qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Python docstrings and licensing/authorship


From: Daniel P . Berrangé
Subject: Re: Python docstrings and licensing/authorship
Date: Wed, 16 Sep 2020 17:37:14 +0100
User-agent: Mutt/1.14.6 (2020-07-11)

On Wed, Sep 16, 2020 at 12:22:37PM -0400, John Snow wrote:
> On 9/16/20 12:05 PM, Christophe de Dinechin wrote:
> > 
> > 
> > > On 16 Sep 2020, at 17:47, John Snow <jsnow@redhat.com> wrote:
> > > 
> > > For some of the Python cleanup work I am doing, I am moving preamble 
> > > comments into docstrings. These docstrings are visible in interactive 
> > > editors and may be visible when using Sphinx to generate documentation 
> > > manuals for Python code.
> > > 
> > > My instinct is to remove the licensing and authorship information from 
> > > the preamble and leave the docstring only with information functionally 
> > > relevant to the module, while leaving licensing and authorship 
> > > information in a comment (above? below?).
> > > 
> > > The end effect would be that using e.g. `help(qapi.parser)` in the 
> > > interactive Python shell would not show licensing or copyright for the 
> > > module, but it would still be visible in the source file, as always.
> > > 
> > > Is this in bad taste? Do we have strong feelings about authorship and 
> > > licensing being visible in help output and generated documentation?
> > 
> > What about putting that in a separate pseudo-entity, so that 
> > help(copyright) would show it?
> > 
> 
> help is a Python builtin that shows metadata about an object. If an object
> has a docstring, it is capable of displaying that as part of the help
> output. I'm not sure what type you are suggesting `copyright` to be.
> 
> So, you could do something like:
> 
> __copyright__ == """
> Copyright (C) 2020 John Snow, for Red Hat Inc.
> """
> 
> And you could then theoretically do:
> 
> >>> import qapi
> >>> qapi.__copyright__
> 
> which will show you that information. However, I don't think there's any
> standard for module-level metadata like this, so the odds of this
> information being seen or used is low.
> 
> Python has some standards for package-level metadata, but I don't think
> there are any standards for module-level metadata *except* the __doc__
> attribute -- which is where the module-level docstring goes when we write
> one.
> 
> The real question I have is if anyone thinks it would be "rude" to separate
> out any of the comment preambles (currently not visible at runtime or docs
> in any way, shape or form!) into two pieces:
> 
> 1. Functional stuff relating to the usage of the module, visible in
> help(module_name), visible in generated docs, visible in IDE popups, etc.
> 
> 2. Authorship/copyright and licensing info, not visible in the above places.

I think this makes sense. IME it is not common to include copyright /
author info the module help text, as that s non-technical information.

Including author info in API docs in particular I think is liable to
be harmful becuase it will encourage users to directly email
individual authors for help, rather than using collective comms
channels like the mailing lists or forum or bug tracker.

And of course all of this author and copyright info is generally
horribly out of date, as people rarely add themselves when modifying
existing files.

As long as you don't remove the copyright info entirely, that satisfies
the license requirements to preserve copyright notices.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|




reply via email to

[Prev in Thread] Current Thread [Next in Thread]