Project fork

[kernc]

Dear Maintainers,

I came across pdoc a couple of months ago and I immediately felt tremendous appreciation for all of its core ideas:

• public identifiers are auto-discovered following common conventions,
• whole documentation relies on a well-known language feature (docstrings),
• documentation hierarchy follows module hierarchy,
• module docstrings are used to introduce the topic,
• identifiers can easily be referred to with backticks, and
• the default documentation language is a flavor of Markdown which, compared to reST, pretty much everyone knows and adores.

This makes it a perfect documentation utility for small, tidy projects.

So it was with an amount of grief I noticed the repository recently took an overhaul which introduced, on balance, quite a few regressions as far as my use cases go. As no new releases have been made, I understand the development is not finished, but it's initial vigour certainly seems stalled at the moment, and the humble pull request I tried to squeeze in in order to restore some of the previous behavior got no reviews at all. Today with open source abound, PR negligence certainly doesn't instigate new developer/maintaner activity.

Thus, I'd like to take the opportunity to announce I had recently started rapidly applying various sensible changes to a diverged fork of this project. The split point was 03b96f4, about right before the package layout was, in my view needlessly, restructured. Elsewhere following your example, I took what I could from the upstream master (like --filter switch and module importing logic).

I wouldn't be as blunt as to propose you reduce the bus factor of the project back to 1.0, nor that you overwrite your history with this augmented one, but should you by chance decide to ever discontinue maintainership, I'll be happy to take over the project I'll at the time quite likely still be most actively using.

Refs: #148

[cortesi]

Thanks for letting us know, and we've discussed this somewhat by email. The bus factor will return to normal as soon as time permits, and in the meantime we'll keep an eye on your repo.

[untitaker]

After I filed a bug report against this repo, I was contacted in private by @kernc asking me to try out https://github.com/pdoc3/pdoc, a fork of pdoc, confusingly also named pdoc (sometimes pdoc3). Almost all information linking back to pdoc3's origins have been removed from its docs, the Python wiki page has been edited to link to his fork (not necessarily by him) while simultaneously removing the link to @BurntSushi's repo (which redirects to mitmproxy/pdoc). Those occurences make me believe that @kernc's fork is both hostile and attempting to take over the name pdoc from the original project without upstream's consent.

That fork is also relicensed under AGPL, which seems to diverge from @BurntSushi's usual license choices. So it can't possibly be called a successor like @kernc tries to make it seem.

For the sake of transparency, @kernc's reasoning for contacting me privately was:

I was asked not to participate on the tracker so I am writing to you personally.

I have no idea what he meant by that.

[cortesi]

Hi there! I'll have a look at the licensing situation. Kernc is very welcome to participate here, but he would have to change his avatar to something that doesn't so closely resemble a swastika.

[BurntSushi]

The relicensing is likely perfectly legal. pdoc does use the UNLICENSE after all. If it were the MIT then they would be obliged to retain that license somewhere in the source repository.

Of course, just because it is legal doesn't mean it's right. I consider @kernc's actions here to be rude and unethical. Keeping the same name in a hostile fork is clearly also an exercise in poor judgment.

[untitaker]

I noticed the swastika as well, but I wasn't entirely sure if mentioning it here would derail the conversation. Thanks for calling him out @cortesi.

I also just noticed that the relicensing removed copyright statements, which is fine for unlicensed code I suppose (just like @BurntSushi said)

I contacted the pydotorg-www mailinglist about the wiki page entry.

[Rosuav]

From the point of view of the Wiki, I'd like to just say that it is rather rude to change the link to point to the new fork without even notifying the present maintainers that you have done this. Please exercise some caution when making changes of this nature. I hope that the maintainers of both the original project and the fork can come to some form of agreement about how the projects are to be advertised (for instance, if the fork chooses a new name, both could be listed on the wiki, with an explanation of the distinctions). If consensus is not possible, please do NOT engage in an edit war on the wiki itself, as this will serve only to confuse end users and give everyone involved a bad name (including the wiki, for being untrustworthy).

For the time being, I have reverted the most recent change. I'm going to let the matter be debated here rather than make a strong stance myself, as I'd like to believe that everyone involved can behave maturely and come to an agreement. :)

[cortesi]

Sorry, this is not a conversation we want to host here.