Norman,
> Such macros really are absolutely useless outside the tree, though they
> have to be in the tree somewhere. Similarly, the tools that Steve uses
> to do the nightly builds, or to construct the CDs, should be in the
> tree somewhere, but are undistributable, since they are meaningless to
> anyone not working with a checked-out version of the code. As such,
> these don't fall into any of the categories `applications, buildlog,
> docs, java, libraries, pubs, thirdparty' -- the autoconf macros are
> not an `application', and if we force them into that category, we'll
> just confuse ourselves and others, and be obliged to have a separate
> way of recording which `applications' are sanely distributable and
> which are not.
>
> C'mon, folks -- is this really this contentious?
Not really. I just think we should be careful to be clear about what goes
where. We want the categorisation to be as orthogonal as possible. So
if the primary distinction of the items in this tree is that they
cannot be distributed, then why not "internal" or "private" or
"undistributable" or somthing? I am assuming here that the concept of
"distributability" is well defined to you CVS experts!
Or maybe I've just been arguing about IVOA data models for too long!
(Conan the Modeller, for sooth!)
David
----------------------------------------------------------------------
Dr David S. Berry ([log in to unmask])
STARLINK project | Centre for Astrophysics
(http://www.starlink.ac.uk/) | University of Central Lancashire
Rutherford Appleton Laboratory | PRESTON
DIDCOT | United Kingdom
United Kingdom | PR1 2HE
OX11 0QX
|