Those outside science are often surprised to learn that we usually have to pay for the privilege of being published. The charges incurred by authors, argue the journal publishers, reflect the cost of the significant value that added when a raw manuscript is converted into a scientific paper. And, at a pinch, I would tend to agree.
I remember seeing my very first paper appear in Biochemistry back in 1990. It was a magical moment. The bulky double-spaced manuscript of text, references, figure legends, tables and figures had been transformed into a slender, formatted paper. I can recall making my way to the library to seek out the fresh, green-covered issue of Biochemistry that contained my article. I don’t think I would go so far as to describe the paper as an object of beauty, but it was nevertheless for me a thing of wonder and joy. To see my words and graphs laid out pristinely on the page, in two regular columns with each table and figure neatly embedded in the text produced a moment of quiet fulfilment.
Science (and history), in the making. From Biochemistry
But that wasn’t all.
An extra smack of satisfaction derived from the knowledge that, until that day, papers were things that I read, that I pored over and struggled to absorb. Always they were articles that had been written by other people. But here, for the first time, was something created by me. I had found something out and now it was being presented to the world. A once impossible mountain had been climbed.
And I knew that my work had attained a kind of permanence (inasmuch as we may grant our civilisation that destiny). As part of the published scientific literature, the paper was indexed, catalogued and sent to libraries all over the world.
That sense of magic remains, even 19 years later. I still love to get the galley proofs of each new paper – there is a distinct sweetness to the first glimpse of the transformation of a raw manuscript into a peer-reviewed publication.
These days of course, newly published papers almost immediately appear in PubMed and are available for download across the planet (subscription or open access permitting). The advent of the Internet has also changed the type of information that can be published. No longer are we restricted to the published page or even to a fixed number of printed pages. These days we can add, affix, annex, append and augment. In a word, we can supplement. It is now common for a scientific paper to tow in its wake a great mass of supplementary online information. This provides a rich resource of additional text, images, sounds and video.
Or, at least, it should do.
Usually the added material consists of more text and figures than could be accommodated on the printed page and it’s a boon to be able to include it in the public record. But too often, alas, the magical ‘added value’ of publication is lacking from the presentation of this supplementary information.
What amazes me is that, despite the evident care taken over the paper, no-one seems to have given much thought to the way that the supplementary information is provided to the reader. As a case in point, I was reading a Nature paper last week, in preparation for a review I’m writing and needed to track down the associated supplementary information. It was easy enough to find but the vision that greeted me on the web-site was a crushing disappointment. There were no fewer than nine separate files to download. Nine!
More or less? You decide.
Six of the files were Word documents and three were pdfs. The information was a benign mix of words and figures but for reasons that are lost to me no-one had thought to combine them into a single unit. Or to integrate the information to make it more accessible — to put the figure legends with the figures, for example.
I don’t mean to pick on the author (or the journal*) in this case, since my experience is that this fragmentation of supplementary information is fairly commonplace. There is presently a degree of chaos that we could do without; it is degrading the reader’s experience. But I am also worried that the random and proliferating set of filetypes being used to store supplementary information is undermining the permanence of the scientific record.
I think we, authors and publishers, can do a lot better. With just a little supplementary effort.
*Update – As Maxine points out in the comments below, the Nature stable of journals have made great strides since 2005 to improve the formattting of supplementary information that they publish.
So, are there any requirements for the submission of supplementary information – or is it just ‘must be in the following file types. . .’ It seems this may be a journal specific requirement and unlikely to adopt a universal standard anytime soon. You would think though (and I don’t know the authors) that the data would be a bit better organized from earlier incarnations of the manuscript. I agree that supplementary effort should be employed.
Why aren’t online supplemental materials so non-Web savvy? I guess all of this stuff could be put up with html, which would make access much simpler.
Nice post Stephen, and a bloody good point.
When’s the revolution?
Stephen – you are not alone. Claus Wilke raised point about the format of supplementary information in a letter to Nature in 2004 – Supplementary materials need the right format, saying that Nature should switch from Word to pdf format. There’s also some discussion at OpenWetWare of supplemental material, calling for a central repository.
Where different kinds of information are involved (movie, image, molecular structure) then of course different formats and files are unavoidable. I’m not sure that HTML is a great format for preservation purposes. I agree it seems unnecessarily messy to have several different files containing Word or PDF content.
I remember many moons ago (early 1980s) being asked to get some supplementary material from the J Am Chem Soc. At that time this meant generating an order to be sent to the ACS, paying the fee (the price of a small book) and waiting for the microfiche to be delivered. It probably took a month or so all told. I was surprised to see that the highly-desired supplementary material was a few typewritten pages, with hand-written annotations, plus one or two wobbly graphs.
@Craig – I think that part of the problem is that journals don’t give a strong enough steer to authors, both about filetypes and the format of the supplementary information. In one way it is good to have the freedom to include a range of information, but too few people take advantage of that to do a good job of putting the package together.
I tend to agree with Frank, Bob, that HTML is not-necessarily the best format. Since the papers themselves are always available as pdfs, an established format that is excellent for combining text and figures, why not apply the same standard to Supp. Info.?
These days it is easy enough to combine text and figures, either within Word or PowerPoint (or their equivalents). On a Mac it is now a trivial step to convert such files to PDFs. It’s odd that the very technology that facilitates the publication of supplementary online material is not being exploited to make the content more digestible. Frank’s story throws into sharp relief how much better off we are now. But we can do better still.
Richard – the revolution starts here and now! As a reviewer myself I have started asking for Supp Info to be put together into more integrated formats (I try to do this on my own initiative as an author). I urge others to do the same.
@Frank – meant to thank you for the links. I’m aware I’m treading a well worn route though. Disappointing to see that not much has changed since 2004.
But surely Nature can’t ignore its own Network of influential bloggers! 😉
Or can it?
One of the columnists of the Journal of Cell Science has a number of interesting observations to make with respect to linkages between the (peer) review process and the online supplementation process. It is too long to reproduce here, but for those interested, it is easily accessible here: Part I and Part II.
Thanks Bart – those articles are very good. A teensy bit windy (cut to the chase in Part II if you’re short of time) and the author, the bizarrely-named ‘Mole’ falls into their own fragmentation of Supp. Info. trp (though perhaps that was meant ironically?).
Mole explores a point I hadn’t considered – namely that the existence of Supp. Info. may be making it too easy for reviewers and editors to ask for additional experiments and comes close to advocating a ban on supplementary text and figures (apart from detailed Methods). And goes on to suggest that such material could perhaps be posted on the author’s own web-site. I certainly wouldn’t go that far. That way lies utter chaos.
We simply need some added value.
Perhaps the solution is simply to move to Open Data, that way all data is deposited in an appropriate repository, in an appropriate format, with appropriate preservation policies, and the article can just link to it.
Easy, no? 😉
@ Stephen: I agree with you that “Mole” goes too far in his radical restriction of Supplemental Materials to methods. Nevertheless, his point in how the use, ease and availability of Supplemental Materials influences not only the process of scientific publishing, but also of science itself, is worthwhile.
The more readable, accessible, scannable and searchable Supplemental Material becomes, the more normal it may become to request or even demand it, by journals, reviewers or readers. This has severe consequences not only for how papers look, but also for the amount and type of scientific labour that goes into them. Supplemental Material is an extra and it should add value, but papers should not derive their value from the Supplemental Material. While there is no harm in requesting Supplemental Material to be kept to a minimum to counteract such trends, ignoring the possibilities digital publishing has to offer would be counterproductive.
@Frank – don’t go there. Just. Don’t.
@Bart – I’m with you. Mole does raise the interesting issue of the possibly ever-expanding size of the minimal publishable unit that Supp. Info. may seem to introduce.
As an author I confess I haven’t sensed that some sort of floodgate had been opened. The remaining constraints on the size of the published paper should, I hope, keep a downward pressure on the amount of work needed to get a single paper.
I have heard some mutterings that online-only journals could allow scientists to publish at great length but I am not at all sympathetic to that view. There are just too many self-important windbags out there. Brevity is a virtue that should be cultivated (even while Supp. Info. gives us some necessary latitude).
Nice post Stephen. I also wondered why supplementary material to a well-presented, formatted PDF is so, well, ugly – double spaced, line numbered MS Word material hastily converted into a PDF…
Apart from that – I would like to congratulate you with that very nice graph! I like the fonts and the presentation in these ‘old’ figures. They are much nicer than the standard MS Excel stuff you see in many journals today.
bq. As a reviewer myself I have started asking for Supp Info to be put together into more integrated format
fantastic, Stephen. This is better than writing to one’s MP, yeah?
@Thanks Raf. Of course the graph was produced in pre-Excel days. Actually I still haven’t mastered Excel for that function. Kaleidagraph is better (and Numbers is shaping up…).
@Richard – no reply yet from my MP about the issues raised in the Singh libel case. How ’bout yours?
Nope, not a frelling sausage since the automated response.
Bah. My current gripe is journals that put the materials and methods into supplementary information. Sometimes, just sometimes, I like to read those to see if the conclusions allegedly drawn from the data were arrived at in a sensible way.
Also, from Nature Methods:
“Authors should note that Supplementary Information is not copy-edited…”.
Not to place any burden on already-overworked editors, but perhaps some copy-editing (or editing for style) would help – or at least if the editorial staff looked at the stuff and said “this is a big pile of nonsensical mumbo-jumbo, please take it away and make it readable” or words to that effect.
@Richard – It is frustrating that M&M are more often shoved into supplementary information but I could live with that as long as the stuff is well presented.
I take the point about copy-editing but only partially. Apart from Nature and some other high-falutin’ journals, I’m not sure how much copy-editing goes on. Reviewers should be able to comment on the overall clarity of the writing and figures for the manuscript and the supp. info. (though it is certainly not their responsibility to correct everything).
I wonder if there’s a role for readers here, through the magic of Web2.0? Perhaps publication of online material could come with a rider that the author has ongoing responsibility to respond constructively to suggestions from readers and amend the material with any suggested improvements?
A few comments from Nature.
(1) The paper you reference was published in 2005. Two or three years ago, we hired a full-time member of staff whose sole job it is to merge components of SI into a PDF for a single download (where format permits, eg excepting movies, excel sheets and other manipulatable data, and so on). So if you look at contemporary Nature, you will find SI in single download format (might have been some research that could have been done before writing the post, possibly?)
(2) Nature’s SI guide to authors, as well as the letters the editors write to them, request them to merge SI components into a single PDF where possible. (Our production editor combines SI for authors who don’t do this.)
(3) SI is free – therefore we don’t have the production resources to put into processing it compared with the rest of the paper. (Yes, it is peer-reviewed).
(4) For Nature and increasingly the other Nature jouranls, Supplementary Methods are sub(copy) edited, formatted and integrated into the main online PDF of the paper, so are downloaded with the main paper.
I was rather sorry to read this post and some of the comments. We have put a lot of effort into improving our SI since 2005 – and note that it is free – so I’m sorry this has not been noticed by anyone. Could it be that nobody commenting here has looked at any 2009 SI before commenting?
Perhaps Nature isn’t publishing stuff people want to read—
hang on, someone’s at the door
@Maxine – What an elegant smackdown! Thanks for setting the record straight.
Two or three years ago, we hired a full-time member of staff whose sole job it is to merge components of SI into a PDF for a single download…
How long did it take that person to burn out/leave for a more rewarding career? 😉
I still hate supplementary online information, not least because it’s not really portable, unless, as Stephen says, you download all umpteen files (or tipping my hat to Maxine’s unfortunate
slavedronejunior editorial person, one big honkin’ PDF file) and take them with you. I like the idea of having all those primary data and things available, just not the mechanics of it. Give me a nice, succinctly written, single manuscript anyday.Actually, Richard (W), she enjoys her job. She does quite a bit of formatting to the SI also. I’m sorry you think it sounds boring.
I do agree that a succinct paper is best – as a reader it has more impact, is more accessible and is just less of a burden to wade through if the paper is crisp and focused about what it is saying. But nobody writes those any more.
One more point of information: Stephen you note above that “journals don’t give enough steer to authors, both about filetypes and format”….
Here are Nature’s SI guidelines, and here is general info about SI across all the Nature journals. Forgive my riposte, but as you have chosen to “name and shame” Nature specifically in your post, I feel honour-bound to point these out.
Frank – re your anecdote about the microfiche, etc. I remember when we first started SI on Nature (an idea of John Maddox’s), which was pre- the web – in the 1980s. The first types of SI were lists of crystallographic coordinates and the like. We used to keep them in a filing cabinet and the editor’s secretary, would send a photocopy to anyone who wrote in for one. All free of course. If you look at old Natures in the archive from that era, the papers concerned have a footnote that indicates this, and that Mary Sheehan will supply copies upon request. (We’ve since scanned it and put it in the online archive. It took ages to go through that filing cabinet!)
Maxine – I’m very sorry if I caused you any grief. My post was certainly not intended as an anti-_Nature_ rant; it was mere coincidence that the paper that sparked the idea for this post came from that journal. I made a point of not picking on the author in this instance and I guess I should have done the same for the journal.
I’m grateful to you for putting me straight on the matter. You are quite right that I hadn’t researched more recent publications from Nature but this was a blog post, not a scholarly analysis of academic publishing. I saw it as a personal account of one aspect of my recent experience that might stimulate a useful discussion, which it has.
I completely applaud what Nature has done is this regard, but of course the problem that I encountered remains in much of the literature elsewhere (as well as in Nature’s archives, though I fully appreciate the cost implications of going back through old files!). From remarks made by some of the other commenters, I am not alone in thinking that there are problems with the presentation of supplementary information. If I might be permitted a further single example (from outside NPG), I recently discovered that the links to the Supp. Info. on one of my own papers in Structure now takes you to a pair of blank files. I have written to the publisher to request a fix (some weeks ago) and heard nothing.
Post updated in a wonderfully Web2.0 way…!
Thanks, Stephen, I can now pick up my little bleached skeleton from the floor and carry on to the next bright day 😉
fight fight fight
>:)
Now I am disturbed: Maxine carries around a little skeleton and, over on Twitter, Henry has been dragging a carcass hither and thither… what is going on
@Ian – for shame! That’s what’s called robust but civil discourse. Thank goodness the BCA weren’t involved… ;-p
I am in Ian’s Camp….NN smackdown – my money is on Maxine 🙂 (sorry Stephen).
Maybe this could happen at that Science-on-line thingy y’all are having across the pond and someone could live blog/twitter/FF/FB update/Youtube video (on the Nature Channel of course) the whole thing.
Now I am stuck, because I want to be on Stephen’s and Ian’s and Craig’s side….is this possible?
There are no sides, Maxine.
We are all in this… together.
I have a solution, Maxine. Since Ian and Craig are so keen on fighting, I’ll nominate Ian as my second and you can have Craig. We can then sit back and enjoy the show (without being the least bit partisan). I’ll get the popcorn…
(Watch yourself, Craig, I think Ian’s quite good. I believe he practises.)
@Maxine: Despite Nature’s efforts to (a) provide guidelines on how to use Supplemental Material and (b) make sure the information presented through Supplemental Material is checked and presented in an aesthetically pleasing way, one key issue remains:
The increased use of Supplemental Material makes the boundaries of what a scientific publication is fuzzy. Where does the peer-reviewed status begin and end, what happens to shared criteria for the size of the “publishable unit”, how does the uninhibited spray of data connect to subsequent publishing attempts. For instance: can Supplemental Material be Supplemental Material to two papers? If so, where does the boundary between the two publications lie? What does that mean for citation or evaluation?
The plain and simple fact that Supplemental Information exists, is facilitated, is used in increasing amounts and frequencies changes what a scientific publication is. Is is a “unit” (identifiable by a set of page numbers, or a single link) or is it a collection (multiple files on, perhaps, multiple servers, identified by multiple links)? Is responsiblity for all the supplemental data shared by all authors on the paper, or can one or more authors have their name attached more to some bits of the data (a trend to individualise responsibility for contributions is clearly visible as well)?
I am conviced that supplemental data is not a footnote in the process of knowledge production, but an agent of change…
Didn’t mean to dump on your colleague’s job, Maxine – just trying a bit too hard to be humorous I guess. What I should have said was “I don’t think I would have the determination and tenacity to do that long-term”. 🙂
Having just looked at one of Barbara McClintock’s papers from the 1940s, which was about 30 pages of dense narrative, I’m thinking that there’s something to the idea of putting some of the material into supplemental information. Classic paper and nicely written, but loooooooooooooong.
Barbara McClintock is my personal hero! Genes do so jump around!
Yes, and chromosomes have “telomeres” too. What an amazing career.
O.K. Stephen – Not that I think Maxine would need defending or a surrogate (she could certainly choose better then me) I would gladly stand in her corner. I know Ian likely has mad-crazy fighting skills but I am willing to risk it.
Perhaps we could do this in Second Life (except where band in Australia)?
Wheee! Gong Sau
Where’s the fun in a 2L fight?! Come on mate, it’ll be a laugh! We could have an official weigh in at the Nature
orificesOffices, and everything!I’m 165lb (fighting weight), 6’3″ with 2+ years of Muay Thai training. I’ve got an amateur record of 0-0, but I have enjoyed having the living daylights beaten out of me on a regular occasion by
complete strangersmy instructors and class mates. I also have a couple of belts in karate too, but that was so long ago I don’t think they count.After the
bloodbathmatch we’ll retire to the nearest pub for a nice pint and some sausage and mash!We could time if for one of Jenny’s geeks-in-the-pub, or whatever its called…
Hm. Ignoring all the talk of
drunkennesspainful beatings above, I just noticed something interesting about Stephen’s graph. He’s plotted the log10(Ki) on the y-axis, but rather than cluttering it up with error bars he’s put the associated standard errors in an accompanying table.Rather a nice way to clean up the graph without omitting/hiding the variability in the data, I think.
What an amazing career.You mean a-maize-ing, right?
Oh, sorry, too corny?
Oh dear, I go away for a few hours and look what happens – out of control. I would have suggested a nice game of cricket as a decider, but with Australia’s score today…perhaps not.
(As a nonsequitur, a couple of online friends are evidently cricket followers and have been tweeting using the #ashes. Out of mild interest I clicked on it and it was full of tweets from people saying “what is ashes?” and “what does ashes mean?” etc. Sad, sad.)
Bart – you raise interesting questions. As things stand at the moment, SI is peer-reviewed “supplemental” information – in its purest form, an underlying dataset. However, take methods and protocols. SI can be used to provide a recipe for others to repeat the experiment. But what about slight changes to the recipe that might improve the result? Nature journals suggest that authors might like to upload their detailed protocols to Nature Protocols network (a free service) upon publication of their papers, and other people can use Protocols network to comment on and add variants to these. So you are right, the distinctions are blurred, and presumably more of this kind of thing will evolve.
For the moment, however, the SI of a paper comes under the paper’s doi so is part of that paper’s unique identity.
First – I wonder how the adoption of SI helps with the concerns of releasing “unpublished” information to the public. Will peoples experience with SI publication translate to a future Open Lab book experience – is this a good way to start people doing/thinking about doing more electronic friendly data recording?
Second – Ian we are fairly matched on size, however, I have zero years of fighting training, but I can single-handedly wrestle a 2year old, a 6 year old and a 9 year old simultaneously – so let’s just call it a draw and head straight to the pub. First round is on Stephen and Maxine!
@Richard – why thank you.
@Cath – {groan} – yes!
@Bart/Maxine/Craig – it is interesting to see the development of SI (and to discover that it predates the internet). I suspect it will stimulate further reflection about archiving problems. Though I still feel strong resistance to the Open Lab concept (apart from Maxine’s protocols point above), I guess this could stimulate new ways to disseminate data within collaborations during the research that will eventually lead to publication.
@Stephen – Part of my curiosity stems from the off-and-on dust up of the use of the Freedom of Information Act (FoIA) in the U.S. To gain access to unpublished data created by government funding. I wonder if/how/when the government agencies will begin to embrace/require electronic documentation of data.
@craig: deal!
@stephen’s graph: I can only imagine the explosion if I suggested putting my s.d. or s.e. in a table… I can’t see reviewers in my
oldfield going for that. Nor mentors neither…My reference to Open Data above was inspired by thoughts akin to Craig’s. Sticking up little bits and pieces of data here and there as SI, in inappropriate formats (e.g. data tables and graphs in PDF) is all very well as a stop-gap solution to the need to give more background. But at some point we’ll realise that we have an unmanageable, undiscoverable, unmanipulatable mass of data on publisher’s servers. That’s when we”ll realise that we needed a data repository all along.
@Ian – Kind of ironic that the standard errors from the graph should have been included as ‘supplementary information’ in Table 1, which is on the previous page in the article. But the reason for this (I think – it was a long time ago) was because the ordinate axis is on a logarithmic scale and the errors are all about 10%, so the error bars are actually smaller than the symbols used for the data points. Plus the logarithmic scale would have meant using asymmetric error bars which was probably beyond the capability of our plotting program at the time.
I agree with your sentiments Frank, but Open Data (immediate publication of experimental results?) and good archiving are not quite the same thing. I’m all for the latter, but am mildly antagonistic to the former.
Some types of SI consist of data that other people want to manipulate – I am straying out of my area of knowledge here so please
take the gloves off and hit me with an Aussie cricket bat when I’m downcorrect me if I’m wrong, but microarrays and various spreadsheets (tab deliniated) are like this – which is precisely why we don’t merge spreadsheets into a combined PDF ;-). (And microarrays are now hosted in an annotated public database and linked from our site, not hosted on our site – public repositories being one desirable way forward for various kinds of data and materials. Like the journal’s doi, these provide the author, as “first discoverer”, with an accession number as a unique identifier.)I may be misunderstanding the points made by Stephen, Craig and Frank about openness, but in that SI is peer-reviewed and published – and should not vanish from a journal’s website*, that’s for sure, Stephen – this seems to me different in principle from the debate about openness of “general data that are not yet published”.
*I have to admit that we have, in the earliest days of the journal’s website, had a bit of trouble along these lines ourselves, eg when the journal web platform changed to a different provider. But once a publisher is informed of missing published material, I hope it would take steps to restore it. Hope you kept those old files, Stephen, in case you have to resupply them!
Strikes me that some people are getting creatively enraged because publishers are not dancing to their tune. Now there are problems, sure, and highlighting them is a good way to get them discussed.
However, attacking the publishers that are actively engaged in trying to find the best way forward is counterproductive and is likely to make them say ‘sod you’.
And reject your papers.
So let’s have a little respect and recognition that things will not change overnight. Nor should they.
Very interesting discussion. In this context I would like to point out a TalkScience seminar at the British Library with John Wilbanks in two weeks time: Scientific findings in a digital world: What is the genuine article?. Incidentally, and for those of us that can’t go to that seminar, we also plan a session on this topic at Science Online London. The original idea was to talk about non-traditional article formats (e.g. methods paper, or a paper only containing data) and media formats, but what to do with supplementary information to me seems also relevant to that session.
@Maxine – I think we agree on the difference between archiving SI and open labbook styles of doing science. And you’re reminded that I really should get back to that journal about my missing SI files (fortunately I still have the originals).
@Richard – not quite sure who your remarks are aimed at… have i missed a point somewhere?
@Martin – thanks for the tip-off. Can’t make Wilbanks talk since I’ll be on hols but look forward to that session at Science Online London. I gather you’re enjoying yourself in California…
Stephen; I’m trying to be subtle. People seem to think that bashing Maxine (among others) is fair game because she’s obviously a corporate shill. If you don’t think my comments apply to you, they probably don’t.
Thanks for your gallatry, Richard – I am very touched by it.
But it’s OK on this occasion anyway, I don’t feel I’ve been bashed in this column (apart from that cricket bat) – I think it’s been a useful debate and it has been focused on the issue not the personality.
Sometimes, though, I wish that we had people on NN who could speak for other journals’ editorial practices, as it would be useful I think to have a comparison. For example, if you are trying to decide which journal to submit to, it must take a long time to go into the minutiae of whether each one meets the grade in, eg (in this case) SI presentation. I don’t think we’d get very far with Ian given our “error bars” checks, either 😉
gallatry? what’s that? Typing on a dodgy keyboard and forgetting to check “preview”, sorry.
But just to reiterate, it is a bit of a hard balance to tread – I like to clarify Nature policies where it is appropriate and useful, but I don’t want to stifle any debate nor do I expect the gloves to be any more “kidded” than they are for anyone else. (I’m not going over to Science Blogs or certain other blogs outside NN, though! I’m not that brave!)
Sometimes, though, I wish that we had people on NN who could speak for other journals’ editorial practices, as it would be useful I think to have a comparison.
Sorry Maxine, just changed jobs 🙂
Seriously though: I’ve seen how a small publisher might
be easily overwhelmed bydeal with supplementary information. They probably do the best they can (here is an example of a paper with a pretty supplement).speak for other journals’ editorial practices
Yes, that would be interesting. I’ve not seen much (anything) about comparing SI in different journals.
Sounds like a job for the RIN to me? Branwen, where are you?
It would certainly be good to have the perspectives of other publishers on this issue. That said, in the case of supplementary text and figures, it shouldn’t be too hard to work out an industry standard. Nature’s practices look pretty good to me, particularly since they favour the use of the PDF format. Are there ever conferences for scientific publishers where people get together to discuss best practice or industry standards?
I had a quick gander at Cell’s guidelines: “In general, every attempt should be made to submit the Supplemental Data in a composite Word file.” Word is a commonly used program but not one that is quite as universal or platform-independent as the PDF.
J. Mol. Biol.‘s instructions make no stipulations on format.
Not a scientific poll by any means but it does point towards rather variable practice.
Lastly – does anyone provide an integrated PDF of the paper and SI? That would make my personal archiving within Papers a bit more streamlined…
Nature Cell Biology has some integrated article/SI PDFs. Not all Nature sister journals do, but they have a single integrated PDF (except in special cases, e.g. where an Excel is needed, and movies are an exception too for now). And supp legends are all under the relevant figure so no flipping back and forth either.
@Frank – i am here now. Will bring it up out our next meeting.
Thanks Sarbjit – Nature Cell Biology leading the way!
I’m obviously getting more and more sensitised to this issue. I have just now been reading a paper from the Journal of General Virology and came across a link to Supplementary Information that is hosted not by the journal but by the authors’ institutional web-site. This practice seems to be in line with the journal’s guidance:
However, the danger in this approach was illustrated when I typed the URL into my browser:
“This page cannot be found.”
Grrrr.
P.S. Just noticed the link to JMB’s site is messed up in my previous comment. Correct link is here.
@Branwen – can you explain what impact the RIN might have on this issue? A big one, I hope.
I really hope y’all are printing out these comments and saving them on paper for posterity. I’m here in the archives at Berkeley reading mountains of Melvin Calvin & Co. correspondence from the 1930s-1960s, and although some of the stuff strays from the time period or subjects I’m covering, the notes and letters reveal quite a bit about a) what the various actors found interesting; b) how they thought; c) the social rules observed among them; d) who was really friends with whom; and e) what they were doing concurrently, since they were of course all working on many projects simultaneously. I’d have a hell of a time trying to write this book without these letters and memos, and I can see why the historians worry about these websites. Page not found indeed.
Stones. The ultimate off-site backup.
I’m suddenly exceedingly grateful for onion-skin paper.
@Amy – it seems clear to me that this first flowering of the digital era is a disaster for future historians in terms of the records that will be left behind. Not only has it been the death-knell of the written letter but the plethora of file formats and short lifetime of most computer systems (and their hard-drives) is going to make their lives extraordinarily difficult.
@Richard – hmmm. But even stones have their limits…
Onion-skin, Amy? Please explain…
I think future historians will have a fascinating time. Tilly Tansey (of the Wellcome Trust Centre for History of Medicine at UCL) says that historians will make do with whatever’s left, jsut as they always have done, and we shouldn’t worry about trying to preserve everything.
Jeremy John at the British Library runs a fascinating Digital Lives programme, in recognition of these factors – as life gets increasingly digital, information is lost. I do recommend checking it out via the BL website, it is fascinating. He wrote an Essay in Nature recently about some of his work, eg in reading Bill Hamilton’s archive. Apologies for lack of links.
It is also sadly true that many scientific databases have become defunct and much valuable information has been lost. Stephen Maurer has published extensively about this.
In terms of SI, if funders have money, the best thing for it is for publicly accessible, annotated and user-friendly databases, which are regularly upgraded. This does not come cheap! But it is far better to do it this way than having all the data in different formats on different publishers’ websites, I think.
Or in Stephen’s case, not on publshers’ websites 😉
Ah, onion skin paper! – that’s my new thing for today.
This is the link to the Jeremy John essay in Nature (subscription required?) that Maxine mentioned.
I like the title of that article – The future of saving our past.
Yeah, and not just the paper, Stephen — those typewriter carbons did a fine job. You’d think that after 50 years the ink would be sliding off the paper, but no, everything’s still crisp.
@ Frank: I think Tilli’s a little cavalier. If you don’t bother with the records all you can do is push the slider towards fiction, and in academic circles the winning fictions will be chosen by some of the most boring politics imaginable.
National Public Radio, over here, is going to great lengths t try and “preserve the present for the future” as it were.
Er …. what Maxine said.
When I was giving a talk over at Frank’s place someone in the auience raised a question that seemed so far-out it had never occurred to me before – does the posibility of dumping more and more data in Supplementary Information mean that referees will use this as a way to keep on demanding more experiments and such from authors, now that the compass of a paper is much more elastic than it once was?
@Henry – so do you detect that as a trend in the manuscripts that you handle?
Mind you, perhaps there’s not much scope to ask for further experiments from paleontologists. Or do reviewers ask them to go out and find a few more fossils…?
@Stephen – no, I don’t. I think the questioner was being a bit
paranoidsuspicious.We’re all paranoid when it comes to submitting to Nature or Science… (HT Grace Baynes, via Twitter)
Anyway, am pleased to report that the Supp. Info.)00274-6 related to my Structure paper has now been re-instated!
One Giant Paper for Man, One Horrible Read for Mankind….
Well done on the restoration, Stephen!