0.6 Lines Of Code Per Hour June 30, 2008
Posted by gordonwatts in GLAST, computers.add a comment
After the GLAST launch the other week I’ve looking at things space and astro when I surf the web (honestly – I don’t do it much! Especially at work! Especially while waiting for an ATLAS build!). I stumbled on this pod-cast on writing software for spacecraft.
It was a fascinating listen. Hans-Joachim Popp is the head of computing (CIO) at Deutsches Zentrum für Luft- und Raumfahrt, a company that makes software for German spacecraft. Some of the things that he said were jaw dropping in the context of the ATLAS software (and DZERO software). For example, averaged over the whole project about 0.6 lines of code are written per hour. They have code reviews, where the developer stands in a room and has to defend every single line of code. Their test code is a x12 longer than the actual code they write. They scan all of their code with a static analysis tool to look for “dumb” bugs.
Apparently the programming environment is similar for devices that run in intensive care units in hospitals. Comforting, I suppose. Modern critical airplane software has dueling versions of the code written (much like the space shuttle). In the case of the A380, apparently, one is written in C and the other in Ada* (other than this podcast this slide show was the only reference I could find).
Of course, the standard modern day project is only 20,000 lines of code. Voyager, which has been up there for 30 years now, has only 4 KB of memory for programming – so only 4000 assembly language instructions. He also mentioned that the ability to update the code in Voyager has been crucial to keeping the missing going this long.
I wonder how much of this formal methodology was followed for the GLAST LAT software? The trigger, for example? I remember Toby talking about how it was impossible to change because of all the paperwork and reviews involved in changing a single line of code.
One thing I’ll miss about returning to Seattle is I’ll loose this commute. Yeah, backwards. I’d rather be without the commute. But one advantage is I get to listen to all these podcasts (normally I listen to news and politics from the USA).
* I was in love with the idea of Ada when it first came out. I never wrote any code in it, but I thought that was the coolest programming language when it came out. Heck – it has multi-threading built in. Was over designed to catch errors early. Now, I’d hate it as being too restrictive.
** Imagine if all of this code was written at 0.6 lines per hour!?
Watch GLAST Blast off June 11, 2008
Posted by gordonwatts in GLAST.add a comment
If you are online right now (about 11:45am Eastern Time), you can watch the GLAST launch realtime video here!
11:59: It sounds like they are working in the D0 control room! They have disabled an alarm! We do that all the time. 
12:08: Already up 35 miles! And fast — 7000 miles/hour.
12:12: The room they track the progress from looks just like one of our movable counting houses in a particle physics experiment. Lots and lots of racks of electronics and displays. I’d feel at home. One difference: the guys I can see are wearing ties! Not tee-shirts!
12:14: 100 miles up there. I’m in an office at CERN watching this on my computer. The offices is crammed full of people — but they are all working on various things ATLAS. I basically had to yell at them to get them to watch the launch they were so intensely involved on the problems they were solving.
12:16: Ha — the guy watching those banks of displays and electronics is getting supplemental information from what looks like a laptop! That is so just like D0!
12:19: Now it is coasting in a low orbit – so it will sit there for an hour. That was cool! I’ve never seen a launch before – so this is the closest I’ve ever come. Sounds like an hour or two from now that will all be done and the solar arrays will open and GLAST infrastructure will start to power up.
1:40pm: Hey — they got it in orbit and the solar pannels out and the thing is under power. That was 90 minutes from launch pad to orbit. It was, oh, about 10 years from design to launch pad. And, it will be another 2 weeks before they start to power up the instrument (the next two weeks will be work on making sure the spacecraft that carry’s GLAST is in good shape). 90 minutes. Wow.
And congrats to my fellow prof at UW, Toby Burnett, who has been working on this from before my arrival at UW. He must be dancing in the streets by now!
P.S. They do have guys there in ties with long hair.
GLAST: Any Day Now!! June 4, 2008
Posted by gordonwatts in GLAST, NASA, physics.add a comment
My friend and colleague, Toby Burnett, is right now down in Florida for a GLAST collaboration meeting. This is a funny place for a GLAST meeting… until you realize the launch date for the satellite is “any day now!”. The last I heard it was supposed to be June 7th (see the official page for news releases). But apparently the members of GLAST have been told that Saturday is now out.
I’ve been urging him to blow off university duties and stay down there if he can. After all he has been working, along with everyone else in the collaboration, towards this goal for the last 10 years!
UPDATE: The official NASA launch schedule – now slated for June 11.
