[GSoC] Bio JavaScript/Node.js idea (bionode)

[Eric Talevich]

Folks,

There is a good reason Google appoints mentoring organizations rather than
accepting GSoC students directly. The primary goal of GSoC is to get young
people involved in open source software development with existing,
established projects. This allows students to learn how to develop in a
group context on a project that already has a significant user base -- an
environment that's almost impossible to create for university students in
their coursework, but an important one to experience.

Our first priority for evaluating project ideas is that they will create
this environment for students. This means an existing community of
developers who know the codebase and the programming languages involved.

For bionode.js in particular:

1. The convention for npm (the standard package manager for Node.js) is to
create smaller packages that do specific things well, and use the npm tool
to fetch and update the necessary dependencies to bundle with your
application. Bioinformatics is a huge subject domain. Wouldn't it be much
better for bionode.js to act more like BioGems and serve as a portal for
finding the right packages, rather than a single monolithic project? The
existing Bio* projects have already has grappled with disentangling a
monolithic codebase into more independent modules, while npm provides a
nice mechanism to avoid this problem from the start. BioHaskell also works
this way, through Cabal.

(See the bio-related Node packages currently available on npm:
https://www.npmjs.org/search?q=bio )

2. Have you at least contacted the BioJS folks? We need not just a mentor
but a community. Yannick is great and I trust his student Bruno is up to
the task as well, but it would be much more in the spirit of GSoC to
connect with a larger group that is specifically focused on using
Javascript in Bioinformatics. Surely the BioJS crew has developed some core
object types that they reuse for visualizations, and worked through some of
the rough patches already. (Wouldn't it make sense to stream the same
objects from the server to the client over JSON as they're used for
visualization?)

3. To be clear, I love the idea of expanding OBF beyond the old guard of
Bio* projects, but launching on a brand new project in a clean repo,
working with just two other devs and no existing user base, is not the
doing GSoC justice -- GitHub already provides that experience to everyone.
A year from now, bionode.js will almost certainly be a thriving project
with a year's worth of experience to fall back on and will be a good
educational experience for OBF students in GSoC 2015.

My suggestions for this project idea are to first contact the BioJS folks,
if you haven't already, to see if they want to take on a Node-based project
-- if they can offer JS expertise and a larger, fully supportive community,
students deserve that. Then (whether or not BioJS is interested in
collaborating on this), can you try to split this idea into smaller
components that are more npm-friendly? For example, a very lightweight core
package for DNA sequences; another one or more packages for sequence IO in
different formats, emitting sequence objects; another for microarray-like
data; another for protein structures; etc. -- whatever specific topics you
feel comfortable mentoring and supervising.

Best regards,
Eric


On Mon, Mar 3, 2014 at 9:27 AM, Fields, Christopher J <cjfields at illinois.edu
> wrote:

> Pjotr, Peter,
>
> In the end I think this has to be a decision that Eric and Raoul make, as
> they are running the show.  We wouldn't have any GSoC 2014 w/o their hard
> work.  My feeling is, in the end, openly discussing these will suss out the
> ones that seem to fit best overall.
>
> My 2c, if it's worth anything: I tend to agree with Pjotr, in that we
> should not limit ourselves to the various Bio*.  If anything, what I
> personally would like to see at the end of the day is a very good set of
> projects overall that we can pick from.
>
> However, I also think a balance needs to be struck, and that whatever we
> accept has to (in the end) benefit OBF.  That, to me, is the purpose of
> GSoC.  If it means we can get a new Bio* started up or we can (even
> tangentially) add functionality to one of the Bio* projects, all the
> better.  But I do think this has to feed back to OBF somehow.
>
> The reality is, we simply can't take every project from other related but
> non-accepted orgs.  In some instances, other accepted orgs might simply be
> a better fit, such as BioJS.  It doesn't mean we won't accept
> Javascript-based projects, but all alternatives need to be explored
> (particularly when there are a finite # of slots for a lot of projects).
>
> chris
>
>
> On Mar 3, 2014, at 5:56 AM, Pjotr Prins <pjotr2014 at thebird.nl> wrote:
>
> > BioJS is browser oriented. It may or may not fit there. See also
> > earlier E-mails on this ML by Chris and myself (Feb 27th).
> >
> > Even so, can we please be open to new project ideas and get rid of the
> > discriminatory programming language notion?  The bioinformatics
> > community is growing rapidly. Way faster than before. We should cater
> > for community needs as is represented by project ideas coming out of
> > the community.
> >
> > I think, as long as the project is about bioinformatics and FOSS, it
> > will fit GSoC/OBF.  We should only filter on the quality of students
> > and mentors.
> >
> > Note also that Yannick is a respected member of BioRuby.
> >
> > Pj.
> >
> > On Mon, Mar 03, 2014 at 11:39:04AM +0000, Peter Cock wrote:
> >> Devil's advocate: Would this not be a better match to BioJavaScript
> >> which was accepted as a GSoC organisation in its own right? What
> >> do they think?
> >>
> >> http://www.google-melange.com/gsoc/org2/google/gsoc2014/biojs
> >>
> >> Regards,
> >>
> >> Peter
> > _______________________________________________
> > GSoC mailing list
> > GSoC at lists.open-bio.org
> > http://lists.open-bio.org/mailman/listinfo/gsoc
>
>
> _______________________________________________
> GSoC mailing list
> GSoC at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/gsoc
>