re[2]: [cgmo-webcgm] UL Review of Tests - WebCGM Compliance

[Lofton Henderson <lofton@rockynet.com>]

At 09:18 AM 3/26/2009 -0500, Don wrote:
>Rob,
>
>I don't see any reason why either.  I think ProfileEd 2.0 (or even 1.0 if it
>conforms) is ok given that we didn't really add anything in the 2.1 profile
>that I can re-collect.

That's true:
[1] 
http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-Appendix.html#webcgm_whatsnew

All the features in 2.1 are DOM or ACI feature.  There are no *additions* 
to metafile content in WebCGM 2.1.  However there are subtractions:
[2] 
http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-Conf.html#webcgm_conformance_deprObs

So for any 2.0 or 1.0 files would be good the ensure that they don't use 
any deprecated or obsolete features.  More strongly, 
s/good/essential/.  The conformance definitions of Ch.7 relieve 
2.1-compliant viewers of the obligation to support such features.  (In 
reality, I suspect most viewers still have their 1.0-compatibility, 
2.0-compatibility, etc).

Long story short, I don't feel strongly about this aspect, allowing 2.0- or 
1.0-WebCGM-valid instances in new tests (as long as they are properly 
vetted for dropped features).  I just think it is "tidier".

I do have stronger opinions about WebCGM-valid versus WebCGM-invalid files:
http://lists.oasis-open.org/archives/cgmo-webcgm/200903/msg00067.html

Regards,
-Lofton.

p.s.  A piece of good news from [1] -- changes to MetaCheck for 2.1 should 
be few and relatively easy.  It only needs to check for the illegal 
deprecated and obsoleted features.

>  >  Lofton,
>  >
>  >  I  would invert the question you pose below and ask it this  way.
>  >
>  >  Does  anyone see a reason why 2.1 should be required in all 
> new  metafiles?
>  >
>  >
>  >  I  don't, not only for the reasons that you allude to, but also for other
>  >  reasons  such as the fact that some of the new tests were added to fill
>  >  gaps in the test  suite, etc.  See my previous message for more detail.
>  >
>  >  http://lists.oasis-open.org/archives/cgmo-webcgm/200903/msg00056.html
>  >
>  >  Rob
>  >  -----Original Message-----
>  >  From: Lofton Henderson  [mailto:lofton@rockynet.com]
>  >  Sent: Monday, March 23, 2009 10:21  AM
>  >  To: 'WebCGM'
>  >  Subject: Re: [cgmo-webcgm] UL Review of  Tests - WebCGM Compliance
>
>  >  While we are sorting through the  details of Ulrich's reviews, here's my
>  >  opinion about one question Ulrich  asked...
>
>  >  At 03:55 PM 3/18/2009 +0100, =?us-ascii?Q?Ulrich_Lasche?=  wrote:
>  >  [...]
>  >  We have finished our review of the assigned test cases and will send  a
>  >  more detailed report soon.  One general remark though:
>  >
>  >  I  ran all 10 CGMs through MetaCheck with the WebCGM  option and all 
> files
>  >  were incompliant.  The first error  message:
>  >
>  >  Error 6589:  WebCGM 1.0 (2R) Profile Violation.
>  >  The METAFILE DESCRIPTION string is invalid; it lacks the  phrase
>  >  "ProfileEd:1.0" required by the Profile.
>  >
>  >  is clear  but still raises an issue.  All 10 files contain either
>  >  ProfileEd:1.0  or 2.0.  In the test files that I constructed manually I
>  >  put in  ProfileEd:2.1 (see rotateAPS.txt).  Is this  correct?
>  >  I think all of the CGM's ought to have 2.1 in  the metafile description,
>  >  just because they are new parts of the new WebCGM  2.1 Test Suite.  But,
>  >  strictly speaking, some of the 2.1 functionality --  for example all of
>  >  the ACI stuff, and the geometric transform stuff -- would  be equally
>  >  applicable to WebCGMs of any version, or indeed to any valid  metafile.
>
>  >  Thoughts?  Does anyone see a reason why we should  not have "2.1" in 
> all of
>  >  the new metafiles?
>
>  >  -Lofton.
>
>  >
>  >  While  this would be a cosmetic issue there are a number of additional
>  >  errors in  these files.  I am wondering if there is some compatibility
>  >  issue  between 1.0 and 2.1.  Could you please look into the MetaCheck
>  >  reports  (grouped into vendor specific zips) and tell me your  opinion?
>  >
>  >  Thanks &  regards
>  >  Ulrich
>
>---------------------------------------------------------------------
>To unsubscribe from this mail list, you must leave the OASIS TC that
>generates this mail.  Follow this link to all your TCs in OASIS at:
>https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php