Re: level of effort: (Was: Re: Draft minutes of the 13/Apr/2004 XKMS teleconference)

Stephen Farrell wrote:

> Yes. In the past people (mainly commercial companies) have put servers
> up and allowed clients to test against those. We'll have to do the same,
> though since we won't all be able to afford to leave a server on the
> net all the time, some testing may need to be co-ordinated (e.g. we
> arrange that you leave your server up for an hour or so at some given
> time and I try run my client against it). We did CMP (rfc2510) testing
> that way some time ago and it worked ok (though it was bad for those in
> unlucky timezones, and we did have some f/w problems to get around -
> easier over port 80 though I guess;-).
> 
> Can those intending to implement servers let us all know what they can
> do in this respect?

The intention is there to build a server.  (Mind you - standard caveat 
around intention vs. reality in open source.  Everything we do is 
volunteer contributions.)

Whether we can get something permanently on the net - I'm not sure.  So 
we may end up with the same kind of thing around co-ordination.

> We'll gratefully take any and all application code that acts as an xkms
> client! (There was talk of a JCP to do an api some time ago, not sure
> if that got anywhere?).

Well, if we can put something together, it'll be under the standard ASF 
license.  Useable by all :>.

But it would be nice to demonstrate (for example) an "ability" to fail a 
signing check on the basis of a failed response from a ValidateRequest. 
  (For example because the key is for exchange, not for signature.)

Cheers,
	Berin

Received on Thursday, 15 April 2004 07:36:59 UTC