VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Thu Jun 03 2010 - 04:59:08 CDT

hi francesco,

On Thu, Jun 3, 2010 at 3:50 AM, Francesco Pietra <chiendarret_at_gmail.com> wrote:
> Hi:
> A from time to time recurring question: any date now for providing
> paratool for free qm suites? Today buying commercial qm codes that are
> slower than certain free qm codes is no alluring perspective. To
> justify my question, try for example free ORCA for (parallelized) CD
> calculations and compare it with commercial codes.

you are raising an important issue. unfortunately, the answer
will not be something to your liking. the problem with "free"
tools is, that they don't really come for free. somebody has to
write them and so somebody has to pay for that person's time.

if VMD (or any other free package) was a commercial software,
the question would be easily answered. customers pay for the
use and new features are implemented according to the (perceived)
need of customers or according to what is considered important
to gain new or keep current customers paying for the software and
it being supported.

and this is where the rubber meets the road: there is currently
next to no proper scheme to finance work on free software that
supports implementing features based on the needs of users.
most of the time scientific software development is cross-funded
by research projects and thus the individual project and the
initiative, needs and resources of the individual researcher
determine what gets implemented and how. there is some
additional work done by people who are just enthusiastic about
implementing scientific tools and supporting them, but that
is certainly not a foundation for a "professional style" development
for a "professional quality" software package.

the situation is getting worse if you look at packages like VMD
who have over time grown into highly capable and flexible tools
with a lot of sophisticated add-ons that have been developed
based on it flexible foundation. people have gotten used to it
and for many frequently used components, it is a taking only
a little effort to keep them up-to-date. however, quite a few
components are ageing or have not been completed and
their developers have since moved on (as is the nature of
our profession) and their work has been fallen by the wayside
and due to the boundary conditions of research funding, it
is highly unattractive to work on completing or maintaining
these kind of tools.

this is made even worse by the fact, that the ratio of users
that are able to make modifications to the code versus those
that depend on others implementing missing features is
changing significantly to the latter group.

bottom line:
if you want something done in a package like VMD, you
have two options: a) learn how to do it yourself and do it.,
i.e. invest time, or b) hire/pay somebody to do it.

the latter is something that people will have to get used to,
since the first option does not seem to be working (anymore).
even worse, it is hard to even get proper credit for the work
being done. people using free software have developed a
habit of taking it for granted and often cannot tell which
component was written by whom, and why should they care
in the first place; all they want is to get the job done, and
this is why tools are written in the first place.

looking back the 15+ years that i have been involved in
using and programming scientific software, the change
is quite massive. 15 years back it was no question that
one first had to learn programming and write tools by
yourself to get anything done. but those tools were limited
and not very sophisticated. when larger integrated packages
were emerging, it was only natural to contribute and a
way to grow and learn and gain the respect of your peers,
and at the same time pay those people back for their
effort by investing your time and effort into their software.

however, this breed of people is dying out and we have
to find a way to put this on a more solid footing or we will
end up with a situation where everybody wants to use
scientific software packages, but nobody knows how to
write and maintain them. there have to be ways to make
it desirable to implement features that people want, but
cannot implement themselves.

cheers,
   axel.

> thanks
> francesco pietra
> 

-- 
Dr. Axel Kohlmeyer    akohlmey_at_gmail.com
http://sites.google.com/site/akohlmey/
Institute for Computational Molecular Science
Temple University, Philadelphia PA, USA.