From: Giacomo Fiorin (giacomo.fiorin_at_gmail.com)
Date: Mon Jul 30 2018 - 13:03:55 CDT

Hi Ivan, I simply wanted to state that using version control mechanisms is
possible for contributions to VMD or NAMD.

Even for VMD, where John or other UIUC developers manually merge
contributions into a CVS repository, one can create a GitHub repository (or
any repo that works best) for additional files that go into NAMD or VMD.
The UIUC license even allows to include a small number of affected UIUC
files, to ensure that merge conflicts are addressed. If the number of
affected UIUC files becomes too big, an alternative option is a private
repository that does not break the license. Both options are not
theoretical: they are already being used.

Ultimately, any problem inherent to a certain version control mechanism or
practice is not really a problem for contributors, but only for the UIUC
core developers.

The main deficiency, which *Git by itself does not solve*, is having more
descriptive commit messages and public discussions about the code design to
accompany the source code comments. Since this would involve a change in
everyone's mindset and decades-old workflows, it is really not a surprise
that nagging doesn't get anywhere.

Giacomo

On Mon, Jul 30, 2018 at 12:56 PM John Stone <johns_at_ks.uiuc.edu> wrote:

> Just FYI,
> Our experience with git on the NAMD side has not been without issues.
> In fact, we have had as much or more difficulity training people to use
> Gerrit than we ever had with having people send us diffs via email.
> The fact is that working on large scientific software is hard work,
> and one is either prepared to do that work, or not. Whether the code
> is in git, mercurial, svn, cvs, or whatever, hardly makes any practical
> difference. I've used them all and most of the issues come down to the
> quality of the code contributions if/when they occur, and the diligence
> of the contributor in testing their code. I would say that something like
> 99% of my time is spent on the latter issue. If someone can't deal with
> diffs, then they may not be willing to properly test their code either.
>
> The revision control system in use is probably the least important aspect
> of software development at this level. Our experience with NAMD and
> Gerritt
> is causing things to move in a new direction there, and when we have
> come to what we think is a better solution on the NAMD side, we'll see
> about
> moving VMD to match. So far the git/Gerrit attempt over the past year+
> has not been a panacea. Several of the key NAMD contributors were
> completely
> blown away by the complexity of working with the code review system, and
> this
> ended up creating more problems for us than it has solved thus far. Until
> things are working well with git on the NAMD side, I don't expect to change
> anything on the VMD side, it would be wasted effort until we feel that
> there's a net positive outcome.
>
> Best,
> John
>
> On Mon, Jul 30, 2018 at 12:33:01PM -0400, Ivan Gregoretti wrote:
> > Hi Giacomo.
> >
> > I think that migrating to Git is a good idea. It feels like the right
> > thing to do in 2018.
> >
> > Git might be more attractive than CVS to the younger generation of
> > users and contributors.
> >
> > I am thinking of student interns as an example. Blessed be all of
> > them. They have great ideas and they speak Git nowadays.
> >
> > What do people think?
> >
> > Thank you.
> >
> > Ivan
> > PS: VMD is jealous because only NAMD got a Gerrit privilege. :)
> >
> >
> >
> >
> >
> >
> > ######################################################
> > On Mon, Jul 30, 2018 at 9:40 AM, Giacomo Fiorin
> > <giacomo.fiorin_at_gmail.com> wrote:
> > > This is not a reply to the previous message, whose author didn't ask
> for
> > > one.
> > >
> > > For other readers of the list, NAMD has a Gerrit portal:
> > > https://charm.cs.illinois.edu/gerrit/#/q/project:namd
> > >
> > > The code base of VMD still uses CVS, but this is easy to track it with
> a
> > > private Git repository, from which John integrates changes.
> > >
> > >
> > > On Sun, Jul 29, 2018 at 3:19 PM Vlastimil Zíma <
> zima_at_karlov.mff.cuni.cz>
> > > wrote:
> > >>
> > >> Hi everybody,
> > >>
> > >> as I'm finishing my PhD and heading out of world of molecular
> dynamics, I
> > >> will most likely no longer use VMD.
> > >>
> > >> I've tried several times to provide patches to VMD, but since it's so
> > >> cumbersome, I'll put my effort into some other project. As I work as a
> > >> programmer for several years now and also participate in open source
> > >> projects I can only recommend to open the VMD source code to more
> > >> developers. Using CVS and having patches send over email is a terrible
> > >> approach when there are services such as GitHub, GitLab, BitBucket and
> > >> others and modern version control systems.
> > >>
> > >> All my molecular dynamics related projects are now open for new
> > >> maintainers, as I most likely will no longer support them:
> > >> * https://github.com/ziima/pyvmd
> > >> * https://github.com/ziima/vmd-debian
> > >> * https://github.com/ziima/namd-debian
> > >> Well, all my projects at github are open for contributions, you can
> check
> > >> any of those at https://github.com/ziima/
> > >>
> > >> Best wishes
> > >> Vlastimil Zíma
> > >
> > >
> > >
> > > --
> > > Giacomo Fiorin
> > > Associate Professor of Research, Temple University, Philadelphia, PA
> > > Contractor, National Institutes of Health, Bethesda, MD
> > > http://goo.gl/Q3TBQU
> > > https://github.com/giacomofiorin
>
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> http://www.ks.uiuc.edu/~johns/ Phone: 217-244-3349
> http://www.ks.uiuc.edu/Research/vmd/
>

-- 
Giacomo Fiorin
Associate Professor of Research, Temple University, Philadelphia, PA
Contractor, National Institutes of Health, Bethesda, MD
http://goo.gl/Q3TBQU
https://github.com/giacomofiorin