[eml-dev] [Fwd: Re: STMML and Ecological Metadata Language]

[Margaret O'Brien]

The stmml group is still interested in it's maintenance, and has 
suggested that the we might become its co-maintainers. In the interest 
of simplicity (and finishing the 210 release), we are probably better 
off sticking with stmml for now. If I interpret this correctly, it 
sounds like the simple solution is preferred; see my email to him, below.

Margaret
BTW, I should be able to test that fix next week (whether or not we 
decide to co-develop).


-------- Original Message --------
Subject: 	Re: STMML and Ecological Metadata Language
Date: 	Wed, 23 Jul 2008 20:21:32 +0100
From: 	Peter Murray-Rust <pm286 at cam.ac.uk>
Organization: 	Department of Chemistry, University of Cambridge
To: 	Margaret O'Brien <mob at msi.ucsb.edu>, Joe Townsend 
<jat45 at cam.ac.uk>, "Rzepa, Henry" <h.rzepa at imperial.ac.uk>, Nico Adams 
<na303 at cam.ac.uk>
References: 	<48863791.9020408 at msi.ucsb.edu>


Margaret,
How exciting to hear from you and that you have been using STMML for 
EML. I remember we corresponded some years ago and I'm glad it's working 
out.

I am not surprised you have discovered a bug - I am pleased that you 
haven't had problems up to now.

There is still a real need for something in the STMML space and we are 
slightly surprised not to have had more activitiy from other groups than 
yours. In practice we have continued to use a subset of STMML in CML 
(see www.sf.net/projects/cml) but we found it too difficult to use more 
than one namespace. So CML has probably slightly diverged from STMML.

If you feel STMML as current is valuable then I think we'd be very happy 
to explore that you become co-developers and maintainers. In which case 
it will be straightforward for us to make changes jointly.

In the current case we are not very keen on complicated content models 
and would be happy to see some relaxation.

P.

Margaret O'Brien wrote:
> Dear Dr. Murray-Rust,
>
> Ecological Metadata Language (EML) is a metadata standard developed by 
> the ecology community using XML schema language. For describing 
> measurement units, the EML schema imports stmml.xsd (v1.0). Since 
> 2003, EML has become widely used for descriptions of both ecological 
> and environmental datasets, and STMML has been an invaluable addition.
>
> EML's development group is currently upgrading the schema. In the 
> process, we have discovered a problem with the STMML schema which was 
> not previously apparent with most parsers.
>
> The stmml.xsd schema is non-deterministic. Here is the offending 
> section, starting at line 1708. The problem is the unbounded 
> "definition" next to the <xs:choice>
>
>     <xsd:sequence>
>         <xsd:element ref="definition" minOccurs="0"/>
>         <xsd:choice minOccurs="0" maxOccurs="unbounded">
>           <xsd:element ref="alternative"/>
>           <xsd:element ref="annotation"/>
>           <xsd:element ref="definition"/>
>           <xsd:element ref="description"/>
>           <xsd:element ref="enumeration"/>
>           <xsd:element ref="relatedEntry"/>
>         </xsd:choice>
>       </xsd:sequence>
>
>
> It appears that the schema can be made valid in one of several ways.
> 1. remove the minOccurs (ie, the "definition" element is required 
> before the choice) or
> 2. remove one of the 2 declarations for "definition", either the 
> first, or the one inside the choice group.
>
> Since the <xsd:choice> group nested inside the <xsd:sequence> allows 
> any combinations of these elements, it seems that removing the first 
> element declaration is the simplest solution.
>
> As I said earlier, this invalidity was not exposed by parsers that 
> were in general use several years ago. However, it is now caught by 
> XSV, xmlSpy (versions later than 2006), and Xerces (with the 
> schema-full-checking feature).
>
> We hesitate to make any changes to our local copy of the STMML schema 
> without first contacting its developers. We had thought that an 
> updated version might be available which addressed this problem, 
> although none has been found. We would like your recommendation for 
> the best course of action, or some insight on the intended use of this 
> part of the schema.
>
> For more information about Ecological Metadata Language, see
> http://knb.ecoinformatics.org/software/eml/eml-2.0.1
>
> Best regards,
> Margaret O'Brien
>
>
>
> ========================
> Margaret O'Brien
> Information Management
> Santa Barbara Coastal LTER
> Marine Science Institute
> University of California
> Santa Barbara, CA  93106-6150
>
> 805-893-2071
> mob at msi.ucsb.edu
> http://sbc.lternet.edu
> ========================



-- 

========================
Margaret O'Brien
SBC-LTER Information Management
Marine Science Institute
UCSB
Santa Barbara, CA  93106-6150

805-893-2071
mob at icess.ucsb.edu
http://sbc.lternet.edu
========================