Caching of discovered Yadis service definitions / best practices?

[OHTSUKA Ko-hei]

Hi,

I released perl module for discovering Yadis Resource Descriptor from
Yadis URL on CPAN.

Net::Yadis::Discovery

And in this case, using cache or not is depending on RP's choice.
I just use URI::Fetch module inside of it, and if RP want to use
cache, just give a cache
object to it's constructor.

OHTSUKA Ko-hei

2006/3/4, Johannes Ernst <jernst+lists.danga.com at netmesh.us>:
> The first time a Relying Party (RP) encounters a Yadis-enabled URL,
> the RP will likely want to obtain the Yadis document associated with it.
>
> The second time the RP encounters the same URL, it may or may not
> want to re-resolve. If a year has passed, it probably should re-
> resolve. If a second has passed, re-resolving is almost certainly
> going to produce the same information.
>
> Of course, we agreed that the Yadis 1.0 spec will be silent on this
> issue, and that decision was the right one.
>
> I have a question for implementors: Do you cache? Or not? Or in the
> next version? ;-)
>
> If you cache, for how long do you cache? Or what is your policy? I
> suspect that most people won't cache at this point in time, in
> particular as resolution loads are still low. But if you did, what
> would your policy be and why?
>
>
>
>
> Johannes Ernst
> NetMesh Inc.
>
>
>
>   http://netmesh.info/jernst