Re[2]: [cgmo-webcgm] Questions

[Benoit Bezaire <benoit@itedo.com>]

Dave, did you make those changes to the draft or to some other
document that you are working on.  I need the latest draft in order to
make my changes.

-- 
 Benoit   mailto:benoit@itedo.com

 
Thursday, January 20, 2005, 6:30:05 AM, David wrote:

CDW> All,

CDW> See embedded comments:

CDW> Benoit Bezaire wrote on Tuesday, January 11, 2005 8:20 AM

CDW> Hi all,

CDW>   While reviewing chapter 5 of the latest spec, I came accross a few
CDW>   issues.  I wrote them down...

CDW>   a) Why are 'version' and 'filename' on the <webcgm> element
CDW>   #REQUIRED instead of #IMPLIED?


CDW> I don't completely remember but I think 'version' was not
CDW> the CGM version, but the version of the WebCGM (file?) (spec?)
CDW> (companion file?).  Help me remember.  I agree version could be
CDW> #IMPLIED.  The "filename" could be #IMPLIED, I guess, but it's
CDW> probably good practice for the user agent to always supply it. 
CDW> It's the link between the companion data and the WebCGM file.  In
CDW> my original architecture I had envisioned that the user agent
CDW> application would issue a load XML companion file data which
CDW> would point to the associated CGM file.  In later discussions
CDW> with the vendors, I came to understand that the user agent
CDW> issues separate commands to load the cgm and companion file (and
CDW> later apply)?  Since there is a potential many-to-one
CDW> relationship of companion files to a cgm file, there can't be a
CDW> pointer from the cgm file to the companion data.  The filename in
CDW> the companion file is the only link between them.


CDW>   b) In the DTD, the subpara element should allow for subparaAttEXT
CDW>   instead of paraAttEXT.

CDW> agreed...fixed

CDW>   c) Replace <!ELEMENT linkrui   EMPTY
CDW>   with <!ELEMENT linkuri   EMPTY

CDW> typo....fixed

CDW>   d) I think we agreed to add viewcontext and region to the WebCGM DTD,
CDW>   is that correct?  If that is right, we should modify the DTD.

CDW> I've added region and viewcontext to grobject, para,
CDW> subpara, bindById, and bindByName.  region and viewcontext don't
CDW> apply to layer, so some wording in the description of the
CDW> bindBy's needs to include some exclusionary words about applying
CDW> region and viewcontext to layer.  Right now, the attributes are
CDW> CDATA.  I know we wanted something more specific, but NUMBER is
CDW> not allowed in XML.  We could have more control over the content
CDW> if they were elements instead of attributes, but we still
CDW> couldn't specify that they be numbers.  XML Schema would allow
CDW> some control on that.

CDW> I'll be supplying r2 of the DTD for publication in the next WebCGM 2.0 document
CDW>  
CDW> thx...Dave