Bill Hooker

Wiki-based, login required, invite-only, real names, custom access levels. User tracking as anti-scooping device. Joe: "I'm hoping that a few people will try the current beta version and let me know where it can be improved." - Bill Hooker

I created an account and copied one of the experiments from our wiki. It looks like it keeps track of versions ok and should be usable as a lab notebook. The one issue I have is that there is no mechanism to access the notebook pages without a login. This means that the notebook will not be indexed by search engines and will not be discovered by researchers who don't already know about your site. I don't think a notebook can truly be called Open if it is not freely indexable and < a href="http://drexel-coas-elearning.b...">accessible by a simple url</a>. Perhaps the term "Shared Notebook Science" would be more appropriate for what you are trying to accomplish. - Jean-Claude Bradley

Copied from comment on Blog - Interesting idea, although as you might expect, along with Jean-Claude I'm not convinced that closing things up is the most effective way to be open. Conceptually though explicitly recording who has seen what is still an interesting route to explore though and if this is a route into full ONS that people feel more comfortable with then that can't be a bad thing. I'm also really interested in your graph of idea propagation as it has a close connection to some things we are trying to represent in our blog based notebooks (see e.g. http://chemtools.chem.soton.ac... and http://chemtools.chem.soton.ac... for a different representation. We've been very interested in network representations of the connections in the system. I'm based at RAL near Didcot so if you want to drop me a line then I'd be interested in meeting up to talk about this. - Cameron Neylon

Don't seem to be able to get onto the system anyway - doesn't like my passwords or something - Cameron Neylon

Coming from OWW, I'm sort of biased but the closed fashion that this "open notebook" is using is not going to work. Not sure that tracking can prevent the risk of being scooped either. I happened to sign in without any delay. At OWW, all registrations are parsed manually and we try (we do try!) to verify that every user is authentic. Do you think a page displaying the list of users/IP's that visited a page would prevent getting scooped? - Ricardo Vidal

Glad to have generated some interest. Let me say that the problem with public access was a bug, not a feature, as was the ability to register without an invitation (I ticked the wrong box in the settings). All should be fixed now. I'm not sure that the tracking will prevent scooping, but I hope it will discourage it. This is simply an experiment to see if it does. As the beta progresses I want to add the option of time based emabargos aswell. The idea is to let people share information however they prefer. - Joe Fitzsimons

Ricardo's right in that a determined scooper (may their entrails knot!) will simply claim to have had the stolen idea independently. Proving otherwise is going to be a nightmare than no one will want to take on -- although user tracking might help to establish a pattern of behaviour. I think Cam has hit the main point, that this sort of Shared system might be something more people would be comfortable with -- for example, creating a group for a few labs that are collaborating. - Bill Hooker

What might be really useful is a bridge: some kind of description/running summary of the Shared Notebook that is fully indexed and Open, so that potential collaborators can still find you, get in touch, log in and become part of the project. - Bill Hooker

Bill, so we'd have an open/accessible project "abstract" and everything else closed? - Ricardo Vidal

My hope is that by making the user tracking sufficiently obvious and accessible that it will discourage scooping. Of course no-one is going to try to prove someone stole their ideas, but if this information is available for the world to see, then I expect that people will be sufficiently concerned with their reputation not to want to be seen to steal ideas. It's a disincentive, not a barrier. - Joe Fitzsimons

I left blogs enabled to act as exactly the kind of bridge you are suggesting. Additionally, wiki structure can be seen, but not content. You can also create shared group wikis for collaborating. - Joe Fitzsimons

If anyone has an ideas on how this can be made better, I'd be delighted to hear them, and if anybody wants to get involved in the development I welcome that too. - Joe Fitzsimons

Joe, have you had the chance to look at OpenWetWare? (There's a link on the main page to the lab notebook part) - Ricardo Vidal

Ricardo - yes I was wondering if OpenWetWare has the same features - I think you can control access pretty well if you wish on OWW. So the added feature here is that it would be by invitation only then. - Jean-Claude Bradley

Ricardo- No, but I'll have a look. Jean-Claude- Actually the user tracking (which I plan on expanding substantially) was my main reason for setting it up. - Joe Fitzsimons

Joe - you would have to check with OWW to see how closely they track pageviews - Jean-Claude Bradley

The point isn't for me to track the page views, but rather for it to be done is a public and open way. You don't even need an account to view what's been happening. I realise that my current directed graph lacks detail, and I'm working on fixing that. As I say, this was just an experiment, and if people like it, great, if not, no big deal. - Joe Fitzsimons

How do you ensure "real names" (a la Amazon.com)? What's to prevent people from using fake profiles (a la BugMeNot.com)? How do you prevent information from leaving the system (e.g. someone could "steal" an idea from a paper someone wrote who has access to Social Notebook)? On the other hand I can see some use in having a credible way to establish that idea x had been mentioned by person y on date z. Don't know how suitable existing preprint servers such as Nature Preceedings are for this? - Eric Jain

Real names should work because it is invitation only. I'm happy to invite anyone who gives me a legitimate institutional address, since that is a reasonable gaurantee of their identity. Alternatively, any use can invite friends. Presumably they know these people. I'm not entirely sure I follow you about information leaving the system though. The idea is to discourage users from stealing other users ideas. This can't prevent a subsequent person from stealing ideas outside of the system. Have I misunderstood? - Joe Fitzsimons

Eric, I guess real names are verified as humanly possible. If someone happens to make it through, it's because they were sneaky and not bots. Also, we ask not only for a real name and email. we also ask for a reason, affiliation, etc. - Ricardo Vidal

Its very hard to work against the active and deliberate cheater, at least in the short term. This is mostly about discouraging I think. I think the central criticism of this approach remains that it isn't properly Googleable. As Jean-Claude and others have said, this is crucial for people to find you. John Wilbanks is also good on the problem of 'closed gardens'. The network effects we want to see depend on large networks and accidental discovery - if it isn't discoverable it doesn't exist. If you can solve this problem and have robust (and visible) view tracking and google searchable pages then I think you've got something quite powerful. And as Bill says, if we can provide bridges for people to ease them into a more open approach then I think that's very helpful - Cameron Neylon

We'll be having another one of our Town Hall meetings at OWW tomorrow and the topic will be none other than Lab Notebooks. I'd like to invite anyone interested in the topic to join in. More info here: http://bit.ly/2esgCF - Ricardo Vidal

Cameron - I think Joe has fixed the issue of Google indexability, at least for the totally public setting. - Jean-Claude Bradley

Joe - there is definitely a need for more people to get involved in Open Science. People are not trying to tell you to stop - they're just giving info about what is going on. We always learn something from any attempts to get something done. - Jean-Claude Bradley

Yes, as Jean-Claude says, it was a bug, not a feature. It's now fixed. I think it would be a bad idea to make stuff that was marked other than 'Public' as searchable via google. There is a built in keyword system, etc. It would be nice to have a search that brings up just the page title if the user is logged out. Not sure how easily I can impliment this though. - Joe Fitzsimons

I was trying to make a slightly different point. A lot of the pull in ONS comes from adventitious discovery. While the halfway house of being visible to known people may be something that people are more comfortable with they then won't get the big benefits (because only a very small proportion of people will be known). Now if there is a way to make people feel safe enough to come in _and_ give them the benefits then its a winner. Access via openid with anyone who asserts an appropriate email address? - Cameron Neylon

Or as an aside as I commented elsewhere tonight. Perhaps if the microcredit problem is solved along with attribution for blog posts, comments, etc perhaps people will want to assert a 'real' identity before looking at things so their contribution is registered - Cameron Neylon

Ah. Open ID is certainly doable. Maybe the email address is enough, but really I want people to be easily identifiable, otherwise the user tracking is essentially worthless. Certainly real name + email address for verification are enough. I am just not sure how best this can be checked. Certainly .edu and .ac.uk email addresses could be automatically accepted. Not sure what to do about the rest. - Joe Fitzsimons

Actually I would drop the invitation system in a flash if there was a simple way to verify real names. Any ideas? - Joe Fitzsimons

Amazon uses credit cards to verify names. Probably not an option here :-) - Eric Jain

The Amazon suggestion is interesting - maybe you could authenticate by charging a very nominal amount (say, 10c) using the Amazon Flexible Payment system. I've never used the system, so don't know it would work for this purpose or not. - Michael Nielsen

A bit of context: my understanding is that Amazon Flexible Payments allows other people to use the existing Amazon system to do payments, paying Amazon a small royalty. - Michael Nielsen

Best is the enemy of the good... Having a site that allows people who would otherwise share nothing share something might be a good thing. But I don't get the logic behind tracking people to discourage plagiarism: If I was worried about people stealing my ideas, would I really want to show them to people whom I don't trust to give proper credit unless they're being tracked? - Eric Jain

Eric - Well, it's the idea of making your ideas public, rather than showing them to specific people. I really don't think people would risk their scientific reputation if there is a non-negligible chance of being caught. Of course you can't stop someone who doesn't care about getting caught plagarising, but I don't think these people represent the real danger. You could blatantly rip preprints off the arXiv and submit them to journals if you really don't care about being caught. - Joe Fitzsimons

Posting a preprint to a server such as arXive feels safe (perhaps even safer than keeping something to oneself) because doing so establishes a date and authorship. Sort of like a "patent pending". Having something similar at a more detailed level (ideas?) could be useful to encourage people (i.e. make it in their own interest...) to share. But this doesn't require tracking access, which seems problematic to me (and not all that robust anyway)? - Eric Jain

Actually, the UK arXiv server does have limited user tracking. - Joe Fitzsimons

http://arxiv.org/help/faq/stat... argues against showing what paper was accessed how many times and by whom, so I'd be surprised if one of the mirrors did so nevertheless? I suppose they do collect some information for internal use (e.g. to detect abuse). There is also the trackback feature (detect and show incoming links), but that's something altogether different... - Eric Jain

Have you never seen CiteBase? Here is an example: http://www.citebase.org/abstra... just click 'Downloads'. lt lists downloads by organization, country or date. - Joe Fitzsimons

The UK arXiv mirror does indeed appear to collect some access stats (for use in Citebase, which happens to be hosted at the same institution, hmm)... Anyway, at this point we might as well merge this thread with Eva's impact factor thread (and check out Maxine's reference to a blog post about download-citation correlation). - Eric Jain