For the CS geeks

[cos]

Ever since grad school, I've been carrying with me a stack of papers by Leslie Lamport, my favorite CS writer of papers. I'd tried to find them online back in the 90s but they were all paper-only publications and I never did... until yesterday, they first time I tried in years. It turns out that Leslie Lamport at some point went and scanned in all of his papers, and posted PDFs here:
http://research.microsoft.com/users/lamport/pubs/pubs.html

If you haven't heard of Leslie Lamport, or think of him only as the guy who wrote LaTeX: His entertaining, fun to read papers laid down much of the fundamental theory behind distributed computing. For example, Time, Clocks and the Ordering of Events in a Distributed System. He's known for unusual ways of presenting the math, most famously in:

• The Part-Time Parliament, in which he described how to do distributed consensus (for example, as in three-phase commit in databases) cast in the form of archaeological findings about parliamentary procedure on an ancient Greek island.


• The Byzantine Generals Problem, in which he re-cast the problem of achieving concensus when some processes/processors may go "rogue" rather than simply stop working, as a story about treason in war.

(Actually, some of these are fun to read even if you're not a CS geek.)

Also, he was at Brandeis at the same time as my stepmother.

I find this archive is sucking my time. It has many more Lamport papers than my short stack from grad school, and it's easy to poke around wondering "what's this one about?", but hard to stop reading once I've gotten far enough into one of them to answer that question. Since it's got me hooked, I feel like spreading the affliction. Err, the wonder. Yeah.

"A distributed system is one in which I cannot get any work done because a machine I've never heard of is down." -- Leslie Lamport

Comments

[rmd.livejournal.com]

!
i hadn't realized that lamport had ended up at msft research. cool.

i don't know how it is now, but back 10 years ago, the research folks got lots of freedom to do weird shit that might be completely unlikely to be productized at any point in the near future, which is a good thing.

[eirias.livejournal.com]

I didn't actually know you went to grad school!

[cos]

I got a masters in computer science at Brandeis, where I just continued after undergrad, so that makes it less obvious in my backstory because I only went to one college/university. I just rediscovered some of my old transparency (overhead projector) presentations on information theory, and consensus protocols, in David Wittenberg's class. He's the one who introduced me to the joy of reading Lamport papers.

[justinjs.livejournal.com]

Excellent timing on this post. Just yesterday I explained Lamport Logical Clocks to a colleague, and suggested that he should go read some of Lamport's papers.

Now I have a nice central collection to point him at. Thanks for the link!

[sauergeek.livejournal.com]

I'd run across references to the Byzantine Generals when reading NTP docs, but didn't know the origin. Neat work from that dude who wrote LaTeX.

[yesthattom.livejournal.com]

Those are also on my list of papers to read/reread! Brilliant.