RE: [cgmo-webcgm] Model Profile version in Chapter 6?

[Robert Orosz <roboro@AUTO-TROL.com>]

Lofton,

GREXCHANGE 2.9 has the same error with respect to NUBS and NURBS in the
Model Profile :-(.

I would prefer that we require the clamped form.  Implementation is easier.

Rob

-----Original Message-----
From: Lofton Henderson [mailto:lofton@rockynet.com]
Sent: Friday, August 26, 2005 5:32 PM
To: Robert Orosz; cgmo-webcgm@lists.oasis-open.org
Subject: Re: [cgmo-webcgm] Model Profile version in Chapter 6?


Hi Rob,

Taking your questions a little out of order:

At 12:16 PM 8/26/2005 -0600, Robert Orosz wrote:
>[...]
>I have a vague recollection that an earlier version of the Model Profile
did
>not require clamped splines which leads me to wonder, are we listing the
>latest Model Profile in Chapter 6?

Yikes!

I just checked my paper copy [1] of ISO 8632:1999 Part 1, Annex I 
(PPF).  In the MP, it is the clamped form.  This agrees with the ISO-hosted 
online (PDF) version at [2].  It agrees also with p1.htm, which is an 
unofficial HTML version that was given to me by ISO and which I have in my 
old editor's directories.  So "clamped" must be considered to be the latest 
MP PPF.

[1] paper CGM:1999
[2] 
http://isotc.iso.org/livelink/livelink/fetch/2000/2489/Ittf_Home/PubliclyAva
ilableStandards.htm
[3] p1.htm

It does NOT agree with the MP of WebCGM 1.0 [4], nor with draft 2.0 [5]:

[4] http://www.w3.org/TR/REC-WebCGM/REC-04-CGM-Profile.html
[5] 
http://www.oasis-open.org/committees/download.php/14161/WebCGM20-20050804.zi
p

It's hard to guess where this happened.  But I recall, in the foggy past 
(probably about 6 years ago), we changed the PPF format from check- box 
style to YES/NO on the Required/Permitted/Prohibited line (to avoid the 
funny empty-box and checked-box characters in the WebCGM HTML 
text).  Someone must have made the error then (let's blame John!).

The good news:  since 1.0 didn't allow NUBS/NURBS, the error in 
transcription of the MP doesn't affect 1.0.  We do need to consider whether 
we want to restrict to the clamped form for 2.0 (seems reasonable -- the 
clamped cubic NUBS/NURBS would be a subset of all cubic ones, which should 
mean easier viewer implementation.  Does "clamped" suffice for the 
potential users?).

The bad news:  did ATA propagate that MP error?  and, are there other MP 
PPF errors?

>I'll try and review more of the PPF to
>see if there are other inaccuracies.

That would be terrific -- very timely and much needed now.  (If you are 
unable to do it, please let me know -- we'll need to give it as an AI 
assignment to someone.)

>I have an open action item to generate some test cases with NUBS for the
>test suite.  In preparation for doing so, I went to sections T.19.24 and
>T.19.25 of the WebCGM PPF to review the requirements.  Both sections state
>"Same as Model Profile: Yes", however, the Model Profile half of each
>section is incorrect.  The Model Profile requires clamped splines which
>means that the first four knots must be identical and the last four knots
>must be identical.  This requirement is not listed in the WebCGM 2.0 draft
>PPF.  Section 6.1 does say, "In case of discrepancies, the Model Profile in
>ISO/IEC 8632:1999 shall take precedence." however, I suspect most people
>won't take the time to cross check with the Model Profile.

Probably not.  However, as I said -- the normative content of WebCGM 1.0 is 
unaffected.  And in any case, the "discrepancies" clause means that 
(formally) "clamped" rules in the MP column.  We will of course fix the 
content of the MP in the 2.0 PPF.  If anyone thinks that 2.0 should NOT be 
"same as MP" (i.e., that we should remove the "clamped" restriction of the 
MP), he should bring it up as an issue.

Thanks,
-Lofton.