Re: Unprefixing CSS properties

As an author, I don't want to be tormented by slightly incompatible
implementations, but I also don't want to be tormented by an
never-ending need to bloat and maintain otherwise-identical prefixed
property declarations. I don't see why one is unacceptable, but the
other is?

It seems to me that prefixes are simply perpetuating incompatibilities
by providing a method of per-user-agent targeting that obviates
vendors of the need to expediently pursue cross-compatibility because
hey, authors have a way to work around differences, and if they
complain you can just tell 'em they shouldn't relying on prefixed
properties anyway. Just wait a few years and you'll be able to use the
unprefixed syntax when it eventually arrives. Convenient for vendors,
certainly, but not for authors, and that doesn't seem to fit the
hierarchy of constituencies as I understand it.

I don't dispute the requirement for *something* in this space, but I
just think that, given all the ability and intellect concentrated in
this group, a better approach than vendor prefixes could be found to
do it. Some kind of mechanism(s) to allow...

 - Opting into unstandardised or experimental features.
 - Targeting individual user-agents.

...that doesn't require the tedium and bloat of prefixes. Whether
that's through some kind of pragmas, strictly optional prefixes, new
syntax or whatever, there has to be a better way.

I do agree though regardless of any better syntax being developed, it
would be preferable if vendors moved more quickly to get standards
finished and interoperable, by concentrating on the boring bits of
finishing up, addressing issues, and writing test-cases for specs that
have been hanging around for years. From reading this mailing list
off-an-off for the past couple of years, I'm not sure how much
appetite there is for this? (Credit where it's due, Microsoft seem to
be the most active in this area).

If that is just never going to happen, and the reality is CSS will
remain a perpetual beta, or a living standard, or whatever the trendy
name is nowadays, then fine. Authors can cope with and work around
incompatibilities, but it would be nice if you gave us more convenient
tools to do so. Right now, you have the insane situation that authors
are writing preprocessors and libraries to work-around the tools
themselves, let alone the incompatibilities! Whatever the theoretical
intentions and merits of prefixing, if people are actively working
around it, it means they don't want it, and a rethink is required.

Thanks,

Jon Rimmer

Received on Wednesday, 16 November 2011 18:28:54 UTC