Re: Draft EOS minutes

>>>>> On Fri, 16 Aug 2002 15:54:19 +0200, "Marek Malowidzki" <[email protected]> said:

Marek> The explanations focus on processing on a manager side. In
Marek> fact, I do not seem any problem here. So, could I ask about the
Marek> real reasons for filling the holes?  If the reasons are
Marek> well-known, I think that new documents should explain them
Marek> shortly, and nobody would ask about them anymore.

The document is hardly finished.  I should have an at least slight
better copy published by Sept 15th.

You don't see a problem with holes an the manager side: good.  Most
people don't see a huge problem and if given the task they would
hopefully write management software that does the right thing.  Now,
can you truly tell me that you've never seen any management software
that *didn't* handle holes properly?  I've seen a lot that doesn't.
I've written some myself that doesn't (with the thought always in the
back of my mind that I'll do that "next").  It's not so much that it's
not algorithmically doable, but rather that it still takes sorting and
processing.  Wouldn't it be better to put the sorting straight into
the protocol rather than have the agent remove the ordering only to
have the manager need to rebuild it on the other side?

-- 
"The trouble with having an open mind, of course, is that people will
 insist on coming along and trying to put things in it."   -- Terry Pratchett