rdbms update

Things have been quiet for some time, but lots of things have actually happened.

I found an excellent beta tester in Rudolph van Graan of Patternmatched Technologies.

Rudolph has patiently waded through rdbms bugs, written test suites and excercised the code in quite demanding settings. From what I gather, things are now working quite well, and especially the ordered_set indexes of rdbms can prove to be a real performance booster.

We have yet to consolidate patches, but if anyone out there is interested in giving rdbms a spin, please know that you can get your hands on a much better version than the published one, if you only make yourself known.

Personally, I’ve been side-tracked in several ways. I picked up and moved to another city (Örebro), and started a new exciting programming project together with Joe Armstrong – the tuple store (we should announce a naming contest to find a better name…) We will continue to work on that in secret a while longer, until we have something that holds together well enough to show the rest of you.

Rdbms beta-testing

I thought I’d use this blog to post temporary versions of the new ‘rdbms’.

The documentation for ‘rdbms’ is incomplete, but what there is, you can read here:

The old documentation, while now slightly outdated, actually describes some things
better than the new docs (e.g. the semantics of referential integrity):

For those who have trouble using sourceforge(*), you can download a copy of rdbms from here:

The code has been “de-jungerlified”, in that the Makefiles have been replaced with standalone counterparts; ‘make all’ should work… at least on unix.

(*) Normally, you should fetch ‘rdbms’ from http://jungerl.sourceforge.net