Re: Sponsorship requirements and copyright files
Sune Vuorela <nospam@vuorela.dk> writes:
> On 2009-03-19, Ben Finney <ben+debian@benfinney.id.au> wrote:
> >> It is a too complex, overengineered solution to a very minor issue.
> >
> > I find it very surprising that someone can be a Debian developer and
> > consider copyright of works to be ???a very minor issue??? in Debian.
> > Perhaps I've misinterpreted this statement. What do you mean by that?
>
> PLease read again. I'm not anywhere discussing the *content* of the
> copyright file, but the *format*.
>
> one of the issues the format should solve is "automatic detection of
> license incompabilities", which is not a big issue today.
That would be a minor issue. I haven't seen anyone propose it.
What I have seen is the proposal that this file should be
automatically *parseable*. This is IMO a big issue, because it means
the difference between hunting manually through unstructured copyright
information that can get hideously complex (as you describe below),
versus the potential with automatically-parsed information to
structure, index, search, and otherwise manage the information more
powerfully and efficiently, thus saving valuable human time.
> I like the current "free for all" format, where you can adapt the
> format to the requirements and differences of the packages.
Do you feel this benefit outweighs the lack of dependable structure,
that currently imposes all the work of filtering and parsing the file
on the human reading it?
> > I find the [proposed format] structure makes it far easier to
> > write and check than the free-form chaos of many existing files.
> > What would you have removed from the format to reduce the time for
> > writing and checking it?
>
> Try do it with a bit larger package. it does not scale.
This surely applies even more to a free-format dump area for copyright
information.
> I think when uploading kde4.2 to unstable, at least 60 developer
> hours was put into working on the copyright files, even with loads
> of help from various scripts.
You are of the opinion, then, that it would be *harder* to work with
the masses of copyright information for that package if the
information was in a regular, well-specified structure?
I'm not saying that the current draft of the specification meets that
description; but I can't see how what you say is anything but an
argument *for* a well-structured copyright file format.
> Is this the right way to spend developer time? as far as I see it,
> developer time is our most valuable resource, and should not be
> treated as such.
Certainly, the time of people is valuable. I would like to see a
format that is structured such that it's easier than free-format text
to navigate, and makes it possible to write tools to automate many of
the associated tasks that you say take up too much time.
> A real gain would be something that made tedious work less tedious.
> Copyright files are tedious to write. making it more complex will
> not improve this.
I argue that a well-designed structure significantly *reduces* the
natural tendency for free-format text to become even more complex.
--
\ “To me, boxing is like a ballet, except there's no music, no |
`\ choreography, and the dancers hit each other.” —Jack Handey |
_o__) |
Ben Finney
Reply to: