W3C home > Mailing lists > Public > www-svg@w3.org > September 2005

Re: [SVGMobile12] UA requirements vs Author requirements

From: Ian Hickson <ian@hixie.ch>
Date: Wed, 21 Sep 2005 21:45:40 +0000 (UTC)
To: Jon Ferraiolo <jonf@adobe.com>
Cc: www-svg@w3.org
Message-ID: <Pine.LNX.4.62.0509212130340.423@dhalsim.dreamhost.com>

On Wed, 21 Sep 2005, Jon Ferraiolo wrote:
> This is the SVG Working Group response to your comment found at:
> http://lists.w3.org/Archives/Public/www-svg/2005May/0131.html
> It is a good general suggestion for the specification to express text 
> using "The UA must..." and "The author must..." (although it is probably 
> better to define a requirement on the content, not the author, much SVG 
> content is generated by software). However, we will not take on the huge 
> task of checking every line of the SVG spec. Most of the existing SVG 
> spec is retained from previously approved SVG Recommendations. However, 
> we have already made a series of spec changes in light of your 
> suggestions and will take this good suggestion into account in future 
> editing work.

I look forward to proof-reading these concerns in the next publication of 
this draft, which I assume will be another Last Call draft. (If the next 
publication of this text is not another WD or LC draft, then I do not 
consider this to have satisfied my comments, as it is not at all clear 
what the WG has decided to change and what is still ambigiuous.)

> If there are other specific comments about particular sections of the 
> specification, such as the comment about 'contentScriptType' in this 
> email, we will respond to those specific comments.

Well, the problem is that there are entire chapters that seem to lack 
conformance criteria. For example I opened chapter 11 (at random) and from 
a quick glance, I see absolutely no conformance criteria related to the 
"fill" feature whatsoever. This would mean that a UA could do whatever it 
liked in terms of painting an element and it would not be non-conformant. 

Similarly, the "stroke" section seems radically under-specified; the only 
normative criteria I could find related to specific cases, such as 
rendering "stroking properties which are affected by directionality" (how 
do you render a property?), using particular algorithms "for stroking 
properties such as dash patterns whose computations are dependent on 
progress along the outline of the graphics element", and an empty 
requirement stating that the behaviour for shapes must be equivalent to 
the behaviour for paths (which itself is not defined insofar as I could 

This problem is maybe best described by the sections on 'stroke-linejoin' 
and 'stroke-linecap', which do not contain any normative statements at 
all, and no definitions beyond images that are themselves written in SVG 
using the very properties being defined.

> Relative to your comment about contentScriptType, we have changed the
> definition of contentScriptType to say:
>   This language must be used for all scripts that do not specify the scripting
> language.


> Thank you for your feedback. Please let us know if this does not address 
> your concerns. We will assume that you accept this response if two weeks 
> pass without further postings on this subject.

This does not address my concerns, as explained above.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Wednesday, 21 September 2005 21:45:46 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 March 2017 09:47:04 UTC