[whatwg] Web Forms 2.0 - what does it extend , definition of same,

James Graham jg307 at cam.ac.uk
Mon Jan 10 09:07:32 PST 2005


Jim Ley wrote:

>On Mon, 10 Jan 2005 14:22:46 +0000, James Graham <jg307 at cam.ac.uk> wrote:
>  
>
>>They're certianly web applications in my definition - they provide an
>>interface which allows me to retrive, view and manipulate data.
>>    
>>
>
>What manipulation can you do on IMDB
>
I'm not that familiar with IMDB

>, and if we ignore the purchasing
>part of amazon, what manipulation do you have there?  To me these are
>simply websites, if they're not, then just about everything is a
>web-application.
>  
>
Isn't that a bit like saying "If you ignore the email part of GMail..."? 
In any case, without purchasing, one can write reviews, rate products, 
etc. So there is certianly an ability to manipulate data. A web forum is 
also a type of web application with similar functionality (indeed, these 
have superceeded news readers in many environments).

>  
>
>>I feel I must have missed your point here. Why does it matter if a
>>(proprietry) web application (under your definition) consumes semantic
>>markup, non semantic markup, binary data or anything else?
>>    
>>
>
>Because consuming semantic mark-up means that user agents can
>understand the semantics of the mark-up
>
I understood that it was the web application itself that ws consuming 
the markup, not the UA. Your original example was a web-application 
based RSS reader. In this case, the UA has no need to ever come into 
contact with the RSS, just with the HTML (or XForm or XUL or whatever). 
Similarly with an internet based mail app; the user (or UA) has no 
interest in the format of the email, just with the data avaliable at the 
app-end.

>>What matters is the markup sent to the client is
>>semantic or not, so that a user can interpret it without requiring a
>>visual rendering.
>>    
>>
>
>Except of course you're ignoring yet again that web document level
>semantics are near useless, you can guess that the <h1> is the subject
>of the email, but how do we guess which is the from address, hmm I
>guess that's the bit that matches a valid email address production, oh
>no that might be the too, or maybe the cc, hmm web-document semantics
>don't do a great job of really delivering semantics of applications.
>  
>
As a user I know the from address because it has the string 'from:' 
before it (or some other such thing). Now that's not enough that I can 
trivially scrape GMail and export the information to a standard mail 
client (the use case you seem to be proposing) but, in most situations 
that's not what I want to do; I just want to read my email. Moreover, in 
many situations that's not what the service provider wants me to do 
either. So if your application-level semantics were ever avaliable they 
would most likely go unused by most developers. The problem for language 
design is that there's no way to provide those semantics for every 
possible format of information that one might want to transmit. RSS/Atom 
can deal with simple weblogs but if I syndicate lists of recently 
published books (say) they have no relevant semantics. So what do I do? 
There's not going to be an existing langauge that meets my requirements 
so I have to invent one. At the point that I've invented a language, 
semantics are irrelevant because no general purpose client application 
is going to know them anyway.

>>Web Forms and more generally "HTML 5" 
>>    
>>
>
>HTML 5, where can I see the drafts of this specification?
>  
>
You have been around the list enough to know that "HTML 5" is the 
half-joking name that has been bandied about by list members for Web 
Forms 2 + Web Apps 1.0 which are avaliable in the usual place. Is there 
a good reason for the feigned ignorance?


More information about the whatwg-whatwg.org mailing list