Re: Copyright from the lcs-projekt!? [dwarf@polaris.net: Re: First cut at testing and validation]
On Wed, 12 Aug 1998, Dale Scheetz wrote:
> On Wed, 12 Aug 1998, Jules Bean wrote:
>
> > I can't quite agree with Dale's approach on this.
> >
> > IMHO, the 'validate' program should be modifiable, but should have a
> > clause dictating that any modified version must be distributed under a
> > different name, and clearly marked as such.
>
> And just what purpose would this "new" forked version serve? Validate a
> different standard? Do some other job instead?
>
> What "useful algorithm" am I keeping out of the hands of the rest of the
> community by not allowing modification? The major portion of this package,
> which is not visible in the script I published are the files containing
> the lists of objects being checked. Change one character in any of these
> lists and the validation proceedure will no longer be useful.
>
> This is a piece of software that serves a specific, narrow, purpose.
> Allowing it to be modified only dilutes the strength of the test and the
> standard.
The question is, are we in the business of making exceptions? If your
program does genuinely warrant an exception, then the DFSG is broken, and
needs to be modified slightly to allow it in.
As to reason to modify: Yes, to validate a different standard would be
one. To correct a foolish bug which you made just before you went on a
two-week holiday would be another. To propose for discussion a possible
alteration would be a third - as in 'Here Dale, I think validate should
work like *this* - what do you think of my version?'.
>
> >
> > Furthermore, the LCS team could have on their websites the sizes and
> > md5sums of 'validate' (and any other scripts, and documents, which have
> > similar issues attached to them). This means that it is possible to
> > verify a given copy of 'validate' as being correct, buyt still distribute
> > it as free software.
> >
> This is already the plan. The tarball of the validation suite will be
> found along with a README.lcs-validate that will give the md5 sum for the
> tarball to ensure the proper result.
>
In which case, given that you are taking these precautions, I don't
understand why you object to making it modifiable. There would seem to be
no danger of subversion, if you create the README file you describe
(especially if you sign it).
> > Dale: Would this not be appropriate?
> >
> It only serves the letter of the Free Software ideal, not the spirit. I
> don't see a single advantage to allowing modifiability, and several
> disadvantages to doing so.
>
I see a few advantages, and no disadvantage...
Jules
/----------------+-------------------------------+---------------------\
| Jelibean aka | jules@jellybean.co.uk | 6 Evelyn Rd |
| Jules aka | | Richmond, Surrey |
| Julian Bean | jmlb2@hermes.cam.ac.uk | TW9 2TF *UK* |
+----------------+-------------------------------+---------------------+
| War doesn't demonstrate who's right... just who's left. |
| When privacy is outlawed... only the outlaws have privacy. |
\----------------------------------------------------------------------/
Reply to: