to import from or export to an XML 
document.  And we don't - the OMEName (and Password) are never exported 
even when exporting the Experimenter ST (they're not displayed when 
browsing experimenter details in the Web UI either).  If an 
Experimenter is imported in an XML document, any OMEName provided is 
ignored and the user can't log in until they get an OMEName and 
Password using the ome admin utility.  This ignoring of OMEName has 
been true for OME software since the Schema's initial release.

Why is it in the Schema at all then?  The XML Schema co-evolved with 
OME software, so there was a point where we didn't make a distinction 
between implementation-level user information and things that belong in 
an ontology.  Currently we do make this distinction, but the required 
OMEName remained in the schema unnoticed.  This was reported by Curtis 
Rueden as bug 497 (thanks Curtis!):
http://bugs.openmicroscopy.org.uk/show_bug.cgi?id=497

As these XML documents are generated by independently developed 
software, a few bugs of this nature are to be expected as this format 
is tested in a greater variety of real-world interoperability 
scenarios.  These documents have been tested this way prior to release 
(thanks to API and BitPlane), but as we all know, more testing reveals 
more bugs.

Note that this "bug" will not affect existing validating XML documents 
- it will only allow new documents without OMEName to be validated 
using Schema validators.  The Schema validator of reference is here, by 
the way:
http://www.w3.org/2001/03/webdata/xsv

If there are problems with making OMEName optional for now, please 
comment.  Unless I hear otherwise, OMEName will be made optional this 
Friday (May 6, 2005).

-Ilya