はてなブックマーク

SBCL was accepted as a mentoring organisation for Google’s Summer of Code 2013 (our list of project suggestions is here). This will be our first time, so that’s really great news. I’m also extremely surprised by the number of people who’ve expressed interest in working with us. I was going to reply to a bunch of emails individually, but I figure I should also centralise some of the stuff here. EDI

EDIT: 2012-08-29: I added a section to compare comparison counts with known bounds for general comparison sorts and sorting networks. In an earlier post, I noted how tedious coding unrolled sorts can be. Frankly, that’s the main reason I stopped at leaf sorts of size three. Recently, Neil Toronto wrote a nice post on the generation of size-specialised merge sorts. The post made me think about that

Searching in a short, sorted vector is a fairly common task. Lots of time, data sets are just small enough that it’s quicker (and simpler) to store them in a sorted vector than to use an alternative with asymptotically faster insertions. Even when we do use more complex data structures, it often makes sense to specialise the leaves and pack more data in a single node. How should we implement such