WebCGM defects

[Lofton Henderson <lofton@rockynet.com>]

Dave (et al) --

Here are some routine defects to be corrected in WebCGM 1.0 Release 
2.  These result from reading WebCGM to extract Test Requirements.  Item(s) 
marked "QUESTIONABLE" are worth reading -- I think the solution is clear 
but I could also see someone disagreeing with that claim.

"Should" versus "shall"
---------------------------------

Language needs tightening to support conformance testing.

3.1.2.2, 2nd pgph:  "Viewers should...".  It should be "Viewers shall...", 
in order to be a firm conformance requirement.

This is repeated a lot.  "Should" should be changed to "shall" in:

3.1.2.2: I count 7 total occurrences, and all should be changed.

3.1.2.4, 1st pgph: 1 occurrence, change it.

3.1.3.1, 1st pgph: 1 occurrence, change it.  (Note:  because this is an 
example, it is non-normative in theory.  But I still think it ought to say 
"shall").

3.2.1.1:  5 occurrences, change all of 'em.

3.2.2.3:  1st paragraph, change single occurrence.

QUESTIONABLE.  3.2.2.3:  4th pgh, "should normally".  What did we intend to 
say here?  I think that "shall" is intended.  The "normally" implies that 
there are circumstances where CGM is the target media type and viewer 
behavior is not encoded in the fragment.  I don't remember any such 
exceptions.  Does anyone disagree?

3.4:  4 occurrences, change all of 'em (note related comment below about 
sentence containing 1st).

Other Editorial
---------------------

2.3.4, 5th bullet:  remove the ":".

2.3.5, 1st pgh:  'gprm' differs from 'cgmprim' in the final sub-picture of 
figure 1, although they reference the same thing.  One of the two must be 
changed to match the other.  'cgmprim' is used in section 3.3, so I think 
you want to change 'gprm'.

3.1.2.2, figure 3.2:  "_self" should be "_replace".

3.1.2.4, 2nd pgph, last sentence:  "Viewers should [shall] ignore all other 
Picture Behavior values."  I believe that the fragment EBNF of 3.1.1, plus 
the 1st sentence of this paragraph, imply that "all other ... values" in 
fact would constitute an invalid WebCGM instance.  Do we mean to prescribe 
how viewers should react to an invalid instance?  That would be a 
precedent, and contradicts the PPF T.26.7, which says "no", viewer error 
behavior is not addressed.  Therefore, I think the last sentence should be 
deleted.  Or, change it to say that all other Picture Behavior values 
constitute an invalid WebCGM instance, and viewer behavior is unspecified 
in such cases.  Does anyone disagree?

3.3, content model: a caveat is needed here.  Some APS attributes like 
linkURI can exist in WebCGM multiple times within one object.  Although 
there is current discussion about changing this (e.g., list of linkURI 
destinations in one attribute SDR in ProfileEd 1.1 or 2.0, and deprecating 
the multiple occurrences), this cannot be done for the editorial/defect 
Release 2.  Suggestion:  add caveat in a "note" after the content model.

CLARIFICATION NEEDED.  3.4, 3rd pgh:  Fractured sentence, I'm not sure what 
it is trying to say (also, probably change "should" to "shall").  Comments?

Regards,
Lofton.

*******************
Lofton Henderson
1919 Fourteenth St., #604
Boulder, CO   80302

Phone:  303-449-8728
Email:  lofton@rockynet.com
*******************