Re: Proposed requirements for checkpoint 6.4 (Response Required)

Jim Ley wrote:
> 
> "Ian B. Jacobs" <[email protected]>
> 
>>Charles McCathieNevile wrote:
>>
>>>Works  for me.
>>>
>>>Does this have to be the rendered form, the author-specified form, or
>>>are either OK? (For example, an author may have specified two different
>>>fonts,
>>>and the User Agent may render the first one, or the second if it lacks
>>>the first, or neither, if it doesn't have them. Alternatively they may
>>>have been overridden by an author style sheet, or a high-contrast
>>>setting might be
>>>causing rendering in black and white when the author has specified
>>>several colours).
>>
>>The goal of this checkpoint is to make available what's *actually*
>>rendered.
> 
> 
> I agree with it, but I'm concerned it may be expensive to implement in
> current scenarios, and will also rely on the end renderer providing the
> information 

Hi Jim,

Thanks for taking the time to comment. Can you say more on what you
think would be expensive to implement? Jon and I reviewed existing
APIs and both MSAA and Java APIs (and probably others) cover the
indicated requirements. In what way would the end user have
to provide information? The user agent ultimately determines the
choice of font families, for example, after taking into account
author preferences, user preferences (in the UA and the Operating
Environment), and available font families (and probably other
factors as well).

 > - so if I wanted to make a UA that re-used some components
> which doesn't provide exact pixel locations, I have no way to achieve it
> (but can get close, it will be what I tell the renderer to do.) - is there
> any implementation experience?

Please expand on your scenario.

Thanks!

  _ Ian



-- 
Ian Jacobs ([email protected])   http://www.w3.org/People/Jacobs
Tel:                     +1 718 260-9447