[uf-discuss] Re: [uf-rest] Microformats for <form> ?

[toydi]

On 3/28/06, Mark Baker <distobj at acm.org> wrote:
> On 3/24/06, toydi <teohuiming.work at gmail.com> wrote:
> > If forms are the only service description exposed to users, it seems
> > not containing enough info to let users to pre-configure their user
> > machine to handle potential response documents. Does it mean that we
> > need *something else* to expose more info? or do I miss some key
> > points here?
>
> The Accept header?  That and standardized formats seems sufficient to
> me.  On the other hand, since these formats would be microformats,
> then Accept might as well be hardcoded to "text/html". 8-)

Accept header is issued by users side (request) about what they prefer
to handle, but I wonder how service side would announce what users
should be prepared to handle.

Several description language proposals [1] (e.g WADL, WIDL) are
designed to announce possible response/output details. However,
interestingly, form-based languages seem to ommit that part
(intentionally?) at all.

Why is it better to delay the knowledge of response document details
to execution time (during a response is returned)? This looks like
what a form does on purpose.. ;-)

Taking an example in [2], we can provide a form to describe the
"Arguements" section, but not "Example Response" section. So, should
we ommit that, or need an extra W*DL document to describe the
response?

[1] http://esw.w3.org/topic/WebDescriptionProposals
[2] http://del.icio.us/help/api/posts

Cheers,
--
Teo HuiMing (toydi)
teohuiming.work at gmail dot com