Good questions and interesting answers on citation of things that might be ephemeral by their nature and not just by being on the web (e.g. how do you cite a particular configuration or trajectory in a 3d simulation of a molecule...actually you probably cite a PyMol script...but you get the idea) - Cameron Neylon

in your turning 3D digital object, Cameron, let's imagine there's some change or development you want to point to specifically, how do you do this? if the starting second is unequivocal, do you say you are referring to second xy in contrast to second yx? - Claudia Koltzenburg

If you were using PyMol you would be able to set up the exact motion and scene you want to display. So in a sense you would be citing the structure, but through the prism of a specific display tool running a specific set of instructions. In principle if the visualisation was a web service that took a path as a set of parameters then a citation might read "http://myvisualisationservice.org/pdbcode..." Actually thinking about it the ability of URLs to pose as both addresses and instructions sets could be really powerful in solving the problem...this requires some more thought... - Cameron Neylon

sounds really interesting, Cameron, also in view of possibly being more independent of specific (proprietary?) display tools, maybe? - Claudia Koltzenburg

It would be tough to make it completely independent but given (at least for 3d movies) an open 3d scene standard and a web based open visualisation engine it could be done. Interestingly I was involved in a discussion here at STFC about trying to build a 3d description standard but it was decided that it wasn't worth the effort to put together an application because no-one would fund such a thing... - Cameron Neylon

re building a 3d description standard, thanks - has this been tried/done elsewhere on the globe? (how long ago is this discussion at STFC?) - any specific reasons why your expert colleagues think no one would fund such a thing? let us ask: in whose interest would such a standard be created? everyone with sufficiently reliable web access, technology - and bandwidth - now or in the future who is doing - and citing - research, maybe. - Claudia Koltzenburg

Google are doing something called O3D (Open 3D) which may solve part of the problem (http://code.google.com/apis...) although porting a large quantity of legacy molecular rendering sounds like hard work to me. Research funders won't touch this stuff because it is basically incremental. It isn't new functionality it is making existing functionality more available. But I agree the applicability is very wide. I need to think more about the idea of functional citations as well - Cameron Neylon

it seems to me that looking into the IPTC standards (used by image archives) might be a track to follow. While this is not used for multimedia (as far as I know), there still seems to be a knowledge gap between the established scientific journals and other fields of business in this respect. At least, looking around in hematology/oncology, we found no journal which uses IPTC standards (nor any other, it seemed). Images from articles seem to be downloadable without any scientifically reliable trace of where they have been taken from or what their orginal scientific context is... I might of course be mistaken and anyone who can point me to best practise examples in any field of research (for any non-text object on the web) is most welcome to do so :-) - Claudia Koltzenburg

i think there are a lot of different points involved in whether/how/if to cite things other than articles. There's if it is archived so it will be preserved. If it has a DOI, handle, purl, or some other sort of unique identifier so that you can reliably point to it (Claudia - I think your comment on my blog points to this). As in above, there's the bit of referring to a point in a video so you can accurately point to it - but we have included supplemental data in journals, including videos and we have annotated pictures in articles already. - Christina Pikas

Incidentally, figures in PLoS journals have their own DOI which is a help. Christina, what's your feeling on the idea of functional citations I referred to? A date stamped version of a dynamic document would be a simple example of such a thing but is there a fundamental objection to a /process/ being part of a citation? i.e. take object x and do this set of things to it to see what I saw... - Cameron Neylon

Cameron, yes, figures in PLoS have their doi, - and what can you find out about the image and its specific meaning after you have downloaded it? does the image viewing tool you use give you any more information about its original scientific context than this cryptic doi? if yes, great: tell me which tool are you using - and which tool do you use for moving non-text objects? Actually, I am not sure if in creative commons licensed research we need watermarks and all that but - coming back to the hurdles listed for NOT CITING multimedia objects (see the NN forum topic link) - what is considered to be a 'valid' reference for which type of resource - and what does this mean technically speaking? btw, I really like your idea of functional citations with a /process/ included. Maybe we could ask someone in who is knowledgeable about OAI-ORE http://www.openarchives.org/ore... ? - Claudia Koltzenburg

Christina, re annotated pictures in articles, could you point me to an example, please? what kind of annotation is it and would it be useful for animated files (or whatever we choose to call them), too? - Claudia Koltzenburg

Claudia, two answers to your first question: What do you get from the doi? Well the same as you get from the doi for a paper. Very little until you follow the link through to the object of interest, then you usually have to evaluate the context manually. ...after you have downloaded it? Well in an ideal world you wouldn't download you would incorporate by reference, that is you wouldn't have a copy of the image/video/whatever you would have a reference that pulled the original in. But the broader question of a valid reference I think its less about the resource and more about the "containers" that are holding them. Seems to me that a doi is the gold standard of a stable reference to a container at the moment. If the concern is about determining context after you land at a reference point then OAI-ORE may be an answer because it allows more sophisticated container descriptions. - Cameron Neylon

@Cameron - intrigued with the idea of process related citations. Pointing to a specific version however they're tracked (date time stamp) should be the rule. I guess the question is: how much is in the citing context and how much is in the citation? I'm sure I've seen papers thay say ... as Smith says in his introduction... There have also been efforts since maybe the '60s to find a way to provide some information about citing context - computational linguistics people come back to this from time to time, too. So how much goes in the citation (this varies by style - page # required? and in practice - whole book or analytic?) and what belongs in the citing context? - Christina Pikas

good points made, very interesting, thanks :-) @Cameron, I think that in everyday lives and practices in global and varied contexts we might consider that scientific images, too, start traveling on their own. Especially important in medicine that these have glued to them the relevant metadata... @Christina, are we sure there really is no concept or usage in between these two? - Claudia Koltzenburg

Claudia, yes I think I have only just really figured out exactly what was motivating your original question. My gut feeling is that given the underlying file types it will be impossible to do this perfectly from a technical perspective. Yes you can embed stuff as watermarks or put it in the metadata of the image file itself but it is easy to lose either through mistake or malice. Evolving good practice in citation and packaging will probably be the key - which brings us back round to the question of how best to cite these things! - Cameron Neylon

hm, my original question is the one of motivation and habits: do we as researchers really give credit to all the sources that are relevant to the outcome we publish? or are some resources more equal than others ;-)? - Claudia Koltzenburg

http://scienceonlinelondon.wik... I submitted a session idea based on this conversation and other similar ones. If you're interested, I'd love to hear about it. - Jennifer Melinn

Fwd: EPub as a way of packaging scholarly resources? - http://ptsefton.com/2010... (via http://ff.im/pbs16) mentions OAI-ORE and METS - Claudia Koltzenburg