[whatwg] Web Forms 2.0 Implementation choices

[Jep Castelein]

thanks for the answers: we'll try to specify more details on spec changes
that would make it easier for us to comply with WF2. I can't promise
anything though, due to several demanding customer projects.

Indeed, WF2 full compliance is not possible with pure JS, but it might be
nice to get started until native browser implementations arrive. And - of
course - to support the browsers that don't have native WF2 implementations,
such as the current generation of browsers (including IE).

With some creativity, there is more possible with pure JS than many people
think: however, we have chosen not to extend the DOM (not feasible of
course), but to develop our own tag-based language (BXML) with its own
execution model, its own events (that bubble), state mechanism, etc. These
tags have their own namespace, and can be freely mixed with existing HTML
pages. There seems to be quite a nice match with WF2.

Jep Castelein
jep at backbase.com


> -----Original Message-----
> From: whatwg-whatwg.org-bounces at lists.whatwg.org
> [mailto:whatwg-whatwg.org-bounces at lists.whatwg.org]On Behalf Of Olav
> Junker Kjær
> Sent: 10 January 2005 16:16
> To: Ian Hickson
> Cc: whatwg at whatwg.org; Jep Castelein
> Subject: Re: [whatwg] Web Forms 2.0 Implementation choices
>
>
> Ian Hickson wrote:
>
> > Web Forms 2.0 is not written with a pure JS implementation in
> mind, but if
> > there are small changes that could be made to the spec today to
> make doing
> > such an implementation easier, let me know. The primary target
> for WF2 and
> > WA1 is native implementations and XBL- and HTC- based implementations.
>
> Custom events cannot (as far as I know) be generated through pure
> javascript in IE. Also, property setter and getter methods cannot be
> implemented in IE, without HTC's. So I think it would require
> significant changes to the spec to allow a pure JS implementation.
>
> However, some features in the spec is impossible to implement even using
> HTC's:
> - bubbeling events. HTC's support custom events, but they cannot bubble.
> - [[DefaultValue]] method on ValidityState cannot be implemented in IE
> - implementation.hasFeature("WebForms", "2.0") cannot be supported
> - new css pseudo-classes cannot be supported fully
> - lots of limitations around the input type=file and the different
> encoding types.
>
> Of course, since an implementation has to support DOM3 Core+Events to
> claim conformance to WF2, a fully conformant HTC-based implementation
> for IE is probably not realistic anyway :-)
>
>
> Olav Junker Kjær
>