Re: What do we get with PUT/DELETE?

[Sam Ruby <rubys@xxxxxxxxxxxxxxxx>]

Roger B. wrote:

> > http://www.intertwingly.net/wiki/pie/DifferentlyAbledClients
>
> Sam,
>
> Actually, I was thinking of the pre-Atom piece you wrote:
>
> http://www.intertwingly.net/stories/2003/01/26/evolve.html
>
> If Atom were to ditch the reliance on HTTP purity and stick to passing
> around chunks of markup, then I could implement Atom with a few trivial code
> modifications. I could literally change a few variable names and be done,
> and it would arguably be more efficient without the XML-RPC library as a
> middleman.
>
> But if I'm going to end up having to implement a pointless digest
> authentication system (I'll believe that's important when Blogger ceases to
> have an HTML login form), fiddle with HTTP response codes, and generally
> rewrite the whole thing... well, the exercise rapidly loses its charm.

If you look closely, the recommendations in that piece are basically the 
same ones made in DifferentlyAbledClients.

Another thing to look at is:

http://intertwingly.net/blog/003/09/03/ and in particular,

http://intertwingly.net/stories/2003/09/03/mstop10.py

Replace <GetTopDownloads/> with <POST/> and this should interoperate 
with XML::Atom:

http://www.sixapart.com/log/2003/09/announcing_xmla.shtml

It truly *is* just data.

Can you explain to me why such an approach would require a general rewrite?

- Sam Ruby