----- Forwarded message from Kragen Javier Sitaker <kra...@canonical.org> -----
Date: Tue, 18 Oct 2011 03:04:13 -0400 From: Kragen Javier Sitaker <kra...@canonical.org> To: Darius Bacon <dar...@wry.me> Subject: Re: radix-sorting rational numbers with an efficient serialization of continued fractions Message-ID: <20111018070413.ga12...@canonical.org> References: <20111017225537.ga1...@canonical.org> <201110180002.p9i02y5h003...@wry.me> <20111018062120.ga9...@canonical.org> <201110180658.p9i6wygg005...@wry.me> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <201110180658.p9i6wygg005...@wry.me> User-Agent: Mutt/1.5.20 (2009-06-14) Status: RO Content-Length: 920 Lines: 21 On Mon, Oct 17, 2011 at 11:58:34PM -0700, Darius Bacon wrote: > > I think the search engine I wrote in 2006 > > <http://lists.canonical.org/pipermail/kragen-hacks/2006-August/000432.html> > > might benefit from using Golomb coding or something similar for its > > pseudo-posting-lists. It currently uses the "Altavista trick" to encode > > integers in variable numbers of bytes, 7 bits per byte, with one bit per > > byte > > used as a termination marker, but I suspect that it can probably use > > substantially less space with Golomb coding. > > Altavista trick? Does it really go back no further? (I thought it was > in ASN.1 BER, but apparently that did something more complicated, as > usual in that design.) It could easily be much older, but I've seen it referred to as that and simply as "variable byte encoding". > > Do you mind if I forward your mail and this one to kragen-discuss? > > Not at all. Thanks! ----- End forwarded message ----- -- To unsubscribe: http://lists.canonical.org/mailman/listinfo/kragen-discuss