Hi Ron, Thanks for your answer!
> There is very rarely a perfect name for a module. > > The considerations you've mentioned indicate a good attempt to take past > and future other module names into account. Often, that's the best we > can do... > > The standard help for such matters is: > > http://pause.perl.org/pause/query?ACTION=pause_namingmodules > > I think your use of Bucket is not very indicative of what you module > offers. I would not want it to give me buckets :-). > > How about Statistics::Descriptive::Approx? > > That makes sense to me because it indicates a close link to > Statistics::Descriptive::Full while distinguishing itself by > specifically stating what's different. > > That still leaves Statistics::Descriptive::ProbResampling. > > As for the length of the name, at this very moment I'm writing > > MarpaX::Demo::OldInterface::JSON. I've had a long think about this name, > too, since there are already 2 other interfaces (BNF, Scanless), even > though they as yet have no demos. > > The longest name in the disto will be > MarpaX::Demo::OldInterface::JSON::Renderer::JSON. > > Such is life. It suddenly occurred to me that no more than a few modules providing exact statistics are needed. Thus the rest would some kind of space/time/precision tradeoff. So I can really use the Statistics::Descriptive:: namespace without worry. I would like to specify algorithm in the name instead of just "approx", though. The "Bucket" part was a total mistake, it's great that I realized it. What about Statistics::Descriptive::LogScale? P.S. I'm crossposting to module-authors@ as Shlomi suggested in his reply to my initial post. -- WBR, Konstantin S. Uvarin
