Re: Why MD5 Headers are Imperative

Eric J. Bowman wrote:
> > URI aliases do solve the problem in my own implementation and I'm quite
> > happy with them.
>
> Do they?  If I have this URI:
>
> http://example.com/2006/aug/9.01.atom
>
> But it's generated by an .xql file, then yes I could put it at some alias, 
> but how would I then manipulate the XQUERY file at the URI using WebDAV? How 
> does my WebDAV client know where to PUT?  It's only an Atom representation if 
> the query executes, bear in mind.

You might well have identified problems in APP, you seem to believe it 
vehemently. But your arguments are not cogent to me.

I'll speculate some. You have an identity and data modeling model issue 
that APP can't help you with. Or HTTP, come to think of it. I note you 
haven't said what the resource denoted by

  http://example.com/2006/aug/9.01.atom

is. You need to decide what that URI denotes, and what representations 
will be issued. If you need to change the state of a resource that is a 
xquery file, you need to give it a URI. If accessing that URI returns a 
query result, and not the source of the query, that is squarely your 
problem to deal with. In short you need to persuade me you are not 
playing Games With URIs*, and swapping out referents based on the 
internal needs of your application.

cheers
Bill

* Tip of the hat to Wittgenstein