W3C home > Mailing lists > Public > public-respimg@w3.org > September 2014

Re: What are the chances of syntax changes?

From: Simon Pieters <simonp@opera.com>
Date: Fri, 12 Sep 2014 12:20:39 +0200
To: "public-respimg@w3.org" <public-respimg@w3.org>, "Jason Grigsby" <jason@cloudfour.com>
Message-ID: <op.xl2kcpxeidj3kv@simons-mbp>
On Fri, 12 Sep 2014 00:14:06 +0200, Jason Grigsby <jason@cloudfour.com>  
wrote:

> Is it common that features like this get tweaked a little
> in the future as people start using them and an oversight is found?

Yes, it is. But only if it doesn't Break the Web.

As a case study, srcset with only 'x' is already shipping in Blink and  
WebKit. Yet we have tweaked srcset parsing several times after if was  
shipped. But the tweaks did not change the behavior of any existing pages  
that used valid srcset. It basically only affected test cases that went  
out of their way to test invalid cases. The changes were mostly to make it  
easier to make backwards-compatible changes in the future.

It is also normal that the first shipping implementations are not  
perfectly compliant with the spec. For instance they might have  
implemented a slightly out of date algorithm and missed that something was  
changed, or simply have bugs. Then it is fixed in a future version and  
that might break your code if you only tested in one implementation.

This is no different from any other feature that is shipped on the Web. To  
avoid issues, test in multiple implementations and validate.

> Should
> we still be hedging our bets a little?

No, that's not necessary.

-- 
Simon Pieters
Opera Software
Received on Friday, 12 September 2014 10:21:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:06:12 UTC