- From: Sylvain Galineau <sylvaing@microsoft.com>
- Date: Fri, 18 Nov 2011 18:57:42 +0000
- To: Brad Kemper <brad.kemper@gmail.com>
- CC: www-style <www-style@w3.org>
[Brad Kemper:] > > > > On Nov 18, 2011, at 8:22 AM, Sylvain Galineau <sylvaing@microsoft.com> > wrote: > > > Google's Alex Russell just posted on this topic: > http://infrequently.org/2011/11/vendor-prefixes-are-a-rousing-success/ > > He makes some good points. Problems with border-radius and box-shadow > might never have been found and fixed and improved upon in time, if we had > not started trying to use it in production Web sites. Tinkerers don't > spend nearly as much time with these things as authors trying to make them > work for their own needs. Yes, there is huge value in gaining real-world experience from web authors. Not just in finding bugs but understanding which use-cases matter and even, sometimes, discovering use-cases we didn't think of. This kind of iteration may take longer but it results in better standards imo. Achieving the same with nightlies or dev channel builds would be a challenge. While those are very useful to browser vendors they just don't gather the same kind of data or generate as much real experience. I just don't see as many authors experimenting with the new and shiny if they can't use the feature for themselves, never mind share the results beyond the small number of people who download these builds*. So I don't think decreasing or constraining the ability to deploy and iterate would be helpful. But it's possible we could streamline the way the iteration happens. *: yes, I know that those builds can go out to millions. Let's just say that not everyone tests them to the same extent.
Received on Friday, 18 November 2011 18:58:14 UTC