[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
More thoughts...
After reading more of the messages from the last couple weeks as well as
Jason's "Future" document...
I know I'll be interested in seeing how PostgreSQL 7 compares in speed
to MySQL. You (Jason) said that if it were 40% slower that would not be
acceptable. But are you sure? Transactions alone are a huge feature,
and could almost certainly be put to good use. The object code updates
several tables at a time -- that REALLY should be done in a
transaction. Do we expect the computer to be maxed out processor wise?
We have big powerful boxes. They should be able to handle it. The
reliability that transactions provide could be worth its weight in
gold. Another big advantage for PG is the license. I still haven't
figured out why so many open source projects have no trouble using MySQL
as their DB. IT'S NOT OPEN SOURCE!!! Not that it has a terrible
license, but it is far from ideal.
The "future" doc. First, I'll say that Jason has done a fine job of
leading this thing, so he doesn't need to come down on himself too hard.
:-) I'm glad to see OODS becoming its own project. I think it will be
quite useful. My only concern is that we're even thinking about making
the LKB more complicated by involving more and more components.
Frankly, we should have been able to set up the LKB in a month or two.
Really. Using regular conventional database layout and straightforward
queries. Of course, it wouldn't have been nearly as flexible, but it
would have accomplished our main goals here. And that's what I had in
mind when I joined this thing.
But I agree that Corba has some benefits. I'm just not sure that it
will improve greatly on what we have. Do we really need all kinds of
language bindings? Do we want another layer of middleware for each
query to go through? And why NOW????
Have any of you guys looked at XUL much? I was hoping to do some
experimenting with that in the next few days. I know there was talk of
a front end C++ or Java app. But it seems like a Mozilla/XUL based
front end would work just as well with a whole lot less work.
Mozilla M15 is incredibly sweet BTW. After relentlessly using it for
about 14 hours, it only crashed once... and not too long after that,
when I was back in Netscape, I had a whole series of weird events
including an X crash, another Netscape crash, and two complete freezes
of the system. So i think something's screwed up with it anyway...
That's about all for now.....