GUIDs/uniqueidentifiers why not?

[Shaun Clark]

Hello,

I am trying to bring my entire organization from a Microsoft
platform to a Linux one, starting with all of the servers we are
currently running. Sun ONE ASP allows me to run ASP on Linux through
Apache, Mono is working on .NET, but the one thing I can't migrate is
SQL and it's because products like mySQL and Postgres have no built-in
GUID support. We have 100 databases with GUIDs as primary keys across
all of them. It works great in a Microsoft world, but as soon as you
want to go to Linux it seems like their is no way. I have seen a lot
of "hack" fixes, but it still escapes me why mySQL won't support this
kind of seemlingly great functionality.

In my case I have a universal history table that allows the
storage of all sorts of events in one table, it's great, everytime I
need to add a new one I add a new history type to a type table and
then just start throwing data in the table, later on I can look it
back up just fine. It works for everything I do. This really helps to
reduce my need to duplicate the same concept all over the place and
speeds programming since I don't have to worry about which table
contains what history. Now, the reason this uses GUIDs is I can key
into the table with a GUID for the history type. Then I can query the
"reference GUID", which is a generic column that only the specific
software that put it there understands, without ever having to worry
about software that doesn't know what something is trying to do
something with it. I would much rather the software error than try to
take an action that is not warranted by the value.

Any help understanding or with a solution to this problem would be
helpful, thanks.

Shaun