Cameron Neylon

The institutional repository. - D0r0th34

as long as it is google accessible...and I think several of these people come from unis without repositories that are good on data but nonetheless, this ought to be the first choice... - Cameron Neylon

Most IRs are. :) - D0r0th34

How many are set up to cope with data? I should know this - but I've always tended to go off site or have our own services. It always seemed difficult for the Cambridge people to get stuff in and out. - Cameron Neylon

What sort of data are we talking about, and what kind of coping do you need? - D0r0th34

Bleh, I should have read the OP. Here's the basic skinny: IRs will do fine with discrete files in pretty much whatever format you care to throw at them. Actual databases are liable to be a problem, however -- the IR won't have built-in query capability for them. Ditto anything (model, visualization, whatever) requiring specialized interaction patterns. Does that help? - D0r0th34

(free) Amazon Public Data sets... (free?) http://www.freebase.com/ (not free) Amazon S3... Microsoft Azure... - Richard Akerman

D, yes that's helpful. Presumably anything with specialised metadata is also unlikely to be exposed? I think the problem is that most of these are collections of files, sometimes with some sort of descriptive file that a specialist could make sense of. Google Data Sets would have been perfect for this....[sigh] - Cameron Neylon via twhirl

maybe archive.org? also think direct p2p release will become an increasingly viable option - Mike Chelen

Specialized metadata may be manageable; talk to the IR specialist. The tech infrastructure will unfortunately determine a bit too much... but even DSpace can deal with novel key-value pairs. (Hierarchical XML metadata, unfortunately, it won't do much with... not without srs Manakin hacking.) - D0r0th34

So key message is talk to your local IR people... ;-) - Cameron Neylon

OAI-ORE is built for this kind of thing of course but I don't know if there are any tools for packaging arbitrary things up that way. - Cameron Neylon

Not yet, I don't think, but yes, one hopes it's coming. (HvdS may have built it already, of course!) And yes, a lot depends on how accommodating and enterprising the local IR folks are. Some are more, some unfortunately not so much. - D0r0th34

What about if odds and ends of data come from different institutions? My PhD, 1st postdoc and 2nd postdoc have all generated odds and ends that might be useful to someone but no longer of interest to me. I think it would be better to have it all in one place, not odds and sods in each relevant IR. - Clare Warren

You need an institutional affiliation to use it, but Dataverse (http://thedata.org/) is pretty good -- easy to set up, web or local, lots of options. For instance, each lab could have its own dataverse for the sorts of "odds and sods" that Clare mentions. - Bill Hooker

I think that a torrent tracker for scientific datasets would be sweet - Marcos de Carvalho

torrent is kind of cool/amusing idea but if no one is seeding the torrent, it is effectively gone - Richard Akerman via BuddyFeed

mininova provides free torrent hosting & seeding through their content distribution program http://www.mininova.org/distrib... however researchers may feel opposed to the use of consumer grade solutions - Mike Chelen via IM

Web seeding from one of the providers cited above could be an alternative, in addition to the swarm. This could help relieve the bandwidth from a unique source and also acts as a backup in case everybody stop seeding a specific torrent or the default provider goes out. - Marcos de Carvalho

Although Mininova is a cool service I agree that people would be a little afraid to use it. However, searching SourceForge I found lots of opensource torrent trackers, LAMP based, ready to install. - Marcos de Carvalho

nice thing about mininova is that they provide the file hosting, and maintain at least 1 seed indefinitely. a web seed can accomplish something similar now too, given how inexpensive hosting plans have become. here's an example of scientific data (fasta dna from ensembl project) being distributed: http://www.mininova.org/tor... - Mike Chelen via IM

another promising p2p filesharing system is made by http://wuala.com using a modified bittorrent protocol, and features filesystem integration allowing the distributed data to appear as a standard network drive. however, it isn't backwards compatible with bittorrent - Mike Chelen via IM

I really like the concept of these images, but I am concerned that you have used a CC attribution license on flickr. Firstly, I am not sure if you can really CC the images that are derived from YouTube and Twitter logos in this way. Secondly, the CC license requires that if anyone wants to use them in their talks they have to put an attribution to you on the slide. Do you really want to require this? - Matt Leifer

All I'm doing at the moment really is to put some ideas out there as to what people could do. The social media images are part of a collection that is described as "free for use". If there is any copyright in what I've done (which frankly is pretty doubtful) with the collection of things then that doesn't necessarily make any difference to the original rights which flow through. And I... more... - Cameron Neylon

That sounds fair enough. Where do the social media images come from by the way? - Matt Leifer

Oh good point. I got them from here - http://www.thisismyurl.com/free-do... - Cameron Neylon

Since the facts of a presentation are not copyrightable, maybe the text should read more in the direction "You are kindly asked to not (immediately) blog, twitter etc. this presentation"? And maybe also make the logos more suggestive of "request" rather than prohibition? (at least for the ones that are transcription rather than recording) - Anders Norgaard

I thought about wording it "Please do not..." but decided against because it seemed somewhat unclear. There would be an argument that such a panel forms the equivalent of "click wrap" agreement (by staying and listening you are agreeing to my conditions) so copyright per se isn't necessarily the only right or limitation. How would show a request rather than a prohibition? Red circle? - Cameron Neylon

"Greyed out"? - Anders Norgaard

IANAL, but isn't emulating "click wrap" agreements a really bad idea? The EULAs of software are pretty much universally despised and ignored. And your sensible point about licensing of data - that assuming rights in databases of facts is on shaky ground must equally apply here? - Anders Norgaard

Probably, and obviously I'm not a lawyer either, but I just meant to make the point that copyright per se wasn't the only potential right that a presenter held. There are also arguably performance rights in any audio and video, broadcast rights in some jurisdictions, and potentially other contractual rights. The reason I said "not permitted" was simply that it is (hopefully) clearcut,... more... - Cameron Neylon

sxc.hu has an open license option, pretty mucht he only restriction is you can't use the images for porn... - Brian Krueger - LabSpaces

I agree that audio and video is different from transcription (twitter, blog, etc.) so it may also be worth trying to distinguish. Oh, well. - Anders Norgaard

Wrt. blob vs live-blog on one of those images: I don't see much of a difference between live-blogging during (or right after) a presentation at a meeting on one hand, and blogging about some topic right after you get home from the conference. If people presenting work wish to limit the audience in terms of how they can or cannot discuss what they are presenting with, then why are they... more... - 'Mummi' Thorisson

It's sloppy work, is what it is. When I leave a lab I leave behind organized and labeled freezer boxes and multiple copies of a cheat sheet showing the location of every important reagent and all original data. It's just not that much work to do this -- it should be expected of every professional scientist. (As should decent notes that other people can follow, but I learned long ago that some things are considered just too much to ask...) - Bill Hooker

2 jobs in the Navy I left very complete notes, points of contact, and had 1-2 week turnover periods... my successors both threw out my notes and got into serious trouble within months and in the first instance, I had to bail them out by stepping back in and cleaning up.... so it sounds good, but if your successor will toss it anyway, then... - Christina Pikas

I think the problem Jenny (and Christina) raises is more one of people, so the only way to improve it is to train scientists to be more methodological. - Bob O'Hara