W3C home > Mailing lists > Public > www-style@w3.org > March 2011

Re: [css3-text] Splitting up text-decoration into orthogonal properties

From: Aryeh Gregor <Simetrical+w3c@gmail.com>
Date: Sun, 27 Mar 2011 13:29:44 -0400
Message-ID: <AANLkTikW1CJ+wJV3f1MqEqN3eGxj2696Sx6VrTSiZheF@mail.gmail.com>
To: "Kang-Hao (Kenny) Lu" <kennyluck@w3.org>
Cc: WWW Style <www-style@w3.org>
On Fri, Mar 25, 2011 at 6:03 PM, Kang-Hao (Kenny) Lu <kennyluck@w3.org> wrote:
> These properties used to be in CSS3 Text[1]. I think the reason why they
> were removed is as you said, lack of use cases.
> (This is similar to an early request to have longhand property for
> 'transform', by the way. ISSUE-177[2])

Hmm, I see.

> Fake CSS properties seem to be rather confusing. Is reusing the
> add(token) and remove(token) bits of DOMTokenList[3] a feasible approach?

Currently my algorithm
has lots of places where it says something to the effect of "set CSS
property X to value Y", where X and Y differ based on the style we're
applying.  Clearly I don't want to say "set text-decoration to
'underline'", because that will unset other properties too.  So if the
CSS properties don't let me set things orthogonally, I have to rewrite
it in terms of made-up CSS properties and define my own algorithm for
setting and getting them, which for my made-up underline property will
involve processing the text-decoration property.  Or I could rephrase
things to not be in terms of CSS properties at all.

But this is about as low-priority use-case as you can get, since spec
writers' convenience is least important in the scheme of things (other
than theoretical purity).  So if there's no author demand, I guess
there's no justification for the feature.
Received on Sunday, 27 March 2011 17:30:38 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:07:57 UTC