- From: Matthew Thomas <mpt@myrealbox.com>
- Date: Sat, 28 Aug 2004 15:59:40 +1200
On 27 Aug, 2004, at 3:51 AM, Ian Hickson wrote: > > On Fri, 27 Aug 2004, Matthew Thomas wrote: > ... >> Narrowing a specification to *forbid* the hitherto-correct behavior >> followed by the 95%-dominant UA may achieve a variety of good and >> useful things, but interoperability is manifestly not one of them. I >> would greatly appreciate receiving a genuine answer. > > We could define it the other way (the IE way) if you prefer, Given the end-user benefits of the IE/NS4 way (the inverse of the three listed in http://listserver.dreamhost.com/pipermail/whatwg-whatwg.org/2004- August/001767.html), I would indeed prefer that. Though since Mozilla, Opera, and Safari don't do it that way, I'd be content with leaving it undefined. > but it seems to me that having the "default button" be successful That's a tautology. It's only the "default button" *if* UAs make it successful even without clicking. > is technically better. Well, I haven't seen any technical reasons so far, except "Fewer choices makes things easier for implementors". I know I'm biased towards end users at the possible expense of other parties, but I really can't see how any benefit from fewer choices for implementors would be greater than the usability hit from longer URIs. > It does improve interoperability, in that new browsers are more likely > to do the spec thing than just pick a random behaviour. > ... What-WG's compatibility policy (provide HTCs and the like for Internet Explorer for Windows, but don't worry quite so much about other browsers) assumes that people using Internet Explorer for Windows will switch more slowly than people using other browsers will upgrade. So even if interoperability is your only goal for this particular detail, it would make more sense to be interoperable with the Internet Explorer behavior than with the other-browser behavior. -- Matthew Thomas http://mpt.net.nz/
Received on Friday, 27 August 2004 20:59:40 UTC