- From: Andrew Fedoniouk <news@terrainformatica.com>
- Date: Wed, 23 Nov 2005 22:46:22 -0800
- To: "L. David Baron" <dbaron@dbaron.org>
- Cc: <www-style@w3.org>
----- Original Message ----- From: "L. David Baron" <dbaron@dbaron.org> >On Wednesday 2005-11-23 12:22 -0800, Andrew Fedoniouk wrote: >> Can we remove mentioning of any UA specific pre- and post- >> style processing procedures from the spec? >> >> In particular items: >> >> > 1. UA stylesheet >> > 2. non-HTML presentational attributes >> >> and >> >> > 11. UA stylesheet !important, sorted by specificity and then order >> >> ? >> >> as it is really up to UA how to set initial values of styles and >> how to implement fixup procedures. >> It is enough to say that UA is in its rights to implement these >> functions in its own way and this process is *completely unrelated* >> to style cascading, specificity and order - scope of the spec per se. >> >> Why we need these UA implementation details there? >We need it because it limits what a conformant UA can do. If we didn't >have that limitation then a UA could do anything and claim that it was >still conformant, but that the "bug" was due to the behavior of its UA >stylesheet. >-David UA can do anything for the sake of user's satisfaction. It even can override user settings by intrinsic !importants despite of the fact that spec does not allow[1] it to do so. I suspect that there is strong reason for doing this but again formal conformance tests are something different from usability features. Conforming UA shall have "pure" mode when it is possible to disable such post-processing. As CSS spec does not define "standard HTML style sheet" as it should not enforce the way how it needs to be applied, IMHO. [1] http://www.w3.org/TR/CSS21/cascade.html#important-rules Andrew Fedoniouk. http://terrainformatica.com
Received on Thursday, 24 November 2005 06:46:54 UTC