Open Source/Open Science

Stephen Adler (the guy who always does really cool write-ups of various events) writes "A one day conference titled Open Source/Open Science will be held at Brookhaven National Laboratory on October 2nd, which will focus on the use of Open Source in science. This is a pre-announcement so that the organizing committee may be able to better gauge community interest in attending. The full text can be found on the conference web page. "

Yay

[jlee]

I think it's neat to see other fields such as Science becoming part of the Open Source arena. This shows that Open Source is not just for Computer Science.

Open Sources

[gargan]

The new book by O'Reilly, Open Sources describes well the relationship between science and the OSS movement.

Re:Yay

[mong]

I think it's really neat that Steve Adler has abandoned music to support the Open Source cause!

Guess he's off the heroin now...

Mong

* Paul Madley ...Student, Artist, Techie - Geek *

Re:Yay

[rde]

Guess he's off the heroin now...

Imagine what the world would be like if heroin were open source. You'd have the option of flying to some poppy field and making your own from the source, or you could buy it from a reseller who'd made his own modifications (say, adding speed).
Of course, you'd have the source, so you'd know exactly what you were getting...

It's vice versa

[stonehenge]

It has been pointed out correctly by ESR and others before that the principles that make the open source movement nowadays so successful, are the same that made modern science so successfull for the last hundred or so years.

BNL truly a leader in open source / open science

[ubiquitin]

I'm not affiliated with BNL, but merely someone who has been to Brookhaven National Lab with the past few months and seen some of the open source apps they have on the synchrotron floor. I highly recommend attendance of this one. Can't wait for Oct. 2. The macromolecular [bnl.gov] guys [bnl.gov] and many others at the NSLS [bnl.gov] have Linux boxes all over the place, folks, doing anything you can imagine. Can't wait to see an organized display of the OSOS stuff being worked on.

Gotta get front-row tickets for this event. Any scientists from the Boston area want to carpool down to BNL via the Long Island Ferry? It's on a Saturday but I can't think of a better reason to play hookie, personally. email me at: [e-mail] [mailto]

an "open lab"

[Bizzaro]

We all know that the open-source development model has been extended to several areas beyond computer programming. But have you heard of it being extended into the realm of scientific research?

"Bioinformatics" is branch of biology that is completely computer-based, and typically involves the analysis of data stored in public databases on the Internet. This is my field of study.

As you may know, scientists in all areas tend to work secretively on projects for many years before publication, often limiting access to discoveries and innovations to those who work in the very same room! And that is something that has always bothered me about science. It seems we are to share our ideas with humanity (which we often do, if and when we publish in a refereed journal), yet the priority is really to protect one's self from being "scooped". It appears to me that we (scientists) care more about getting credit for the discovery than the discovery itself. Does this remind anyone of the mindset of the proprietary software developer?

A thought I had was that bioinformatics, being computer-based research, may present an opportunity to make an "open laboratory" or "collaboratory". Following the model for open-source software development, an open lab would place on the Internet all of the work in progress for a particular research subject (analyses and/or new techniques). Communication between collaborators, which would include anyone who comes along and decides it is an interesting project, can take place via mailing lists, and so on.

The only worry, even still for me, is that someone can come along and swipe the whole project and take credit for it. But of course anyone who gives away software code would have the same concern. It really comes down to human nature: Just how much are we willing to share with others, at the risk that someone may take it all away?

But I think even temporary measures to prevent yourself from being scooped are restrictive and contrary to the scientific ideal. Why? Because these measures very much squelch free thought and exploration, what science is supposed to be all about.

The one thing I particularly enjoy about free software development is the ability to brainstorm with anyone anywhere over a new project, without fear of idea-theft. To me free software is not just about seeing the code or downloading free binaries--something even a closed organization can accommodate. What I like about GNU- or Linux-style development is the open exchange of ideas--the "bazaar" model Eric Raymond refers to.

My whole point is that it's something VERY MUCH MISSING in all research projects prior to publication--when the scientist feels there is nothing to lose and nothing left to be gained. I just want to see that change.

So I started The Open Lab, or "The Open Collaboratory of Molecular Bioinformatics," a non-profit organization established to promote open collaborations for research and free software development in the field of molecular bioinformatics.

http://theopenlab.uml.edu/ [uml.edu]

The Lab supports Internet-based collaborations that will advance knowledge in this scientific field. This means not only GNU-licensed software development but pre-publication research projects that are completely open for the public to view and contribute to.

Jeff

This sort of thing has cropped up before. And it has always been due to human error.

Maybe this would help me in Chemistry

[extrasolar]

I was looking on freshmeat last night and came upon some molecule editor and view. And on the home page there was mention of this open science thing. The really nice thing is that this program is in Java and I could, with the teacher's permission, install it on the computers in class and let the class know the values of open source (Okay, they probably won't care, still, I think the school has a football program to run ;). I could give the teacher the web site so he can keep up to date provide some feedback.

I know my beginning Chemistry teacher noted the lack of such a program and am not sure how much such a program would cost otherwise (especially for the dozen machines in the room).

Maybe I am being optimistic.

--

"Open Source" and Scientific Publications

[Thorsett]

One example of "open source" thinking in science is the increasing importance of preprints over journal publication. Traditional scientific communication relies on institutionally sponsored (commercial or nonprofit) journals publishing papers that have been carefully vetted by referees. But in many fields (including my own, astrophysics) this system has been almost completely supplanted in recent years by electronic distribution of unrefereed preprints through the server at http://xxx.lanl.gov/ [lanl.gov]. The questions raised by this shift should sound familiar to open source software advocates. For example, how can quality of product be maintained in a system without identifiable, responsible publishers? My own experience suggests that quality is enhanced by widespread distribution of results before publication, for exactly the same reason that distribution of source results in better code: lots of eyes looking for bugs (or neglected references, etc.). Paul Ginsparg has written thoughtfully about these issues, for example at http://xxx.lanl.gov/blurb/pg96unesco.html [lanl.gov].

Open Source is not new in science

[-=Hastur=-]

While I don't think it has ever been formally called "open source", there has long been a free exchange of software in the scientific community. In fact, there are quite a few parallels between the Open Source Movement and science:

1) In both, reputation is a valuable currency. In most sciences, your livelihood depends as much on citations as publications. Writing software to perform a valuable task and then distributing it to others means more citations in their papers, increasing the reputation of the software's original designer.

2) In both, individuals are suspicious of anything they can't verify themselves. For this reason the code of scientific software is often made available for users to tinker with as they see fit, and so they can be certain that the program is doing what they think it is doing.

In the free release of scientific software, I also see a few potholes which highlight the potential weaknesses of open source:

1) Even when science software source is freely released, typically much of the development has already been completed. If a scientist were to make a project open-source from the beginning, she sacrifices her most valuable potential gains--the reputation she is accorded when people use her software. No one gets grants for starting open source software movements in science!

2) Science relies on reproducibility, and heavily modifying the source of a popular program hampers this goal. If two scientists use the same software to analyze the same data, they should expect to get the same results. Poorly-organized open source software risks fragmentation into a myriad of different versions, which dilutes the value of the program and raises as many suspicions as open source was designed to allay.

I think the most successful open source software has learned these lessons as well: open source works best when there's a shepherd to guide the flock, and this shepherd (or shepherds) must maintain a strong handle on "official" and "unofficial" versions.

-=Hastur=-

How useful will it really be?

[scheme]

Right now I'm working with some people on the superk project and I wonder how useful the code I'm writing is to others. A lot of times the code written in High energy physics is used to interface with custom electronics. As such the code probably isn't very useful to others not working on the same project. For example, I'm writing is used to interface with either an $8k or $22k oscilliscope attached to some custom electronics. I'm not sure how useful the code would be to others without the same setup.

It isn't just credit.

[Brett Viren]

While it is true that much secrecy in science is in order to insure getting credit, I would say that it is more to do with not releaseing false information and making a fool of oneself. It is irresponsible for any scientist to publicize their work before it is finalized.

Publicizing unfinished science is like releasing pre-alpha software and calling it golden. That is, it is a lot like the marketting strategy of a certain wealthy software company.

Sometimes keeping the source or the science quietly closed is the Right Thing to do.

-Brett.

Re:It isn't just credit.

[fiori]

The one premise that has kept modern science respectable is peer-review for all work before it is published. Beyond the political problems inherant in any field, this prevents false and/or misleading information from being mistaken as fact. Every scientific scandal (e.g., polywater, cold fusion, base metal transmutation) has been the result of untested, unsubstantiated prelimary results being released as fact without peer review. It is important to remember than scientists were among the first to collaborate on projects via e-mail and later the web. Scientific projects do not occur in the vacuum of a single lab: there is most often communication of preliminary results to other labs around the world is the norm. In many ways the current development process of the Linux kernel, etc. is very similar to any scientific research project. The Brookhaven National labs established the Protein Data Bank [bnl.gov] in 1967. Every protein structure that has been determined, and published, has been place in the PDB for public access. This is the scientific version of open source.

Has been used over 2 years by ALife

[MoobY]

Artificial Life is the science of modelling biological phenomena in artificial media such as computers. Two years ago, on our biannual european conference in brighton, there was a discussion on how to conduct experiments on computers.

It was agreed, that if someone wants to redo an experiment, that the source code of the original experiment should be available, therefore open source.

Since we're mostly dealing with models of phenomena that are studied, it is pretty logical that these experiments have to go opensource. Other programs that can be used as regular lab programs, but are not needed to conduct one or another experiment, may go opensource, but are not needed by the scientific community as opensource packages.

Re:Maybe this would help me in Chemistry

[stonehenge]

What you saw were announcements of JMol and JChemPaint - a 3D viewer and an editor. The www.openscience.org site mentioned in the freshmeat postings is not associated with the Open Source/OpenScience meeting at BNL, as far as I know. After all, "inventing" the term Open Science after all that (well justified) Open Source hype seems to be quite obvious.

Re:It isn't just credit.

[Brandon S. Allbery]

Actually, early announcement is common. It has to be: that's part of how things get peer-reviewed. As soon as it's preannounced, labs all over the world set out to duplicate the results, in order to prove/disprove its existence and if possible find out more about what's really happening and why.

The problems occur when it's over-publicized (as with "cold fusion").

Re:It isn't just credit.

[Bizzaro]

The problem isn't the publication or peer-review process. These are needed to help verify results and ensure the researcher performed the work in an ethical ("right") manner. They are both extremely valuable.

The problem is secrecy. And I don't mean secrecy to keep one from making a fool of himself. It is entirely a matter of "intellectual property". People protect their ideas for just long enough to ensure that they are given credit for it. Fame, fortune and human nature; that's what it is all about.

And I think one of the best (or worse) examples is the concealment of scientific discoveries by pharmaceutical companies. (Note that I come from the biological sciences. This issue may not be so well known in physics.) They deliberately hide information to protect their bottom line, and that's it. This should be unacceptable to the scientific community.

And yet another problem is the non-disclosure of all data relevant to the results after publication. This is done so that the researcher has first dibs on finding new leads. We should not stand for it.

I do realize that science is the inspiration for the free software movement, and I suppose there are many more negative things to say about the software community than the scientific community (at least from the point of view of a non-scientist). But that does not mean there are no problems to resolve.

Jeff

This sort of thing has cropped up before. And it has always been due to human error.