Abhishek Tiwari
Science and Web 2.0: Talking About Science vs. Doing Science - http://scholarlykitchen.sspnet.org/2010...
In fact too many talkers and evangelists! - Abhishek Tiwari from Bookmarklet
This hits me right where it hurts! All I do is talk about these ideas, never put them into practice. Crotty, you bastard, I'll get you for this. :-) - Bill Hooker
Well the OS glass will never be full - it isn't even half-full but there is water - lets keep adding more drops as opportunities arise http://en.wikipedia.org/wiki... - Jean-Claude Bradley
I am conflicted. After all I did the complaining, but not sure I agree with everything in there :) - Deepak Singh
I am very curious on how "doing" is defined - Deepak Singh
And I'm looking at all the cool stuff from Galaxy, and all kinds of collaborative apps that may not quite be public - Deepak Singh
If "doing" means randomly selecting and reading scientific articles then this is easily done. If doing means listening to my fellow friends and scientists in various networks for allowing me to rank scientific articles, then this is better. In this case "doing" requires communication, one form is talking, and the major goal is increasing knowledge by reducing information overload. BTW, on the contrary, do we not already have enough scientists doing science without communicating at all? Sure, they might publish, but do they really talk or communicate with each other? Think about it: Knowledge=People+Information. - joergkurtwegner
"Tools for Work: Every journal is looking for a leg up on the competition, looking for offerings that make them more attractive than other journals. Instead of offering yet another suite of communication tools likely to be ignored, we need to instead focus on the priorities and needs of our readers. Can we create new resources that support communities or that aggregate information in valuable ways? Can we open up our journals and let others tinker with our content and data to create something new (think along the lines of the Guardian’s open API)? Can we create new efficiencies for scientists, ways to make them more productive rather than tools that ask them to take time away from their research? Can we develop tools for doing science, rather than tools for talking about science?" - Daniel Mietchen
Neil - if the question is "are Web2.0 tools good enough to allow people to do Open Science" then the answer is yes - at least for applications like sharing solubility data. It might not work as well for applications with massive amounts of data - but we have to evaluate that on a case by case basis by at least trying with existing tools first. My concern is that people who might be interested in doing OS might get the impression that the tools aren't up to the challenge and not even try. Instead they'll think they have to wait for the killer "facebook for scientists" app to come out. Of the people I know who have tried radical sharing using Web2.0 tools and stopped - technology was not the problem - IP and collaborator issues were. These are much greater obstacles and have nothing to do with technology. For human-to-human collaboration I think existing tools are fine for OS. Where I see advances in technology becoming important is with automation, the semantic web, etc. - Jean-Claude Bradley
JC, that's where I disagree. If there's one area where modern web technologies do very well it's handling massive amounts of data. It's just that no one is building the tools or (based on what I've seen), the best ones are either commercial or internally focussed. I am talking about data managements, etc. Why is there no Flightcaster for genomic data or platforms like Twilio? - Deepak Singh
I'm not so sure, Deepak. If nothing else, transferring massive data sets is a blocker in many cases. Or do you mean web based interfaces/applications to massive datasets? - Rajarshi Guha
Deepak - with the docking work that Rajarshi did for our molecules the massive (relative to what I'm used to in organic chem) amount of data was hard to share on free hosted platforms. I was very hopeful about Google Research Data Service - we got some early accounts at SciFoo and Rajarshi did upload a large number of conformations from our virtual library. But that didn't pan out. We (Cameron Neylon, Tony Williams and I - were you there too D?) had a meeting with the Research Data people at SciFoo to try to convince them to do some data processing on Google servers since upload and download times were really a bottleneck. I'm feeling a little better with the ability to upload arbitrary files to Gmail accounts and we'll probably use that for some archiving but we can't run JSpecView from there so that kills what would have been a nice way to host JCAMP files. Of course Amazon could always donate some server spaces and services to OS D? :) - Jean-Claude Bradley
@JC, your last point is what makes me think that web tech is not necessarily a panacea for large datasets. The situations you describe are basically all ways to avoid transferring large datasets. Furthermore, they lead to some extent of "implicit" vendor lock-in. Not 'explicit', since you can pull the data down - but if it's so massive would you really want to pull it down? And of course there's the issue that code then needs to run on the storage system. - Rajarshi Guha
Rajarshi, I mean the latter. You should not be sending data sets to multiple locations, but shared dataspaces accessible by multiple people spinning up ad hoc compute resources or using APIs/services to pull in required data is the way to go (and being done today across any number of fields) - Deepak Singh
JC, Gmail isn't meant to front end compute. The design point for those storage systems is write few, read occasionally, so that's not surprising per se. - Deepak Singh
Jean-Claude makes a good point here: "IP and collaborator issues were." That is why people like John Wilbanks, Heather Joseph (both of whom are speaking at Science Commons Symposium – Pacific Northwest http://sciencecommons.org/events...) and Victoria Stodden (see her papers here http://www.stanford.edu/~vcs...) are so important. You guys all need to talk more to the Open Access people. Come to the symposium (and hear Jean-Claude and Cameron, too!). But I don't agree that the existing tools are good enough--we need to get more search and non-Os Web developers involved, too. - Hope Leman
Reading Jean-Claude's comments 5 posts up: Replace "Google" or "Amazon" with The Library sounds like the future to me. Libraries have the advantage of sharing the same mission as the open researcher (i.e., eliminates need to make it profitable for Amazon). Disadvantage is they don't kick as much ass yet (as far as I know). Have libraries across institutions begun collaborating on this problem yet? Hmmm...thinking about that, maybe all the libraries should pay Amazon or Google for a data hosting / processing service that is free to all researchers (both as contributors and consumers). - Steve Koch
Steve, well we made a small step forward with our library this weekend by having them host our notebook archive on DSpace. There is a chance they will agree to do more sophisticated hosting in the future- we'll see. - Jean-Claude Bradley
Great to hear, Jean-Claude. Maybe we're about a year behind you (no data hosting till at least the fall), but hopefully can catch up some day. - Steve Koch
Steve, as to your last point. Libraries cannot and will likely never be able to manage infrastructure at scale and really shouldn't even try. They should (and are) leveraging 3rd party infrastructures today. It's up to them to figure out the models around it. (example: http://duraspace.org/duraclo...) - Deepak Singh
Steve I would be curious to know if it is possible to export your whole notebook from OWW or is that code that needs to be written? - Jean-Claude Bradley
That code seems not to exist yet - I had posed Steve the same question by email, and here is what he replied yesterday: " OWW does provide data dumps. It is described here, http://openwetware.org/wiki... and supposedly the most recent data dump is here: http://openwetware.org/dumps/ Presumably you could write a script that would download the data dump daily to back it up? It doesn't include uploaded files, though. If you keep locally all uploaded files, when you upload them, that may be a sort of fix. Also, a problem is that you don't want to necessarily get everyone else's data daily to backup!" - - Daniel Mietchen
thanks Daniel - it would be nice if you could export your pages and data only - Jean-Claude Bradley
Bill (developer at OWW) could probably write these scripts relatively easily (and perhaps already has). I've been unable to reach him lately, though. - Steve Koch
Steve if you end up using lots of Google Spreadsheets to store data Andy has written some nifty code to make archiving them with the notebook fairly painless. - Jean-Claude Bradley
Thanks Jean-Claude -- I've been noticing that. Anthony uses google docs almost every day, so I've been thinking about it. There's talk of "OWW 2.0" (with that kind of thing integrated), so I've been holding off a bit. - Steve Koch
Chiming in late. Quick reply to Steve & JC's thoughts on libraries as partners for data archives/curation thru lens of UMich. Library reporting structure is thru Public Goods Council http://www.provost.umich.edu/publicg... , includes partnerships with institutional data repository. I did a couple blogposts on this with loads of links. Data plans: http://is.gd/jCiSet & data sharing http://is.gd/3FEirR Perhaps a model for other schools? - Patricia F. Anderson