PG is great and m$sucks

hi, guys,

I almost finish porting a schedular from linux/pg/apache
to nt/m$sql/iis (there is going to be a headach of iis security --
any ideas? ).

I understand that some people may feel uneasy about this, however, note
that we are engineers, not the boss, nor the customers. So, even
in closed-software :-) situation it's other end's problem(in this
case, M$shit's "issue"), but aren't oss totally different from THEM!
also, it is good for marketing to get people into PG by saying that
after all, if you are not satisfied, to get out of PG is easy!

I'll give a report of this experience. -- I can not develop the db
itself now (ya, someday, I will try the oodb or data warehouse/datamining
feature, someday :-), at lease I can do this.

However, now I just want to say: I now apprecaite PG much much more than
before. PG is great! perhaps only second to Oracle (perhaps after
I port it to oracle, I will find PG is the best!). But surely, it
is better than M$sql. you see what, M$sql can not support two sequences
in one table, it's varchar field can not be longer than 255 chars, and,
its text type does not work right if you use common sql, what's more,
also, the programming interface of "select currval(...)"
(select @@identty) does not work. (I have to lock the table's insert
and max(productId) ). ...

thanks for your great work.