Re: [Axiom-developer] Re: [Axiom-commit] SF.net SVN: axiom: [426] branches/wh-sandbox

[Waldek Hebisch]

> On Thu, 1 Feb 2007, Waldek Hebisch wrote:
> 
> | Gabriel Dos Reis wrote:
> | > On Wed, 31 Jan 2007 address@hidden wrote:
> | >
> | > [...]
> | >
> | > | Modified: branches/wh-sandbox/src/etc/asq.c.pamphlet
> | > | ===================================================================
> | > | --- branches/wh-sandbox/src/etc/asq.c.pamphlet  2007-01-27 16:50:57 UTC 
> (rev 425)
> | > | +++ branches/wh-sandbox/src/etc/asq.c.pamphlet  2007-02-01 05:09:31 UTC 
> (rev 426)
> | > | @@ -2,1559 +2,929 @@
> | > |  \usepackage{axiom}
> | > |  \begin{document}
> | > |  \title{\$SPAD/etc asq.c}
> | > | -\author{Timothy Daly}
> | > | +\author{Waldek Hebisch}
> | >
> | > I believe the conventional wisdom would suggest that you add your name
> | > as co-author as opposed to replacing the existing author, even if you
> | > have rewritten it. If this were a separate new program not building on
> | > previous ideas, I would that it is OK.
> | >
> |
> | Hmm, I undersand that conventional wisdom is to err on side of giving
> | more credit for work than too little credit, but I find this suggestion
> | a little extreme.
> 
> I'm not so sure.
> 
> | Gaby, did you look at both programs: the new asq is a _new_ program.
> 
> I read over your new "asq" and the old asq and followed your
> explanation, *before* I sent my mail.
> 
> My opinion is that your name should be added co-author, not replace
> the original author.  Or, call the program something else and
> acknowledge that this new program greatly benefited from the existing
> one, in particular from its documentation and its errors. But having
> two proram doing almost the same thing will be very confusing.  So, I
> believe the conventional wisdom is good here.
>

I fail to see how filename affects authorship status of a file.  FYI
I developed under different name outside of Axiom tree.  I started
copying a few lines from old asq but quickly realized that old code
do not fit into my design.  So I wrote the program essentially from 
scratch.  <heresy> After the program was finished I have put literate
wrapper around it and dropped it into the source tree </heresy>.
IMHO the few copied lines are not more significant than a few declarations
that I copied form man pages.

Note that I do not claim any credit for the old asq program.  And
I think that we should prominently state that previous version was
due to Tim Daly.  But the \authors{} construct seem too limited
to correctly express such information.

AFAICS changing file name limits possibilty for
misunderstanding authorship, but is an inconvenience for users
(still, if you guys think that renaming is a way to go I will do
it).  I do not consider keeping the old program as an option:
main reason to write the new one was that the old one does not
work with other changes that I am making.
 
> A ways to kepp a project attractice and collegial is to be cautious
> with credit.
> 
> Look at GCC for example.  Most parts of it have nothing common with
> the GCC from 1996 or 1997.  In particular, look at the C preprocessor;
> it has been *completely* rewritten from scratch, with readically
> different ideas.  Nonetheless, the authorship displays the people who
> did not "directly" contribute to the new program in any form.
> 

I do not see information about authors in preprocessor program files,
only note like:

   Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
   1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006
   Free Software Foundation, Inc.
   Contributed by Per Bothner, 1994-95.
   Based on CCCP program by Paul Rubin, June 1986
   Adapted to ANSI C, Richard Stallman, Jan 1987

IMHO such note is misleading by ommision, but at least dates allow
to infer that this information is highly incomplete.  I see
Richard Stallman mentioned as a coauthor of cpp.texi, but comparing
cpp.texinfo from gcc-1.35 with cpp.texi from gcc-4.1.1 I can find
a few passages are taken almost verbatim from the old version.
I would guess that careful examination would show more connections.
So I consider it reasonable to state that Stallman is a coauthor
(assuming he did wrote version contained in gcc-1.35). 

Concering programs: I did not look at preprocessor files.  But I did
found surprising similarities between some parts of gcc-1.35 and 
parts of gcc-4.1.1 -- after many edits code looks very different,
but clearly is still the same code.

-- 
                              Waldek Hebisch
address@hidden