[d4rt]

Disclaimer: I work for a Splunk Partner.

Interesting technique, although I think a domain specific extension of Haskell for analysing log records might be more useful.

I'm sure there are more complex examples, but all the tasks in the screencast could be done easily with Splunk. It might, if you want to develop the idea of log analysis further to look at their search cheatsheet. http://www.splunk.com/base/Documentation/latest/SearchRefere...

Most of the syntax in the screencast is relatively guessable or explained, however NubBy means remove duplicates using the provided test condition. http://www.zvon.org/other/haskell/Outputlist/nubBy_f.html

[mightybyte]

I've used and developed on Splunk at work and you're completely correct. You can also get this information using AWStats or something similar to analyze Apache log files. It's not surprising that software marketed for the task of log analysis can do better.

The point of the post is to demonstrate how a practical task can be done easily in a programming language that has been accused of being too academic and impractical.