Re: Late binding

Mark Baker wrote,
> Hey.  Redirecting to www-archive because I said I'd shut up. 8-)

Thanks for redirecting to a list I'm not subscribed to ;-)

> > You can ask for "something" and get it, yes. But what can you do
> > with it once you've got it?
>
> Right.  I don't claim that either approach does anything special
> here.
>
> But I will point out that HTTP permits a client to ask for a specific
> format, and that there are several orders of magnitude less data
> formats than there are possible methods.  Plus, data formats are
> standardized, whereas methods rarely are.

So you get back application/xml ... so you can parse the result. Is that 
useful without knowledge of the semantics associated with the document?

> > As far as I can see, all you've done is moved the uncertainty
> > (alternatively, the requirement for some kind of agreement or
> > shared understanding) from the invocation of an operation to the
> > interpretation of its result.  Can you persuade us that anything
> > very much has been gained?
>
> Intermediaries such as firewalls also understand what "1"/GET means.

Sure, but can they tell the difference between,

  GET /benign-uri

and,

  GET /malicious-uri

any more than they can tell the difference between benign and malicious 
request entities associated with a POST? Granted the parsing costs will 
favour the URI over the entity, but it's not clear to me that that's 
likely to be all that big a gain.

> > To me it looks like you're squeezing the semantic balloon in in one 
> > place only to have it bulge out in another.
>
> It's similar, but this game of balloon squeezing isn't zero-sum. 8-)

That might be true, but you haven't persuaded me yet. And who's to say 
that non-zero-sum doesn't pan out as negative-sum?

Cheers,


Miles

Received on Sunday, 30 June 2002 08:40:05 UTC