paying for software

Christoph Gartmann GARTMANN at IMMUNBIO.MPG.DE
Wed Oct 13 07:52:38 EST 1993

In <12OCT199317092534 at> mathog at writes:

> 1. Is it reasonable for government agencies to fund software development
> and then cede ownership of the resulting software to a University?  In other
> words, didn't we already pay for FLEECE? 

These are in fact two questions: If the funding agency keeps the ownership,
that is the ownership is separated from the development, this could result
in problems concerning further development, especially when the funding agency
stops funding. On the other hand this does not mean that the agency should
cede ownership without any conditions.
Concerning the second question: yes, I feel you paid already for FLEECE
(but not me, I am German ;-) ).

> 2. What are the appropriate criteria for saying that a government agency
> did or did not pay for the development of a particular piece of software?
> This one is not quite so simple as it might appear - for instance they may
> have purchased all of the machines and compilers, but not paid for the
> graduate student stipends, or vice versa.

I think something like "who pays has claims". If the software is a direct
result of the funding then the ownership belongs to the agency (if it payed
the machines/compilers just to create this piece of software). Otherwise
I offer a few compilers to MicroSoft for free and wait... And even if the
funding is only partial (but somewhat substantial) the funding agency has 
some claims like the software should be distributed freely or may be used
by the public (but remains on a single host). This applies to governmental
or "public" funding. But the type of claims is a matter of negotiation.

> 3. What is a reasonable distribution cost for software retrieved over the 
> Internet?

This may depend. It is mainly a question of support. But provided that the
Internet connection has already been there the additional costs should not
exceed a nominal fee.

> 4. Can anybody think of a justification for charging per user license fees
> under the NSF software guideline shown above?  After all, the software is
> distibuted only once, so there should be no further duplication or
> distribution costs once a single copy has been set up on a multiuser
> machine. 

Don't forget the main issue of support. Just providing the software and
documentation as it is, will not require much additional costs. But when
you start to answer questions, to fix bugs,... this can result in quite
a lot of real work. And this work depends upon the number of users. Next,
it is a question of further development. You may need the money to ensure
the development. But the costs of development don't depend on the number
of users.

> 5. For publicly funded software, are there any cases where it should not be 
> required that software be made available via anonymous FTP over the Internet?

This could result in additional costs. Next, what about programs for
research? Lot's of genomic sequences are retained until publication. Some
of them have been patented. It seems to be the same issue.

> 6. Should we be giving grants for software development to researchers at
> Universities that have policies like the one in question?

If you can get the same software (same quality, functionality, reliability
of further development and support) from somewhere else, no.

  Christoph Gartmann

| Christoph Gartmann            Phone   : +49-761-5108-465   Fax: -221       |
| Max-Planck-Institut fuer      PSI     : PSI%(0262)45050160374::GARTMANN    |
| Immunbiologie                 Internet: gartmann at           |
| Postach 1169                                                               |
| D-79011  Freiburg, FRG                                                     |
+------------- Do you know MENUE, the user environment for VMS? -------------+

More information about the Bio-soft mailing list