W3C home > Mailing lists > Public > public-webapi@w3.org > April 2006

Re: ISSUE-72: should XMLHttpRequest be an EventTarget?

From: Robin Berjon <robin.berjon@expway.fr>
Date: Sun, 9 Apr 2006 15:14:11 +0200
Message-Id: <CCB3720C-D2DB-469E-BC46-82F242F54A03@expway.fr>
Cc: Web APIs WG <public-webapi@w3.org>
To: Jonas Sicking <jonas@sicking.cc>

On Apr 08, 2006, at 02:56, Jonas Sicking wrote:
> What is the advantage of doing this now as opposed to doing it in a  
> version 2?
>
> I can see two advantages:
>
> 1. It means we can add a pile of other features like onload, onerror
>    onprogress and so on.
>
> However, this is straying far from the original intent of the spec  
> which was to define what already works.

Indeed, it's straying very far from our intended goal which was to do  
the smallest thing that works, and ship it in a very timely manner. I  
don't see anything compelling in the EventTarget proposal that  
justifies delaying the spec. Yes XHR is ugly, everyone knows that.  
This is not the time to fix that. There'll be the whole of v2 for  
that. And the quicker we get v1 out, the faster we can start on v2.

We've now covered 1/8th of the WG's lifetime, let's not start getting  
creative on the specs that do take up WG time until we've got at  
least a fair share of our deliverables out of the door.

> 2. It solves the problem of "what type is onreadystatechange in the  
> idl"
>
> It does feel like this is a pretty poor argument. We would be going  
> against the original intent of the spec for what feels like  
> editorial reasons. If this is the only reason then I'd like to find  
> another solution.

I'll note once more that this problem was solved a long time ago by  
DOM Events, so it's not even an argument.

-- 
Robin Berjon
    Senior Research Scientist
    Expway, http://expway.com/
Received on Sunday, 9 April 2006 13:14:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:54 GMT