Kim Rostgaard Christensen and I had spent most of the day testing LDAP servers in a replicated setup and the LDAP bindings provided with AWS. It had been a good day, with us practically walking around knee deep in success. And it just wouldn’t stop. At 1730 the first Ada-DK guys joined us at the Responsum offices for yet another open Ada-DK meeting. Attendance was not at an all time high yesterday, but still we had a pretty solid showing:
We talked at some length about the wonder that is OpenID, and naturally Kim and I showed of our LDAP stuff to the crowd. The general consensus was that OpenID rocks and that the Ada code needed to interact with LDAP is functional but bordering on the odd. Kim would really like to implement the LDAP protocol in pure Ada, ie. no binding to libldap, but seeing as that would probably take quite a few days to do, we decided to just stick with the LDAP binding found in AWS, and then simply abstract our way out of the weird stuff.
After having talked about OpenID and Ada/LDAP, we moved on to information hiding in Java – or the lack thereof. Jacob showed us some examples on how easy it was to break information hiding in Java, without the programmer actually knowing that that was what he’d done. It was pretty scary actually. The same thing in Ada was of course also possible, but here the code explicitly told the programmer that things were no longer safe. So both languages could in fact produce the same unsafe code, with the big difference being that the Ada code clearly told the programmer that something was amiss, whereas the Java code still appeared safe, despite the fact that it was completely broken. I must admit a great deal of it was a bit above my level, but it was still very interesting.
We also “wasted” a good 30 minutes checking out the cloud gaming service OnLive. I played the latest Duke Nukem game on my little MacBook Air, streaming it from America. It worked flawlessly, albeit with a little bit of lag. It is an amazing piece of technology the OnLive people have come up with, and we all hoped for a native Linux client as soon as possible. And some servers in Denmark. With a native Linux client, I really do believe a service like OnLive can help push more people towards Linux. So go go go OnLive – get that Linux client out to the gaming hungry *nix geeks! :o)
I started a discussion about how to do connection pools with the GNATColl Database_Connection objects. I’d tried to build a solid system myself, but all my efforts had failed. I had hoped for some good ideas from the Ada-DK guys, but it appeared I had already emptied out most options. In the end it was decided that it wouldn’t be possible to just create a simple pool. Instead a full abstraction would be necessary, to avoid copying and/or reusing Database_Connection objects by mistake. I admit I was a bit disheartened by that, but it was at least nice to know that I hadn’t missed some straightforward solution to the problem.
During dinner we talked about the latest Microsoft scheme to own the world: Secure boot. Interestingly enough, opinions on what will happen if this scheme is brought to fruition varied a great deal, and so did the suggestions for what to do if Microsoft pulls it off. We could though agree on one thing: The whole concept stinks!
The evening ended with Kim and I looking for a decent CalDav server, and perhaps even an Ada based CalDav client. We found no Ada CalDav stuff, and honestly the CalDav servers all seemed a bit “meh”. Ulrik and Per talked at some length about formal methods (at least that’s how I remember it), and at about 2200 we started wrapping up.
Another great meeting, with a great bunch of guys!