Re: [cgmo-webcgm] ISSUE: change WebCGM 1.0 degeneracy specs? [was: Initial values, % sub-issue]

[Lofton Henderson <lofton@rockynet.com>]

At 02:55 PM 5/18/2005 -0400, Benoit Bezaire wrote:
>Wednesday, May 18, 2005, 11:51:29 AM, Lofton wrote:
>[...]
>I'm not sure what you mean by 'keep track', but an implementation
>already has to keep track of the style state to be able to revert back
>to the original look (i.e., to implement clearStyleAttr() )

I think of this as a high-level operation, managing the data 
structures.  At a lower level, ultimately, you are going to hand line-width 
0.0 to your graphics back end or some low-level interface close to 
it.  Having written an implementation in the old days, I like the idea that 
both the zero-valued CGM attribute and the related zero-valued 
style-property resolve to the same thing, for the drawing back end.

But that's my (ancient) implementer bias, which isn't a trump 
argument.  (Fortunately, I don't have to implement what we're inventing 
anymore </wink>).

Personally, I think it's a bad idea, thinking about the user now, to say 
that zero-valued LINE WIDTH attribute means something different than 
zero-valued stroke-weight property, when they are so closely related:  lw' 
= sw * lw

I can see use cases (hey, I can make up a use case for anything!).  But I'm 
unconvinced that the use case is compelling enough make this new rule, 
"sw=0% means don't draw any line primitives in the APS".

Btw, does sw=0.0 (nvdc) have the same drawing result as sw=0%?  Or are we 
talking about a special rule only for %?

-Lofton.