W3C home > Mailing lists > Public > public-html@w3.org > May 2007

Re: HTML forms, XForms, Web Forms - which and how much?

From: Maciej Stachowiak <mjs@apple.com>
Date: Tue, 1 May 2007 11:57:17 -0700
Message-Id: <B420732C-720D-4E70-9C1F-E2A88206E9F2@apple.com>
Cc: "Preston L. Bannister" <preston@bannister.us>, mark.birbeck@x-port.net, preston.bannister@gmail.com, public-html@w3.org, public-html-request@w3.org, Simon Pieters <zcorpan@gmail.com>
To: John Boyer <boyerj@ca.ibm.com>

On May 1, 2007, at 8:06 AM, John Boyer wrote:

> It's pretty hard to understand how you (and Daniel) come to this  
> conclusion given that the two working groups are *chartered* to  
> come up with a hybrid approach, i.e. something that is NOT equal to  
> XForms as it currently is written.
> The WF2 crew have many times said "Since we already have a spec,  
> why don't we start with that and you can point out what's wrong  
> with it."  I would say the same thing.  XForms already is a W3C  
> recommendation for forms.  Why don't you start with that and point  
> out what's wrong with it for your use cases?

That's been done many times. To give a quick summary of some of the  
recurring complaints:

* XForms content cannot degrade gracefully in existing HTML4 UAs.
* XForms content has a significantly incompatible syntax, meaning  
that to adopt any of its features for a given form, authors would  
have to rewrite the whole form.
* XForms has content constraints which limit layout flexibility.

Since Web Forms 2 advocates have done XForms advocates the courtesy  
of pointing out what problems they have with XForms many times, I  
request that XForms advocates give the same courtesy.

Are you saying that the Forms Working Group advocates incorporating  
XForms into HTML, rather than XForms Transitional, or something similar?

> I think the biggest complaint is not wanting to use well-formed  
> XML.  Well, the new forms technology is supposed to cover that.   
> The XML serialization is called XForms transitional, but there  
> should be a tag soup variety.  It's the tags and their meaning that  
> we want to be consistent regardless of minor lexical variations.

XForms Transitional is not an XML serialization of anything, and it  
does not have the same tags and attributes with the same meanings as  
XForms. It is also not a W3C recommendation for anything. Indeed, it  
is not even a W3C Working Draft.

At this point, I am confused. Do you advocate including XForms in  
HTML, or including XForms Transitional? The two are, as far as I can  
tell, completely different other than the name.

> I think the second biggest complaint is wanting to reason "on the  
> glass", i.e. wanting to let the UI controls suggest the data via  
> their 'name' attribute.  This seems to be part of the plan too.  My  
> main concern here is that it would be nice if UI hierarchy then  
> implied hierarchy in the data, at least as an option, so that XML  
> can come out of the end of the form fill process if XML data is  
> desired. So, this is where we talk about including attributes like  
> readonly, relevant, datatype and calculate directly on the  
> controls.  It's also where we talk about what the repeat structure  
> should look like.

Web Forms 2 includes a repeating structure, the ability for forms to  
output XML, readonly, and the equivalent of relevant. Data type is  
implicit in the type of control. Calculation is handled by responding  
to events. Have you read the Web Forms 2 spec?

> Ah, now that I read ahead on the list, I see that Dave Raggett has  
> asked essentially the same thing of everyone: let's talk more about  
> design patterns and use cases [and less about sound bites].

When I posted a list of use cases, you dismissed them as "trivial".  
Dave Raggett has also not addressed these uses cases (the ones  
actually common on the web) and said we should talk about design  
patterns instead.

Received on Tuesday, 1 May 2007 18:57:36 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:20 UTC