Re: Generated content via URI

On Monday 16 April 2007 20:38, Spartanicus wrote:
> Bert Bos <bert@w3.org> wrote:
> >The CSS WG agreed with my reply[2] to Spartanicus's comment[1] about
> >ignoring vs flagging an image that fails to download. Thus, the CSS
> >level 2 spec is unchanged and still allows the UA to decide whether
> > to ignore or flag the missing image (but we expect an enhancement
> > to the 'content' property in CSS3 that lets the author decide).
> >
> >[1] http://lists.w3.org/Archives/Public/www-style/2006Dec/0086.html
> >[2] http://lists.w3.org/Archives/Public/www-style/2007Feb/0105.html
> >
> >Spartanicus, should we note in the forthcoming "Disposition of
> > comments" that you disagree with this resolution, based on [3]? Or
> > can you in fact accept it?
> >
> >[3] http://lists.w3.org/Archives/Public/www-style/2007Feb/0108.html
>
> I initially inquired about the rationale behind this being changed in
>
> one of the 2.1 drafts. The response provided a reason:
> >there were people who believed there exist document formats that
> > rely on images being displayed (no ALT text, e.g.) and in that case
> > it would be safer to warn the user of a failure."
>
> I'd like to see descriptions or preferably examples and usage stats
> of such "exist[ing] document formats that rely on [CSS generated
> content] images being displayed" before I can form an opinion on
> whether this is an actual problem, and if so if it outweighs the
> problem which I described.

I'll ask what formats people were thinking of.

I think people referred to document formats that were originally made 
for printing, and thus do not provide fallbacks for network errors or 
for different devices. Such documents may require a certain element 
(e.g., <danger/>) to print a symbol (e.g., a warning triangle), which 
can be emulated with CSS through 'content', except that CSS doesn't 
guarantee that the symbol will actually be rendered. Thus, while 
waiting for the fallback control of CSS3, it falls to the UA to handle 
error cases.

>
> Based on another part of your response I raised an additional concern
> regarding the proposed enhancement of the "content" property in CSS3,
> how it potentially encourages authors to abuse CSS to provide a
> function that IMO should be provided by the markup instead. There has
> not yet been a response to this concern.
>
> Specifically I'd like to learn why the WG feels that it is
> appropriate to code content images (as opposed to decorative images)
> via the CSS generated content mechanism. (I'm assuming that the WG
> thinks this is appropriate since it is the only justification I can
> come up with to explain this new feature).

It's not that the WG thinks it is *appropriate* to put different 
semantics in the style than in the document, but simply that the WG 
thinks there is no way to make it impossible, without also making 
correct usage impossible.

Even with only CSS1, you can omit parts of a message and rearrange other 
parts to completely change the meaning of a document. But we can't 
forbid 'display: none' and negative margins, because they have 
perfectly legal uses as well. With the added power in levels 2 and 3 
comes more possibilities for abuse, but that is inevitable. We need the 
extra power to make documents easier to read in more circumstances 
(different screen sizes, paper vs screen, etc.)

All we can try is to make it as easy as possible to do the right thing. 
That will help good-willing people to avoid mistakes. We can't stop 
people of ill-will from abusing the technology. That's a social 
problem, not a technological one.

The 'content' property allows designers to make the structure of a 
document explicit and to adapt the contents to different circumstances. 
E.g., with 'h2::before' you can make headings easier to find and with 
the generalization of 'content' in CSS3 you can provide images that 
replace text without affecting the accessibility of the document and so 
that you can show different images for different window sizes. That 
normally improves the usability of the document, but, of course, there 
is no guarantee.



Bert
-- 
  Bert Bos                                ( W 3 C ) http://www.w3.org/
  http://www.w3.org/people/bos                               W3C/ERCIM
  bert@w3.org                             2004 Rt des Lucioles / BP 93
  +33 (0)4 92 38 76 92            06902 Sophia Antipolis Cedex, France

Received on Tuesday, 17 April 2007 14:06:33 UTC