Hello, some groups of packages in Debian share introductory pieces in the package description. For example, most pike packages have this:
Pike is an interpreted, object-oriented, dynamic programming language with a syntax similar to C. It includes many powerful data types and a module system that, for instance, provides image manipulation together, with support for graphics formats like SVG, JPG, PNG, GIF, XCF and many others, database connectivity, advanced cryptography, XML/HTML parsers and others. To learn more about pike, please visit http://pike.ida.liu.se/ While these blurbs are informative, they provide information not strictly related to the package itself. As a consequence, if you do "apt-cache search image" you get all of pike, including irrelevant things like "pike7.6-public.network.pcap" or "pike7.6-public.protocols.syslog". This is not normally annoying in simple apt-cache search queries, but it becomes nasty when trying to do some smarter text mining on the package descriptions. Think bayesian tools[1], or my new algorithms for mapping a keyword search to a tag search[2]. Many thanks to the KDE developers for removing the similar blurb that they used to have. They did it nicely, and in a way that others could follow. Would it be worth adding a mention to this to the package description part of the developers-reference? Ciao, Enrico [1] for example, we have something cooking up as a Summer of Code project: http://wiki.debian.org/SummerOfCode2006 (see DebtagsAI) [2] http://lists.alioth.debian.org/pipermail/debtags-devel/2006-July/001292.html -- GPG key: 1024D/797EBFAB 2000-12-05 Enrico Zini <[EMAIL PROTECTED]>
signature.asc
Description: Digital signature