Bug#843328: O: dimbl - Distributed Memory Based Learner for Natural Language Processing
Hi Bálint, On Mon, Nov 07, 2016 at 08:45:07PM +0100, Bálint Réczey wrote: > 2016-11-05 22:17 GMT+01:00 Joost van Baal-Ilić >: > > > > I am no longer interested in maintaining dimbl. Upstream (Maarten van > > Gompel) > > feels shipping dimbl with Debian is of very limited use: people interested > > in > > dimbl likely are better served always using the very latest version of it. > > > > I'll request for removal of dimbl from unstable and testing, soonish. > > I suggest not requesting the removal. The package was practically > unmaintained in Debian for years (last upload: 2013-09-17) while upstream > is not exactly fast moving (last release was in January). > Please delay the removal request by at least 6 months to let others step > up as maintainers. > > I may fix the RC bug with a new upstream during the freeze after all > packages under my name are ready for Stretch. Great, consider this package yours. Thanks a lot, Bye, Joost > Joost wrote: > > Unless someone else steps up RSN, that is. Be aware there is > > serious/RC/FTBFS > > Bug #833905 open (which is relatively easy to fix).) > > > > (From the description: The Dimbl Distributed Memory Based Learner is a > > wrapper > > around the k-nearest neighbor classifier in TiMBL, offering parallel > > classification on multi-CPU machines. Dimbl splits the original training > > set, > > builds separate TiMBL classifiers per training subset, and merges their > > nearest-neighbor sets per classified instance. If you do scientific > > research > > in Natural Language Processing using the Memory-Based Learning technique, > > Dimbl > > will likely be of use to you.) signature.asc Description: Digital signature
Bug#843328: O: dimbl - Distributed Memory Based Learner for Natural Language Processing
Hi, 2016-11-05 22:17 GMT+01:00 Joost van Baal-Ilić: > Package: wnpp > Severity: normal > > Hi, > > I am no longer interested in maintaining dimbl. Upstream (Maarten van Gompel) > feels shipping dimbl with Debian is of very limited use: people interested in > dimbl likely are better served always using the very latest version of it. > > I'll request for removal of dimbl from unstable and testing, soonish. I suggest not requesting the removal. The package was practically unmaintained in Debian for years (last upload: 2013-09-17) while upstream is not exactly fast moving (last release was in January). Please delay the removal request by at least 6 months to let others step up as maintainers. I may fix the RC bug with a new upstream during the freeze after all packages under my name are ready for Stretch. Cheers, Balint > > Unless someone else steps up RSN, that is. Be aware there is serious/RC/FTBFS > Bug #833905 open (which is relatively easy to fix).) > > (From the description: The Dimbl Distributed Memory Based Learner is a wrapper > around the k-nearest neighbor classifier in TiMBL, offering parallel > classification on multi-CPU machines. Dimbl splits the original training set, > builds separate TiMBL classifiers per training subset, and merges their > nearest-neighbor sets per classified instance. If you do scientific research > in Natural Language Processing using the Memory-Based Learning technique, > Dimbl > will likely be of use to you.) > > Bye, > > Joost >
Bug#843328: O: dimbl - Distributed Memory Based Learner for Natural Language Processing
Package: wnpp Severity: normal Hi, I am no longer interested in maintaining dimbl. Upstream (Maarten van Gompel) feels shipping dimbl with Debian is of very limited use: people interested in dimbl likely are better served always using the very latest version of it. I'll request for removal of dimbl from unstable and testing, soonish. Unless someone else steps up RSN, that is. Be aware there is serious/RC/FTBFS Bug #833905 open (which is relatively easy to fix).) (From the description: The Dimbl Distributed Memory Based Learner is a wrapper around the k-nearest neighbor classifier in TiMBL, offering parallel classification on multi-CPU machines. Dimbl splits the original training set, builds separate TiMBL classifiers per training subset, and merges their nearest-neighbor sets per classified instance. If you do scientific research in Natural Language Processing using the Memory-Based Learning technique, Dimbl will likely be of use to you.) Bye, Joost signature.asc Description: Digital signature