[Pkg-doc-linux-devel] Checking licenses

Frank Lichtenheld frank@lichtenheld.de
Thu, 6 Nov 2003 23:51:48 +0100


On Thu, Nov 06, 2003 at 10:06:56AM -0700, doug jensen wrote:
> Hi,
> 
> Just some ideas on license checking, with a few questions at the end.
> 
> Automatically checking the licenses of new howtos would be very difficult

indeed

> Checking updated howtos for license changes is:
>   - difficult <= not fun, to do manually.
>   - Easier to do automatically and could improve the checking accuracy.
>     (Documents that have no license still need to be checked manually,
>     but extraction from the tar file could be automated.)
> [some thoughts about checking]
> 
> Questions:
>   - Do you have other or better ideas?

Before creating something like this we should evaluate what's the status
of approches like the OMF XML indices
(see http://ibiblio.org/osrt/omf/ and 
http://www.ibiblio.org/pub/Linux/metadata/)
The link in the debian/rules to http://www.lampadas.org/ seems to be
dead btw.

Something like this would be far more preferable. If we find such an
index that is maintained good, we ideally just have to check the
<rights> tag (or something similar). Perhaphs we just end maintaining
it ourselves ;)

Colin, you seem to have already worked on this but then postponed it.
Can you perhaps elaborate a bit on what you have planned and what you
know about the existing solutions/indices?

>   - Would you be in favor of this approach?
>   - Does the grep file seem appropriate?
>   - Is the format ok?
>   - It would require maintenance, is that ok?

The problem is not the maintainance (There are about 10-20 HOWTO updates
in a month, even manually checking them is a task for half an hour or so)
but the work you have to do on the beginning for all the existing
HOWTOs.

If you want to work on this start with a script that extracts the
tarball with the text HOWTOs, gets a list of updated ones
(debian/diff-tarballs, *hint, hint* ;) and outputs the license we
currently assume for them (by searching in debian/copyrights).
This should not be to complicated at all.

The _second_ step would then be to try to add a check if our information
is still correct.

Gruesse,
-- 
Frank Lichtenheld <frank@lichtenheld.de>
www: http://www.djpig.de/