Re: A List Apart: Articles: Prefix or Posthack

Eric A. Meyer wrote:
> At 7:28 PM +0200 7/9/10, Anton Prowse wrote:
>> Your point seems to be
>> that once an author has embarked on exploiting a particular behaviour
>> (eg achieving a box shadow through pure CSS) then they have the right to
>>  expect that that behaviour be always available.
> 
>    I think it's a natural assumption, personally.  And it's one I've 
> seen throughout other people's writing and conversations about using 
> prefixes.
> 

Perhaps that's the problem.  After all, if browser X is released after
the unprefixed property is "released" (ie the property reaches CR or
whatever) then that browser won't get box shadows either, so the author
would still have to go back and revise their stylesheet if they intended
to support that browser.  (This scenario is perhaps more important than
it used to be, what with mobile devices and entertainment devices being
shipped with a wider range of browsers than we're historically
accustomed to supporting.)  Personally, I don't see much difference
between doing revisions to perpetuate support for future Firefoxes and
doing revisions to introduce support for new browsers, although I do
accept that many authors still have the "I support IE, Firefox and one
other browser" mentality.(*)

(*) whereas the correct mentality is "I support IE, Firefox and all
other standards-compliant browsers (not necessarily testing in them)

>> Things I do on websites for prominent brand clients aren't the same
>> as things I do for granny's knitting pattern website.  I might use
>> -moz-box-shadow on the latter, but not the former.  How I tailor my
>> work for a given situation is part of what I get paid for.  So the
>> high value client doesn't get -moz-box-shadow, but that doesn't mean
>> I don't experiment with that property. Web authors tend to be highly
>> experimental and so I don't think we should automatically assume that
>> prefixed properties will be ignored.
>
>    Not 100% ignored, but much less used.  I think it's only common sense 
> to expect that.
> 

Well, that depends on whether authors are playing the game properly. If
they are, then the proportion of authors experimenting with these
properties might not drop so drastically.  Instead, the proportion of
sites on which they /use/ the prefixed properties might drop.  I think
that is a much less serious problem as far as author feedback is
concerned.  In my experience, any given author tends to be quite uniform
in how they build sites, and if they only had 5 sites using the
unprefixed property instead of 15, I'm not sure we're losing much
empirical data.  It's the difference /between/ how different authors
build sites that will give us the rich feedback.

Cheers,
Anton Prowse
http://dev.moonhenge.net

Received on Saturday, 10 July 2010 11:06:34 UTC