views:

61

answers:

3

Specifically in relation to this article on 'A List Apart':

http://www.alistapart.com/articles/prefix-or-posthack/

I was hesitant initially, as they did seem a bit hack-ish, but I've started working with the prefixes here and there for progressive enhancement and I've got to say that for me the technique has actually gotten me more excited about web development again.

I'd be interested in hearing strong arguments for and against - I think Eric Meyer presents some really good arguments for, but are there others on either side not presented in the article?

+4  A: 

Prefixes are necessary, for exactly the reasons Eric states. I'm surprised there is any controversy over this at all. Even today many of the prefixed styles have browser syntax differences that have to be resolved before they can become unprefixed standard properties.

But you have to write the same rule four times? Well boo hoo. That's the price of using a feature before it is standardised. If you don't like it, just wait and you won't have to do it in future.

If we got rid of the prefixes, we'd end up with all the CSS compatibility nightmares and hacks of the past. That's not worth it just to make your life as a web author marginally less copy-and-pastey.

The problem with the prefixes is not the prefixes. The problem is that the CSS specs are developed too slowly. box-sizing, for example, has been a done deal for many years, but since the CSS3 Basic User Interface Module has stalled at Candidate Recommendation stage we are still unnecessarily typing -moz-box-sizing.

That's a failure of process on W3's part, not a technical issue to do with prefixes. CSS3-UI needs to be finished and a full Rec. If CSS3-UI cannot go forward due to technical reasons in another part of it, those features either need to be killed and scheduled for CSS4, or the module needs to be split/reorganised so that the stable bits can go through.

Arguably the whole Module system is not granular enough and needs to be split into one-standardisable-document-per-feature. There is a lot of stuff in CSS3-UI which has no impact on box-sizing or outline features. We shouldn't be waiting on that stuff to standardise the uncontroversial features.

bobince
Does the fact that what seems like a fairly high percentage of users can't/won't update their browsers to something that can actually take advantage of the new specs influence the process?
DHuntrods
No, the [Process](http://www.w3.org/Consortium/Process/) doesn't care about old browsers. It is only interested in there being working implementations of a standard before advancing it to PR/Rec (see section 7.4.3).
bobince
That makes sense. Thanks.
DHuntrods
+1  A: 

http://stackoverflow.com/questions/1962697/what-are-pros-and-cons-to-use-vendor-specific-extesions-which-are-not-included

http://stackoverflow.com/questions/2437502/why-do-browser-vendors-make-their-own-css-properties

metal-gear-solid
Thanks - I did search for this question as I figured it must have come up before, but didn't come across the first.
DHuntrods
+2  A: 

There's been an interesting discussion on www-style following the publication of the article. One of the sub-threads has been about whether and when browser vendors should drop support for the prefixed versions, especially when the the standard changes significantly after the prefixed version was introduced. For instance, -webkit-gradient vs linear-gradient - should webkit implement -webkit-linear-gradient but continue to support -webkit-gradient? How long should they continue to support the old property for? It's a good discussion to read through as there's contributions from developers working on the major browsers.

robertc
Thanks, I'll go check that out.
DHuntrods