Re: What search.cpan.org PAUSE produce (Fork from: what to do with dead camels?)
On Aug-04, James E Keenan wrote: Searching via All: 1st distro appearing is Test::More as part of Palm-Progect-2.0.1 by Michael Graham. Schwern's Test::More appears 4th on list. Note: In recent weeks 1st distro appearing under this search was Test::More as part of parrot-0.0.010 by Steve A Fink. Tonight, this distro is 2nd on list. More data for the pot: when I uploaded that file, the indexer sent me back a long list of all the modules it found within it, including this: Test::More version: 0.41 in file: parrot-0.0.10/lib/Test/More.pm status: Not indexed because permissions missing. Visit PAUSE and click View Permissions to find the legitimate maintainer(s). Shortly thereafter, I received mail from a maintainer saying that he'd help me resolve the huge number of indexing problems. (Unsurprisingly, Parrot contains lots of modules.) I wrote back, saying I'd rather none of it got indexed in the first place, and he said he'd looked at what it was and agreed that nothing needed to be done because it wasn't appropriate in this particular case. (I'm still looking for the actual email; sorry for the vague description.) He also said there's no way to suppress the indexer. This suggests that there should perhaps be ways to say things like I'm just repackaging this module for convenience; I am NOT an authoritative source for it. Or, as in my case, I ship with this module but installing my software will not make the module available to anything outside my software so please forget you ever noticed its presence. But I won't suggest that, because I don't understand the system to know whether that would make any sense. :-)
Re: What search.cpan.org PAUSE produce (Fork from: what to do with dead camels?)
* Steve Fink ([EMAIL PROTECTED]) [05 Aug 2003 19:12]: [...] He also said there's no way to suppress the indexer. At the time, yes. A format using the META.yml file has sprung up. Check out your local cpan mirror at the file: $cpan/modules/by-authors/id/A/AN/ANDK/Test-Private-PleaseIgnore-0.001.meta Note the 'private' keys. This will at least ensure search.cpan.org won't index the modules. The PAUSE indexer, I don't know -- I think it's being worked on at present? (Note that the *.meta files are the META.yml files from within the distribution and are automatically extracted, just like the README files.) If you check some of my distributions ( http://search.cpan.org/~spoon/ ) you'll find other example META.yml files. cheers, -- Iain.
Re: What search.cpan.org PAUSE produce (Fork from: what to do with dead camels?)
On Mon, 4 Aug 2003, James E Keenan wrote: May I begin a separate thread for a line of discussion coming up under the dead camels? Sure. I'm going to present empirical observations only; it would be premature to make suggestions for changes until we heard from more contributors. It's never stopped us before. :) Bru-hahahaha. Searching via All: 1st distro appearing is Test::More as part of Palm-Progect-2.0.1 by Michael Graham. Schwern's Test::More appears 4th on list. (A) Do we have any idea what tha algorithm it's using is? (B) Could we make an exception/improvement to the algorithm which makes the real module come up first? Searching via Modules: 1st distro appearing is Test::More as part of CPAN-1.76 by Andreas J Konig. Schwern's Test::More does not come up at all among 471 entries found. Searching via Distributions: Test::More does not come up at all (C) Couldn't we have a seperate box that comes up on the side of the search results that shows the version/date of the last three releases of a given module X if the search string put in is reasonably obviously referring to a specific module. So it might look something like this when searching for Foo::Bar, Foo-Bar, or foo bar: 1) Irrelevant ResultModule Foo::Bar 2) Useless Result v0.1 4-1-1870 3) Misleading Result v0.2 4-1-1970 4) Ancient Result v0.3 4-1-2003 5) Irrelevant Result 6) Irrelevant Result 7) Useful Result too low to see 8) Distracting Result Issue 2: Which links PAUSE builds when an author uploads a module whose POD contains links? I'd love to know the answer to this too. Do others observe similar phenomena? Is it problematic for you? Ironically, search.cpan.org is phenomenal for browsing perl module docs, but it's really strange and bewildering when it comes to searching them. -- /chris The death of democracy is not likely to be an assassination from ambush. It will be a slow extinction from apathy, indifference, and undernourishment. -Robert Maynard Hutchins, educator (1899-1977)
Re: What search.cpan.org PAUSE produce (Fork from: what to do with dead camels?)
On Wed, Aug 06, 2003 at 04:06:34AM -0500, elaine wrote: On Tue, Aug 05, 2003 at 01:09:52AM -0400, Christopher Hicks wrote: (A) Do we have any idea what tha algorithm it's using is? Soundex for everything but Authors and Docs which then uses CPAN::WAIT. I suppose that explains why searching for lbdb with All takes me directly to Acme::Pr0n. (by contrast: kobesearch just returns no results on the default search. ) Is there is a Soundex::FamilyFriendly::Filter out there? I wasn't expecting the Pr0n to popup there like that on my innocent query. :) Mark
What search.cpan.org PAUSE produce (Fork from: what to do with dead camels?)
May I begin a separate thread for a line of discussion coming up under the dead camels? The discussion suggests we should also look at (1) what search.cpan.org comes up with in searching for modules; and (2) what links PAUSE builds when it extracts POD from an uploaded module. I'm going to present empirical observations only; it would be premature to make suggestions for changes until we heard from more contributors. On Sunday 03 August 2003 17:45, Andy Lester wrote: Do a search on search.cpan.org for HTTP::Response, a pretty common module. The first hit that comes up is the one from lcwa, and if you're not paying attention to the distro name (or you're a relative newbie who doesn't realize he needs to), you're going to be looking at 6-year-old docs for the module. Today, Fergal Daly [EMAIL PROTECTED] added: Try Test::More, it's true home is Test::Simple but that's 5th on the list. Issue 1: What search.cpan.org comes up with when searching for a module. Tonight I searched for Test::More via each of search.cpan.org's 4 search modes. I was looking to see where MSCHWERN's Test::More appeared in each of the modes. Results: Searching via All: 1st distro appearing is Test::More as part of Palm-Progect-2.0.1 by Michael Graham. Schwern's Test::More appears 4th on list. Note: In recent weeks 1st distro appearing under this search was Test::More as part of parrot-0.0.010 by Steve A Fink. Tonight, this distro is 2nd on list. Searching via Modules: 1st distro appearing is Test::More as part of CPAN-1.76 by Andreas J Konig. Schwern's Test::More does not come up at all among 471 entries found. Searching via Distributions: Test::More does not come up at all Searching via Authors: Schwern does not come up; 1st author listed is Huascar Tejeda (HUSOFT) -- who has no modules listed at all, so it's not clear why the search engine came up with him. Issue 2: Which links PAUSE builds when an author uploads a module whose POD contains links? This issue came up when I was studying Fergal's POD for his Test::Deep module (http://search.cpan.org/author/FDALY/Test-Deep-0.06/lib/Test/Deep.pod#SEE_AL SO). The POD appears simply as: LTest::More PAUSE links this to the version of Test::More described above as part of parrot-0.0.010 by Steve A Fink. I suspect that on the date this distro was uploaded, Steve's Test::More was still the 1st appearing when searching via All and that this may be why PAUSE linked to this version, when Fergal would probably have preferred the link be to Schwern's version. Do others observe similar phenomena? Is it problematic for you? Jim Keenan
