Re: proposed changes to ACI chapter

I approve the proposed changes.

Dave

On Fri, May 22, 2009 at 7:53 AM, Lofton Henderson <lofton@rockynet.com>wrote:

> All--
>
> Action:
> -----
> Please reply on list to this message:  approve proposed changes; or
> disapprove and state your reasons.
>
> What:
> -----
> Following the Wednesday telecon, Dave and a small quality verification task
> team have looked at Ch.9, the ACI specification and its dtd.  The good news
> is that it appears to be sound now.  However, the group recommends some
> simple changes before 2nd LCWD review.
>
> Details:
> -----
> For seven elements under the (XML) defaultAttributes element, the (XML)
> attributes associated with the element are optional and a default is given.
>  Consider for example the lineJoin ACI element [1]:
>
> [1]
> http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-Config.html#ACI-linejoin
>
> <!ELEMENT lineJoin EMPTY >
> <!ATTLIST lineJoin
>          lineJoinInd ( 1 | 2 | 3 | 4 ) "1"
> >
>
> While this is not incorrect or illegal, strictly speaking, on the other
> hand it is not sensible when viewed from the perspective of the purpose of
> the defaultAttributes element.  For example, it would allow:
>
> <lineJoin></lineJoin>
>
> and that construct would mean that the viewer should use the dtd's default
> value for lineJoinInd, "1" ('unspecified'),  in rendering.  That value is in
> fact the CGM:1999 default value for the LINE JOIN element, so the element is
> essentially an no-op.
>
> But  the ACI defaultAttributes element was added to webcgm precisely to tie
> down such underspecified values, and enable uniform viewer results when
> dealing with the underspecified CGM:1999 defaults.  I.e., if someone is
> putting a lineJoin element into the ACI file, then the goal presumably is to
> nail down "unspecified" and tell the viewer to use a particular one of the
> other 3 well-defined values (which are the legal ways to handle
> "unspecified").  So the optionality and defaulting of the lineJoinInd
> attribute does not make sense for this group of elements.
>
> It makes more sense to require the lineJoinInd attribute whenever the
> lineJoin element is present.
>
> Proposal:
> -----
> Change the specification to:
>
> <!ELEMENT lineJoin EMPTY >
> <!ATTLIST lineJoin
>          lineJoinInd ( 1 | 2 | 3 | 4 ) #REQUIRED
> >
>
> If the ACI has a lineJoin element, it must have a lineJoinInd attribute.
>
> This proposal would similarly be applied to:  lineCap, edgeCap, lineJoin,
> edgeJoin, lineTypeCont, edgeTypeCont, restrTextType.
>
> Additional:
> -----
> Each of these elements also has something like this at the end of it
> definition:  "The default value is '1' or 'unspecified'."
>
> I would change these occurrences to:
> "Note (informative):  in the CGM:1999 specification, the default value for
> the associated CGM LINE JOIN Attribute element is "1 (unspecified)."
>
> It would also be acceptable to simply delete them (speak up if you have a
> preference here.)
>
> Tests:
> -----
> No ACI tests are affected.
>
> Summary:
> -----
> Please reply with:  approval of this proposal; or, disapproval and your
> reasons.
>
> Regards,
> -Lofton.
>
>
>