The exciting news to me is that the "public" is interested in this. The article is being, clipped, dug, etc. - Nice Fish Films
Most excellent -- one quibble, I'd like to see "no insider information" credited to Jean-Claude. - Bill Hooker
well thanks Bill - the author removed it after some reformatting of the original paragraph :) I didn't ask for it to be removed - I have no problem with that paragraph - Jean-Claude Bradley
so you think that private company will always permits to their scientist to share data with the rest of the world? - Piero Giacomelli
@Piero: some will, some won't. The hypothesis is that the sharing model will prove more efficient, so early adopters will realize an advantage and eventually the mainstream will adopt their methods. There are no guarantees but I think it's worth testing the hypothesis. - Bill Hooker
Depends on the information. Pharma companies are already sharing pre-competitive information. Stuff that they would need to repeat in the absence of any data availability. With things like GWAS, it might be impossible to do for a company, so they will collaborate with universities and make the raw data public (e.g. Novartis) - Deepak
Expression data (eQTL) too. See Rosetta/Merck. - Chris Cotsapas
@Chris, oops. Should have included that one esp cause I have used that in talks :) (I was at Rosetta Biosoftware) - Deepak
Dorothea: not sure what type of answer you're looking for. Are you asking what the quantum computers are likely to be good for? Or something else? - Michael Nielsen
Partly that, partly "how the hell do you PROGRAM a thing like that?" - Dorothea Salo
For the example in the essay (qubits stored in atoms that are suspended in a trap), you manipulate the information in the qubits by shining lasers on the atoms, doing what are called "quantum gates". A quantum program is made up of lots and lots of quantum gates. Hope that helps(?) - Michael Nielsen
I got that bit. What I don't get is how a programmer conceptually sets up a problem and wraps his/her mind around how to twiddle the qubits to solve it. Something tells me the old saw about building an OS by flipping switches in octal and getting it right the first time isn't going to apply... - Dorothea Salo
The honest answer is "with a lot of difficulty"; only a few people in the world have really come up with genuinely new quantum algorithms, although lots of others have contributed extensions to existing algorithms. Because we don't have full-scale working devices, everything is done mathematically. Our ordinary intuition doesn't work so well, which complicates matters further... - Michael Nielsen
... (cont) towards the end of the essay I wrote down a series of equations showing the transformation done when a gate is applied. Believe it or not, but the way things work is that people have to figure out how to compose those transformations into something useful! This is, obviously, not easy... - Michael Nielsen
... (cont) The situation isn't quite as difficult as that, though. There are a lot of shortcuts and trick that make things easier. People have spent a great deal of time finding such shortcuts. It's worth finding even very minor shortcuts, because it can lessen the overall difficulty... - Michael Nielsen
... (cont) one thing that helps, I think, is to keep in mind that even experts find this very difficult. It's a bit like trying to solve the Rubik's cube. At first it just seems hopeless. But then you start to build up some tricks that simplify your life. You may never solve the cube, but things do get easier, and you start to build up an understanding! - Michael Nielsen
Thank you! Your patience is much appreciated. - Dorothea Salo
Amund - yeah, that would be possible, as a followon to this essay. But any such essay would need to be several times as long as this one, which I'm not sure I'm up for right at the moment. - Michael Nielsen
awwwwwww! *makes great big puppy-dog eyes* - Dorothea Salo
Amund - Scott Aaronson has a nice explanation at http://www.scottaaronson.com/b... It's at a somewhat higher level, and uses a somewhat different strategy than I'd take, but you might find it helpful. - Michael Nielsen
Dorothea - :-) I must say, that's not the usual way people respond to explanations of how quantum algorithms work :-) - Michael Nielsen
I am nothing if not a maverick. :) Seriously, though, I learned something from you today, and I wouldn't mind repeating the experience. - Dorothea Salo
Really good post Michael :) I'm studying computer science and I've been always interested in quantum computing, your post is really explanatory :) - Carlos Mestre
One thing that is annoying about all quantum computing story is that nothing changed from the time i became interested in qc (7 years ago or so). I assume a lot changed, but from public perspective (non-scientist), there hasn't been a great breaktrough like a largescale qc. The latest buzz was about that SV based startup which claimed they made a quantum computer. And, your book is on my read-some-day shelf. Keep up the good work. - Mladen Srdić
Mladen - There have been a few interesting developments, especially the discovery of measurement-based quantum computing, and a series of breakthroughs related to simulating quantum systems classically. None of this has made the popular press... - Michael Nielsen
... the overlap between breakthoughs and what makes the popular press is quite small. Breakthroughs often don't make the popular press until years or decades later; a lot of what makes the popular press isn't a breakthrough, even when it's sold as such (the company you mention is an example). Partially, that's because it can be hard to recognize breakthoughs as such when they happen. This was true for both the examples I gave above. - Michael Nielsen
Victoria's talk at SciFoo, about licensing and open science. Victoria has also done some interesting work on reproducible research in computational science. She's on FriendFeed - http://friendfeed.com/vcs - Michael Nielsen
Actually it would be excellent to have a discussion about the license here. Would be interested to get others' perspective on this. As I understand it the idea is that it wraps up a series of existing licenses (CC-BY, BSD) and while the license isn't viral, attribution is. In the session there was a discussion about technical attribution stacking that this might lead to but having moved on from data licensing (I hope) this is an interesting concept. - Cameron Neylon
I must admit, from a scientists' point of view, I don't see why everything shouldn't just go into the public domain when published, or something close. I can understand why the publishers don't want that, but I don't think their concerns should be primary, here. In any case, a lot of published material is public domain - anything coming from the US Government, for example, by law. - Michael Nielsen
Continuing on, because copyright and patent law was designed for commercial use, not use in scientific research, it addresses a completely separate set of concerns. Those concerns are addressed normatively in science - when someone copies your work, you don't sue them (what money would you recover?), you instead publicise it, and ruin their reputation. I guess that's why I think the need for licenses is limited. - Michael Nielsen
Cameron - I'd love to have such a discussion. Just to clarify, when you talk about "technical attribution stacking", you're referring to the fact that as stuff gets passed around, more and more people may become "authors", at least in some machine-readable sense? - Michael Nielsen
The attribution stacking problem is, to my understanding, why the FSF recommends not using the original BSD license. The list of "authors" got absurd quickly, causing problems with the prominent display clause. - Christopher Granade
Seconding Michael's view of public domain, at least for government-funded research. This would still leave a role for OA publishers, who can still charge money for co-ordinating peer review and providing a stamp of authority as well as whatever value-add they come up with (nice layout, metadata markup, submission to PMC, etc). Even private research should have no objection to their published papers going into the public domain; anything they want to keep closed they don't publish anyway. - Bill Hooker
@Christopher - my understanding is that with Victoria's license, the attribution would be handled in a machine-readable fashion. I think we're going to have to move to that anyway, eventually - as science becomes more open, contributions will naturally come from larger and larger groups of people. - Michael Nielsen
Oh I'd love everything to go into the PD. And (in my usual homage to John Wilbanks who says all of this much better than I do) I think it is the only long term practical route forward. But I think there is a place for comforting people a bit along the way, and also getting advice on what will work both socially and legally. - Cameron Neylon
Michael - sorry not clear, I meant the 'technical problem of attribution stacking' what Christopher describes well, as opposed to the 'social problem' of 'but I'm not citing that waste of space!'. Whether it is via licensing, legal positions, or by social norms, what should we actually cite and should that citation train track back through different objects? And if so how? - Cameron Neylon
Cameron: there's a sizeable fraction of the current literature that is already public domain. Anything produced by a US govt employee, for example. It certainly doesn't seem to hurt those people. As I said above, I think the main reason for the licenses is to protect the publishers, not the scientists, at least in basic research. - Michael Nielsen
I still think there is a different between data, derived data and content. "data" especially anything publicly funded should just be public domain. Derived data is the tricky part, since that's where there could be commercial value, so you need some licensing options. Content is where publishers come in. IMO, they do not own any rights to the data, so should have limited to little say in the matter - Deepak
The US govt case is a really interesting one of just busting the system down to the ground and then making it work really well. It is a wonderful example of what can be done. But can we actually make it happen in other areas? - Cameron Neylon
Actually I think the core of the problem in IP/content and everything is that the whole underlying legal framework is badly broken. Patents are causing more problems than benefits. But I don't have any good ideas about how you replace them with something the provides a good 'right to exploit' without unnecessarily blocking the option of others to innovate on the same platform. - Cameron Neylon
On the attribution stacking issue, I think it was Myles Axton who said that if he didn't get a solution to the apparently simple problem of how to represent citations in journal articles then his journal was going to run into serious format problems. He's running into figure legends that are twice as big as the figure (I think - I can't remember the details precisely) because of citations to datasets - Cameron Neylon
Cameron - I find it very useful to analyze two cases separately: research that may be commercializable, and research that is almost certainly not. In the latter case, I think copyrights and patents etc are just plain irrelevant, at least to the scientists themselves, and that a switch to pub domain could be made en masse. In this realm, it's all ruled by cultural norms, not by law. In the commercial realm, I agree that there's an issue. - Michael Nielsen
Michael just said what I was trying to a lot more clearly :) - Deepak
There are very real legal concerns for scientists releasing their work on the web, as noted by Michael Nielsen, but I agree this isn't the main barrier we face to freely revealing our work. But the license idea rescinds the strictures of copyright, which as scientists we normatively get around by not suing each other. But the concerns are real and should be addressed in my opinion. The nice aspect to addressing these legal concerns is it brings attention to the nonreproducibility of much of comp. science. - victoria stodden
On attribution stacking: the big particle physics and astronomy collaborations routinely have hundreds of participants, and sometimes thousands. They're starting to employ shorthands, essentially inserting not author lists, but rather pointers to author lists. This is still in flux, but it's an interesting bellwhether. - Michael Nielsen
In that sense norms about what we consider appropriate to communicate openly (right now, essentially just the paper) to encompass our entire research compendia - code, data, paper, figures etc. Encouraging a license that rescinds copyright on all aspects of our work also encourages reproducible research. - victoria stodden
Having work in the PD is a great goal, but i agree we get there by steps. In talking to scientists about barriers they feel to releasing all their research work on the web, a common one is fear of work being stolen. So leaving an attribution aspect attached to work you release - that remains attached even as your work is incorporated into others' research - is a mechanism to assuage concerns about attribution. And the way this dovetails with machine readability, as noted by Michael above, is great. - victoria stodden
In today's production level science, does attribution stacking make sense. it is one thing to come up with an original idea or a new method, but quite another to sequence a genome. There are always a few key people, but it's a team effort and the whole authorship model breaks down. The only reason it exists is cause we have a publish or perish model. That's like saying everyone involved with bringing a drug to market should be on the patent or some other doc describing the drug - Deepak
You know you've got a good conversation when you have to page back up for the comment button (hey guys when do we get one at the bottom as well!). I agree attribution stacking is to be expected and supported. I think we need to have that discussion about 'what is appropriate' and perhaps bit a bit more critical ourselves about what we cite and why. Is there a place for the 'semantic citation' and how important is granularity of that citation? Apparently ecologists still routinely cite page numbers - Cameron Neylon
I think the key question though is how to combat this issue that Victoria pins down as 'fear of stuff being stolen'. Many of us simply take the view that it wasn't ours to start with but we have to accept that that isn't general. Do we change the view socially or provide a little comfort for people to ease them into a new way of thinking. Or more accurately what is the right mix? Do people think there is an actual need for enforcing citation? - Cameron Neylon
I agree, Cameron. Citation in research today has very strong enforcement through norms and it seems plausible that this can extend to the web. So the license can help with this by clearly establishing that these norms also apply to work on the web (where there is a convenient time stamp), as well encouraging openness on the web. The license also corrects the real legal concern with copyright on research - and as we saw in Jacobsen can provide legal remedy via injuctio. http://www.cafc.uscourts.gov/o... - victoria stodden
Cameron - Neither did I :-( We definitely need more community building! - Michael Nielsen
I'm going to try and collect together email addresses I think and send around a circular. We desperately need to get all these people onto the same wavelength. Part of the motivation behind yesterday's blog post as well. - Cameron Neylon
@Deepak - well in this case it was the social discovery mechanisms via Thomas that got us to this point - Cameron Neylon
I'm certainly finding social mechanisms providing more interesting content past my front door. But its not comprehensive and somehow the natural time frame of e.g. friendfeed doesn't work well with how I find things when I need them. - Cameron Neylon
Cameron, I've always felt that we need a combo. Search for on-demand retrieval, perhaps filtered by trust/social mechanisms, and social mechanisms for more real-time discovery - Deepak
Agreed - I just haven't yet figured out how to make that work for me in real time yet. I would like to get more social into my on demand retrieval I think. And a bit more built in search and filtering in the real time stuff. - Cameron Neylon
Seriously, try Lijit. It's built on top of Google, but looks at your delicious network etc to try and get more relevant hits. - Deepak
"Early stage venture firm Y Combinator, which has funded over a 102 young startups, has “open sourced” the legal documents that they provide to their startups to use as they seek additional funding. The documents were created with their law firm, Wilson Sonsini Goodrich & Rosati" - Paul Buchheit via Bookmarklet
Paul Graham is doing a terrific job. FEw days ago, for thr first time, they published a document showing what kind of startup ideas they particularly look for, and now this!!! Cool. - Hayk Hakobyan
very cool. You can imagine doing that with scientific data. plots and everything , embed and write discussion around it. I will be out of job soon enough ;) - Pedro Beltrao
Remote scripting API for easy embedding in other scientific websites - Todd Harris via twhirl
link user's reading habits with scientific expertise and social graph to evaluate and sort articles or other scientific info. - Pedro Beltrao
A clear use case for why I should participate in a vertical social network. Most don't add enough value or are too fuzzyily described to interest me. - Deepak
@Pierre - well it's true isn't it? I think Deepak summed it up well - there needs to be a compelling reason to use a vertical social network, and the best reason usually isn't features, it's the right users. It's quite possible to build a useful social network without ANY features (eg twitter), but no number of features can save a social network without users. - Nick Lothian
I want it to learn from the success of FriendFeed - Neil Saunders
Nick's comment is right on track. I think a user base is the most important part of a social science network - but the features are the reason those people join (or at least the early adopters join) in the first place. - Andreas Matern
This will get fairly high visibility once BL puts it in more open beta. "The Research Information Centre. In close partnership with the British Library, this collaborative workspace will be hosted via Microsoft Office SharePoint Server 2007 and will allow researchers to collaborate throughout the entire research project workflow, from seeking research funding to searching and collecting information, as well as managing data, papers and other research objects throughout the research process." - Richard Akerman via Bookmarklet
It's not really clear, their article http://ieeexplore.ieee.org/iel... talks more about the different things it can link out to, not about how to integrate it with other networks or link in to it. - Richard Akerman
We did some beta testing of this - it's really ambitious. Doesn't work on Macs or *nix, though, needs ActiveX. - Euan
Yes it is really ambitious, possibly too much so. My understanding of it from a presentation by Roger Barga was it would capture the whole scientific process from ideas, grant writing, collaboration, experimentation through to publication. - Duncan Hull
I saw a talk on this earlier this year atBioIT World and the part that concerned me was that the folks at Microsoft, with a few exceptions, seemed somewhat confused about what the real challenges were - Deepak
+Deepak. Somehow seems too fantastic to get a revolutionary end-to-end solution in one shot, without involving evolutionary pressures of the real-life user community. Especially given ill fames of the ActiveX. - Yaroslav Nikolaev
Wonderful!!! I should add that it's also clear that the open science discussion has really matured since last year - Deepak
I would have loved to be there for just one reason ... to debate Chris Anderson :) - Deepak
You would have had to get in the queue :) - Cameron Neylon
lol ... maybe I would have broken line only to find I had the wrong Chris Anderson (the TED one) - Deepak
Could you amplify what is meant by "diversifying sources of credit" (first bullet point)? Are you speaking about broadening formats deemed acceptable for purposes of tenure? Or is it something else? - Jill O'Neill
I will write on this at some length soon (so hold me to that) but broadly speaking I mean that the discussion we often have is about 'why can't people get credit for blog posts' (or data, or twitter streams, tool development, or podcasts or whatever). The point I wanted to make was that we can't simply say 'give credit to people for developing tools'. We need to say 'we need more tool developers in tenured positions'. - Cameron Neylon
A report with far reaching consequences and a clear eye into the future. Thanks!! - Aarthy
This is something that should be posted every two weeks. We tend to focus on things that make us more efficient in the current system, instead of questioning its foundation (publication pressure) and seeking long term solutions. - Pawel Szczesny
One person is not as smart or smarter than all of us. The western ideal of the lone genius (be that one person or one organization) is nothing more than a romantic notion one that works to support non-disclosure. Your suggestion has merit in that data disclosure would potentially encourage deeper peer review, more comprehensive analysis and richer understanding. Kudos on presenting a compelling and cogent argument. The future is not proprietary. - Dave Martin
Science is, as such, generally not proprietary, at least not the life sciences. However, we do have a tendency to be protectionist about our data and as many remind me, it's not always because of choice, but because people need to work within the confines of a publish or perish academic research system. There's enough people who want things to be different, many active in academic science (unlike me). It's easy for me to write down something. The change will have to come from within, and it will - Deepak
One problem is that in academia, we train young scientists to believe that they own the data that they generate. This explains, amongst other things, resistance to open notebooks. The notion of "my personal notebook" is deeply entrenched. We need a mental shift away from "I generated this and it's mine" to "I uncovered what was there already and now it's everyone's". - Neil Saunders
I suppose we need a balance between the two. We need the humility to realize that what we are doing is a lot bigger than any individual, and the pride in our work to have a sense of ownership and desire to make it as good as it can be. Somewhere along the way, we seem to have lost some of those values - Deepak
That's the same Bill Joy that proposed a moratorium on Molecular Biology and Nanotechnology studies because they would end up creating tiny replicating machine-virus? - PauloNuin
Yep, same fellow. I believe I once called him a luddite. That said, he has his moments. Also founded Sun along the way - Deepak
I remember a lunatic, who was at the same time a professor at the Mathematics Institute of the University of Sao Paulo, almost got into a fight with some biologists based on Joy's article. - PauloNuin
re Neil Saunders - sure it's everyones, just after the publication gets accepted :) - Rajarshi Guha
I wouldn't spend too much time making appeals to lofty high ideals, it boils down to the incentive system. The instant NIH grant admins take into consideration the number of data sources you've made publicly available when making awards, pretty much overnight every bit of data academics can find will be made available. Write your congressmen, get the NIH to require it. - Daniel Kluesing
My argument is that the grant system is part of the problem and needs change. That kind of approach (looking at number of sources) is completely bogus. - Deepak
Let's start with some definitions: what is a "network", what is a "collaborative tool", who are "scientists" and what does it mean "social"? What current sites, services and pieces of software count as "social networking tools for scientists"? Just those emulating Facebook? Databases? Journals? - Bora Zivkovic
All good questions, maybe we can have a first hack at them tomorrow at BBC/online here? - Cameron Neylon
I was thinking about writing a blog post about it, but had no time this week and will be in NYC while you guys are at SciFoo, so perhaps we can chat here and then post later on our blogs? - Bora Zivkovic
Great points by Bora - we don't even think about how these things are defined, we just parrot the buzzwords - Neil Saunders
Just don't say "social media" for scientists, or you'll be able to hear me scream from Seattle - Deepak
This is a central problem of all sites that implement social features, not just science sites, most of which are hopelessly web 0.9. Hence the rise of things like Facebook Connect and Google Friends. The best that we can hope for is that developers employ things like openID and better yet shared (professional) profiles like from epernicus.com. - Todd Harris
The contrary view would be that there are already significant efforts underway in standardizing this, and effort would be better spent in writing software and/or building communities - Nick Lothian
I think centralisation and decentralisation are inevitable ebbs and flows of computer systems. For a network to hit critical mass it needs to either manage to be a strange attractor for an interesting group, or offer enough new functionality (like FriendFeed) to bring people in who are already doing things elsewhere. - Richard Akerman
@Neil this is why I thought focussing on the existing sites might be useful rather than taking some grant philosophical position. @Bora I think chat here and/or on Google Doc and then update as we go around. Question as to whether it is worth putting this up as a session at BBC/SciFoo as well? - Cameron Neylon
I second what Nick said. The open social graph will happen, science networks really shouldn't be doing their own thing too... OTOH microformats and FOAF being available along with hashed email addresses would help. - Euan
@Euan @Nick I agree absolutely, that's why I'm pushing for adopting more widely used platforms, while looking at useful features and what people actually use. - Cameron Neylon
I think we should start supporting the standards and APIs that already exist, mainly the standard features that every social networking site needs. Those that don't will probably be left behind. - Martin Fenner
@Cameron - what do you mean by "adopting more widely used platforms"? Platforms as in standards? Or platforms as in software products? I agree with adopting standards, but not about software products (I'm somewhat biased, as I develop a product in this space). I don't think the market is ready for product consolidation yet - there are still too many new features being created, which makes it hard to pick winners. Plus, there's the whole hosted offering (Ning etc) vs installable software debate to consider. - Nick Lothian
@Nick good questions. I didn't mean software products because I don't think the right ones exist, but standards certainly. As for installable versus hosted my personal view is that something along the Wordpress model is a good route. There will be plenty of people with the skills and wish to deploy locally and plenty without the skills or resources to do it. In the end though I don't see this as a single product but as a platform. Essentially that means a data standard and tranport mechanisms of some sort. - Cameron Neylon
@Cameron Well that's easy then. HTTP+Atom+XFN+The OpenSocial REST APIs and datamodel+OpenID+OAuth - Nick Lothian
In terms of open standards/transports, it might make sense to also discuss XMPP Publish-Subscribe/Jabber (http://www.xmpp.org/). There's been some recent discussion on how it may be a better fit for some social applications than HTTP/ATOM. I found this presentation instructive: http://www.slideshare.net/rabb... - Fitzgerald Steele
Igor, I also really like it. It was already known in 90s that people using mouse better memorize angles than distances (therefore it's much easier to learn menu options for radial menu than standard one) - and it looks like it's the first major approach to implement that discovery. - Pawel Szczesny
I like some of the ideas, but the presentation was kinda odd. What was with the low-frame video? Also, that hand cursor is really freakin' ugly. I also get the feeling that many/most "normal" users would be completely overwhelmed by that interface. But I do like the radial menus and the idea that the browser has no button bar or visual menus. Oh, and that weird wheel/mouse thing she was using? Looks really complicated and expensive. - Tad - just Tad
@Tad "...complicated and expensive" Looks like that concept will be transformed in a future - Igor Poltavskiy
Not sure what you mean Igor - it just didn't look simple to manufacture and looked like it had a lot of moving parts. Modern mice are pretty simple devices and I think that's one reason we use them instead of complicated chording devices... - Tad - just Tad
Tad, I'm not sure but it resembled modern haptic device people use for example in molecular research. It's getting pretty inexpensive nowadays. - Pawel Szczesny
Looks awesome! And with UI - think it all goes to the fusion of display+sensor.. - Yaroslav Nikolaev
I don't know what computer they use, but when you look at the caroussel she's got a lot of opened windows & applications. The look is great, but as far as I tested graphical navigation, it's not so efficient when compared to menu & tabs & shortcurts - Olivier
Nice concept video. I think there are a few flaws in it though. For one, while the animations on the radial menus look nice, they decrease performance. Imagine having to wait for the menu to unfold every time you want to use it. Pawel, have a look at marking menus (http://www.billbuxton.com/MMUs...) that allow users to seamlessly switch between novice and expert modes. - Jo Vermeulen
Kinda reminds me of Gibson's world of cyberspace but in its infancy. - zoblue (Zulema)
@zoblue First:Neuromancer by Gipson,second:concept by Adaptive path&Mozilla:) - Igor Poltavskiy
Jo, to me it resembled all these animations on Vista :) I expect one could turn that off. Thanks for the link - I didn't know that research on marking menus were already back then. Olivier, thanks a lot. I found one that works with FF3 - Pawel Szczesny
I've always thought that the greatest waste of taxpayer research money was in competing projects where two groups compete on the same question, instead of collaborating. Open science would prevent/reduce competition, wouldn't it? Thus, open science saves the taxpayer money! :-) - Björn Brembs
I don't think open science will reduce competition per se. What it would hopefully do is make the competition more effective and efficient. Its not enough just to do something faster, you've got to do it faster and better and smarter. What it should prevent is fruitless exact or near exact replication. - Cameron Neylon
ok - what I meant was competition in the sense that two groups are trying to find out the same thing, e.g. "doe
