Per-Entity Metadata Working Group

["Cantor, Scott" <>]

> AFAIK, no one has voiced such a requirement. In any case, I'm not sure
> how valuable such a requirement would be. It's unrealistic to think a
> server will never fail. The question is: what happens when it *does*
> fail.

There are pretty much millions of lines of code written that basically assume 
a DNS server will never fail. And when they do, it all just breaks.

It's not about whether something can fail, it's about the expense of working 
around it and the economic trade-off of asking everybody to code a 
workaround. The question is to what degree a standard "serve static 
documents" web use case has reached this level of commoditization and 
"assumed robustness".

> Btw, InCommon metadata has never had an explicit cacheDuration value.
> Should the MDQ server include one? A realistic value would be 24
> hours. Not sure if this is useful or not.

It's the same as DNS (I sound like a broken record, don't I?). The TTL value 
is used not only to tell a system how often something might change, but more 
importantly to control how often you can count on a client asking again if 
something needs to change.

> I claim they are definitely different (but at least Scott thinks otherwise).

I'm not claiming they are the same (that's an objective matter, and if we 
have constant minutes-long downtime, then no, they're not the same), but I am 
arguing they should be.

-- Scott