Open dstufft opened 10 years ago
From the flipside: classifiers are basically useless as a consumer, I never read them, I just read the prose.
I don't know who decided on the classifiers, either. Are they even possible to use as a search filter?
Occasionally a project doesn't mention Python versions in the README, but does use versioned classifiers. Other than that I've never really looked at them either.
Classifiers seemed like a good idea and were inspired by similar systems at the time like freshmeat.net (you kids probably don't remember that, lol). The list of Trove classifiers were derived from freshmeat's and some others from the FSF IIRC.
Seemed like a good idea, at the time. This was before tagging got organised.
They are searchable, using the "browse" pages. This is not combined with a text search though. I understand some projects use this facility to display a list of packages related to them.
I like the idea of replacing the Trove stuff with a smart tagging system, but it's hard to give users tag suggestions (to help curate the morass) when they're submitting via "setup.py upload" or similar mechanisms. Hence the restricted set dictated by PyPI.
@jonparrott as far as you know, is this the kind of thing that could or ought to go into the Python Packaging User Guide?
Absolutely
On Mon, Mar 12, 2018, 6:49 PM Sumana Harihareswara notifications@github.com wrote:
@jonparrott https://github.com/jonparrott as far as you know, is this the kind of thing that could or ought to go into the Python Packaging User Guide?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/pypa/packaging-problems/issues/31#issuecomment-372519212, or mute the thread https://github.com/notifications/unsubscribe-auth/AAPUc8paMcXziDD1y3UB4sMJ1XPsPYC-ks5tdyWhgaJpZM4BKiY- .
There is confusion about what some of them mean and when you should use them. Currently there is no explanation about what the classifiers mean or should be used for, just a massive list of them.
One example is the natural language ones If a project is coded in English and the primary language of the project/community is English, should they add a Language classifier to it or do those classifiers only apply to things that do natural language processing?