Showing posts with label Data Re-use. Show all posts
Showing posts with label Data Re-use. Show all posts

Wednesday, 31 March 2010

Linked Data and Reality

I have a copy of the really interesting book “Data and Reality” by William Kent. It’s interesting at several levels; first published in 1978, this appears to be a “print-on-demand” version of the second edition from 1987. Its imprint page simply says “Copyright © 1998, 2000 by William Kent”.

The book is full of really scary ways in which the ambiguity of language can cause problems for what Kent often calls “data processing systems”. He quotes Metaxides:

“Entities are a state of mind. No two people agree on what the real world view is”
Here’s an example of Kent from the first page:
“Becoming an expert in data structures is… not of much value if the thoughts you want to express are all muddled”
But it soon becomes clear that most of us are all too easily muddled, at least when
“... the thing that makes computers so hard is not their complexity, but their utter simplicity… [possessing] incredibly little ordinary intelligence”
I do commend this book to those (like me) who haven’t had formal training in data structures and modelling.

I was reminded of this book by the very interesting attempt by Brain Kelly to find out whether Linked Data could be used to answer a fairly simple question. His challenge was ‘to make use of the data stored in DBpedia (which is harvested from Wikipedia) to answer the query

“Which town or city in the UK has the highest proportion of students?"
He has written some further posts on the process of answering the query, and attempting to debug the results.

So what was the answer? The query produced the answer Cambridge. That’s a little surprising, but for a while you might convince yourself it’s right; after all, it’s not a large town and it has 2 universities based there. The table of results shows the student population as 38,696, while the population of the town is… hang on… 12? So the percentage of students is 3224%. Yes, something is clearly wrong here, and Brian goes on to investigate a bit more. No clear answer yet, although it begins to look as if the process of going from Wikipedia to DBpedia might be involved. Specifically, Wikipedia gives (gave, it might have changed) “three population counts: the district and city population (122,800), urban population (130,000), and county population (752,900)”. But querying DBpedia gave him “three values for population: 12, 73 and 752,900”.

There is of course something faintly alarming about this. What’s the point of Linked Data if it can so easily produce such stupid results? Or worse, produce seriously wrong but not quite so obviously stupid results? But in the end, I don’t think this is the right reaction. If we care about our queries, we should care about our sources; we should use curated resources that we can trust. Resources from, say… the UK government?

And that’s what Chris Wallace has done. He used pretty reliable data (although the Guardian’s in there somewhere ;-), and built a robust query. He really knows what he’s doing. And the answer is… drum roll… Milton Keynes!

I have to admit I’d been worrying a bit about this outcome. For non-Brits, Milton Keynes is a New Town north west of London with a collection of concrete cows, more roundabouts than anywhere (except possibly Swindon, but that’s another story), and some impeccable transport connections. It’s also home to Britain’s largest University, the Open University. The trouble is, very few of those students live in Milton Keynes, or even come to visit for any length of time (just the odd Summer School), as the OU operates almost entirely by distance learning. So if you read the query as “Which town or city in the UK is home to one or more universities whose registered students divided by the local population gives the largest percentage?”, then it would be fine.

And hang on again. I just made an explicit transition there that has been implicit so far. We’ve been talking about students, and I’ve turned that into university students. We can be pretty sure that’s what Brian meant, but it’s not what he asked. If you start to include primary and secondary school students, I couldn’t guess which town you’d end up with (and it might even be Milton Keynes, with a youngish population).

My sense of Brian’s question is “Which town or city in the UK is home to one or more university campuses whose registered full or part time (non-distance) students divided by the local population gives the largest percentage?”. Or something like that (remember Metaxides, above). Go on, have a go at expressing your own version more precisely!

The point is, these things are hard. Understanding your data structures and their semantics, understanding the actual data and their provenance, understanding your questions, expressing them really clearly: these are hard things. That’s why informatics takes years to learn properly. Why people worry about how the parameters in a VCard should be expressed in RDF. It matters, and you can mess up if you get it wrong.

People sometimes say there’s so much dross and rubbish on the Internet, that searches such as Google provides are no good. But in fact with text, the human reader is mostly extraordinarily good at distinguishing dross from diamonds. A couple of side searches will usually clear up any doubts.

But people don’t do data well. Automated systems do, SPARQL queries do. We ought to remember a lot more from William Kent, about the ambiguities of concepts, but especially that bit about computers possessing incredibly little ordinary intelligence. I’m beginning to worry that Linked Data may be slightly dangerous except for very well-designed systems and very smart people…

Wednesday, 10 June 2009

Need to read a CT scan from an old Magneto-optical disk

Philipp Westebbe from Berlin wrote to me:
“I am working at Rathgen Research Laboratory (National Museums in Berlin) in Germany. I have some valuable old data (computer tomography images) on MOD [Magneto-optical disc – CR] format. I am aware this is not the usual topic of the list, but we cannot think of a better place to start with. I am working on a project using an old MOD (Laser Drive Media LM1200-002) for recording computer tomography images. I'm looking for an old PC and the belonging software with which I can read this old MOD from 1992.”
In a later email, he wrote:
“At the moment I am coordinating the project about the Egyptian Queen Nefertiti bust [see Wikipedia- CR]. 1992 the Charite hospital in Berlin made a CT from the bust to analyse the internal structure. The CT images were saved at that time on a MOD LM 1200-002. The data are unfortunately no longer readable, because the Charite hospital doesn't have the specific computer any more. The only things I have in my research lab are the drive and the MOD with Nefertiti's scan on it. I got over long paths to your email. I am looking for someone who has such a computer with which I can read my old MOD from Nefertiti, and for a possibility to copy this old format into a new digital format. Nevertheless, I hope to get a contact, which helps me to read the old MOD from Nefertiti. It would be very nice if the person who reads this mail could help me handling my problem, or could send the email to someone who is still working with such an old CT system. It is quite difficult to get an old MOD from 1992 back into a common digital format.”
In a third email, he made clear that he has one disk, on which the CT scans are stored, and that the data are regarded as very valuable, as the bust is too fragile to scan again. He didn’t say, but perhaps the importance (and hence the value) of these data is increased by a current controversy over the authenticity of the bust.

Documents from NARA and other places show that the LM1200 disk is 12” Write-Once, Read-Many (WORM) dual-sided media, capable of holding 1 GByte each side. A catalogue from Siemens identifies it as Plasmon (Philips) WORM Media. These media are reputed to have a long life expectancy.

As far as I can tell, this drive will have been part of an embedded computer in the CT machine, although we don’t at this stage know the brand of the CT machine (or the type of embedded computer; many such machines of that vintage would have been “mini-computers” rather than PCs). The UK-based Plasmon PLC bought Philips LMS, the company that produced Philips 12” optical disks and drives, in 1999. It looks like Plasmon failed at the end of last year, and was in its turn bought by Alliance Storage Technologies in the US.

Their technical support points out that reading the data off the disk may not be enough, because of the proprietary formats involved at the time. “Your best bet is to contact a data migration company that has experience with the application that wrote the data or the company that wrote the software. If you have trouble reading the media because of a problem with your drive we may be able to help but there are few parts left for the drive… Sorry to provide such a bleak outlook. Medical companies tend to maintain legacy equipment the longest so there is some hope.”

Nevertheless, before there is a chance of interpreting the data, Phillip needs to get it off the disk. Does anyone know of a contact who could read and interpret such a disk?

Wednesday, 18 February 2009

Notes from NERC Data Management workshop 1

David Bloomer, NERC CIO (and Finance Director) opened the workshop, and talked about data acquisition, data curation, data access, and data exploitation, in the context of developing NERC Information Strategy. Apparently NERC does not currently have an Information Strategy, as the last effort was thrown out in Council. Clearly from his point of view, the issue was about working out whether data management is being done well enough, and how it can be done better within the resources available.

There were some interesting comments about licensing and charging: principles that I summarise as:
  • encouraging re-use and re-purposing,
  • all free for teaching and research,
  • that the licence and its cost depends on the USE and not the USER, and
  • that NERC should support all kinds of data users equally.
Not sure yet the full implications of this; it clearly doesn’t mean that all data is freely accessible to everyone! However, it sounds like a major improvement over recent practices, with some NERC bodies charging high prices for some of their data.

In the second session, after my talk which I have already blogged (although not the 20 minutes of panic when first the PowerPoint transferred from my Mac would not load up on Windows, and second my Mac would not recognise the external video for the projectors!), there were two short talks by Dominic Lowe from BADC on Climate Science Modelling Language, and Tim Duffy from BGS on GeoSciML.

My notes on the former are skeletal in the extreme (ie nonsense!), other than that it is an application of GML, and is based on a data model expressed with UML. However, I picked up a bit more about the latter.

The scope of GeoSciML is the scientific information linked to the geography, except the cartography. It is mainly interpretive or descriptive, and so far includes 25 internationally-agreed vocabularies. Taken 5 years or so to develop to this point. Based around GML, using UML for modelling, using OGC web services to render maps, format, query etc. Provides interoperability and perhaps a “unified” view, but does not require change to local (end) systems, nor to local schemas. Part of the claimed value of the process was exposing that they did not understand their data as well as they thought!

Tuesday, 17 February 2009

Data Curation for Integrated Science: the 4th Rumsfeld Class!

I had to talk today to a workshop of NERC (Natural Environment Research Council) Data Managers, on data curation for integrated science. Integrated science does seem to be a term that comes up a lot in environmental sciences, for good reasons when contemplating global change. However, there doesn’t seem to be a good definition of integrated science on the web, so I had to come up with my own for the purposes of the talk: “The application of multiple scientific disciplines to one or more core scientific challenges”. The point of this was that scientists operating in integrated science MUST be USING unfamiliar data. Then the implications for data managers, particularly environmental data managers, were that they must make their data available for unfamiliar users. What does this imply?

By some strange mental processes, and a fortuitous Google search, this led me to the Poetry & Philosophy of Donald H Rumsfeld, as exposed to the world by Hart Seely, initially in Slate in April 2003, and now published in a book, “Pieces of Intelligence”, by Seely. These poems are well worth a look for their own particular delight, but the one I was looking for you will probably have heard of in various guises, The Unknown:
‘As we know,
There are known knowns.
There are things we know we know.
We also know
There are known unknowns.
That is to say
We know there are some things
We do not know.
But there are also unknown unknowns,
The ones we don't know
We don't know.’
—Feb. 12, 2002, Department of Defense news briefing
Now this insightful (!) poem (set to music by Bryant Kong, available at http://www.stuffedpenguin.com/rumsfeld/lyrics.htm) perhaps defines 3 epistemological classes:
Known knowns,
Known unknowns, and
Unknown unknowns
Logically there should be a 4th Rumsfeld class: the unknown knowns. And I think this class is especially important for data management for unfamiliar users.

The problem is that in many research projects, there are too many people who “know too much”; with so much shared knowledge, much is un-documented. In OAIS terms, we are looking here at a small, tight Designated Community with a shared Knowledge Base, and consequently little need for Representation Information. In integrated science, and particularly environmental sciences, as the Community broadens and time extends, effectively the need for Representation Information increases. I’m using the terms very broadly here, and RepInfo can be interpreted here in many different ways. But the requirement is to make more explicit the tacit knowledge, the unknown knowns implicit in the data creation and acquisition process.

Interestingly, subsequent speakers seemed to pick up on the idea of making explicit the unknown knowns, so maybe the 4th Rumsfeld Class is here to stay in NERC!

Sunday, 23 November 2008

Edinburgh IT futures workshop

I spent Friday morning at a workshop entitled “Research, wRiting and Reputation” at Edinburgh University. Pleasingly, data form quite a large part of the programme (although I’ll miss the afternoon talks, including Simon Coles from Southampton). First speaker is Prof Simon Tett of Geosciences talking about data curation and climate change. Problem is that climate change is slow, relative to changes in the observation systems, not to mention changes in computation and data storage. Baselines tend to be non-digital, eg 19th century observations from East India company shipping: such data need to be digitised to be available for computation. By comparison, almost all US Navy 2nd World War meteorological log books were destroyed (what a loss!). He’s making the point that paper records are robust (if cared for in archives) but digital data are fragile (I have problems with some of these ideas, as you may know: both documents and digital data have different kinds of fragility and robustness… the easy replicability of data compared with paper being a major advantage). Data should be looked after by organisations with a culture of long term curatorship: not researchers but libraries or similar organisations. The models themselves need to be maintained to be usable, although if continually used they can continue to be useful (but not for so long as the data). Not just the data, but the models need curation and provenance.

Paul Anderson of Informatics talking about software packages as research outputs. His main subject is a large scale UNIX configuration system, submitted by Edinburgh as part of the RAE. It’s become an Open Source system, so there are contributions from outside people in it as well. Now around 23K lines of code. Interesting to be reminded how long-lived, how large such projects can be, and how magically sustainable some of the underlying infrastructure appears to be. What would happen if SourceForge or equivalents failed? However, it does strike me that there are lessons to be earned by the data curation community from the software community, particularly from these large scale, long lived open source projects. Quite what these lessons are, I’m not sure yet!

Four speakers followed, three from creative arts (music composition and dance), and the fourth from humanities. The creative arts folks partly had problems with the recognition (or lack of recognition) of their creative outputs as research by the RAE panels. Their problems were compounded by a paucity of peer-reviewed journals, and the static, paper-oriented nature even of eJournals. The composer had established a specialist music publisher (Sumtone), since existing publishers are reluctant to handle electronic music. Interestingly, they offer their music with a Creative Commons licence.

The dance researchers had different problems. Complexity relates to temporality of dance. It’s also a young discipline, only 25 years or so, whereas dance is a very ancient form of art (although of course documentation of that is static). Dance notation is very specialised; many choreographers and many dancers cannot notate! Only a very small set of journals are interested, few books etc. Lots of online resources however. Intellectual property and performance rights are major issues for them, however.

The final researcher in this group was interested in integrating text with visual data and multimodal research objects. Her problems seemed to boil down to limitations of traditional journals in terms of the numbers and nature (colour, size, location etc) of images allowed; these restrictions themselves affected the nature of the argument she was able to make.

OK, these last few are at least partly concerned at the inappropriateness of static journals for their disciplines. Even 99% of eJournals are simply paper pages transported in PDF over the Internet. Why this still should be, 12 years after Internet Archaeology started publishing, as a peer-reviewed eJournal specifically designed to exploit the power of the Internet to enhance the scholarly content, beats me! Surely, just after the last ever RAE is exactly the time to create multimedia peer-reviewed journals to serve these disciplines. They’ll take time to get established, and to build up their citations and impact-factor (or equivalent). Does the new REF really militate so strongly against this? What else have I missed?

Thursday, 20 November 2008

Comments on an obsolescence scale, please

Many people in this business know the famous Jeff Rothenberg quote: “It is only slightly facetious to say that digital information lasts forever--or five years, whichever comes first” (Rothenberg, 1995). Having just re-read the article, it’s clear that at that point he was particularly concerned about the fragility of storage media, something that doesn’t concern me for this post. What does concern me is the subject of much of the rest of his article: obsolescence of the data. There’s another very nice quote from the article, accepted gospel now but put so nicely: “A file is not a document in its own right--it merely describes a document that comes into existence when the file is interpreted by the program that produced it. Without this program (or equivalent software), the document is a cryptic hostage of its own encoding” (ibid).

Now the first Rothenberg quote is often assumed to apply in the second case, but there’s clear evidence emerging that data are readable well beyond 5 years. Most files on my hard disk that have been brought forward from previous computers (with very different operating systems) remain readable even after 15 years. Most, but not all; in my case, the exceptions are my early PowerPoint files, created with PowerPoint 4 on a Mac in the 1990s, unreadable with today’s Mac PowerPoint version. However, these files are not completely inaccessible to me, as a colleague with a Windows machine has a different version of PowerPoint that can read these files and save them in a format that I can handle. It’s a comparatively simple, if tedious, migration.

I have spent some time and asked in various fora for evidence of genuinely obsolete, that is completely inaccessible data (Rusbridge, 2006). There are some apocryphal stories and anecdotes, but little hard evidence so far. But perhaps complete inaccessibility isn’t the point? Perhaps the issue is more about risk to content, on the one hand, or extent of information loss on the other? Maybe this isn’t a binary issue (inaccessible or not), but a graded obsolescence scale? Is there such a thing? I couldn’t find one (there’s a hint of one in a NTIS report abstract (Slebodnick, Anderson, Fagan, & Lisez, 1998), but the text isn’t online, and it appears to be about obsolescence of people)!

So here are a couple of attempts, for comment.

Here’s what might be called a Category approach. Users would be asked to select one only from the following categories (it worries me that they are neither complete nor non-overlapping!):
A Completely usable, not deemed at risk
  • Open definition with no significant proprietary elements
  • multiple implementations, at least one Open Source
  • high quality treatment in secondary implementations
  • widespread use
B Currently usable, at risk
  • Closed definition, or
  • Open definition, but with significant proprietary elements
  • More than one implementation
  • imperfect treatment in secondary implementations
C Currently usable, at significant risk
  • Proprietary/closed definition
  • single proprietary implementation
  • Not widespread use
D Currently inaccessible, migration path known
  • Proprietary tools don't run on current environment
  • Potentially imperfect migration
E Completely inaccessible
  • no known method of extraction or interpretation
Here’s a more subjective, but perhaps more complete scale. Users are asked to select a number from 1 to 5 representing where they assess their data lies, on a scale whose ends are defined as
1 Completely usable, not deemed at risk, and
5 Completely inaccessible
And here’s a more Likert-like scale. Users are asked to nominate their agreement with a statement such as “My data are completely inaccessible”, using the Likert scale:
1 Strongly disagree
2 Disagree
3 Neither agree nor disagree
4 Agree
5 Strongly Agree
Hmmm. How could you neither agree nor disagree with that statement? “Excuse me, I haven’t a clue”?

Comments on this blog, please; but if you’re commenting elsewhere, can you use the tag “Obsolescence scale”, and maybe drop me a line? Thanks!

Rothenberg, J. (1995). Ensuring the longevity of digital documents. Scientific American, 272(1), 42. (Accessed through Ebscohost)
Rusbridge, C. (2006). Excuse Me... Some Digital Preservation Fallacies? [Electronic Version]. Ariadne from http://www.ariadne.ac.uk/issue46/rusbridge/.
Slebodnick, E. B., Anderson, C. D., Fagan, P., & Lisez, L. (1998). Development of Alternative Continuing Educational Systems for Preventing the Technological Obsolescence of Air Force Scientists and Engineers. Volume I. Basic Study.

Monday, 17 November 2008

Just sharing

Like several others in the blogosphere, I really enjoyed Scott Leslie’s post on EdTechBlog: Planning to Share versus Just Sharing. It is in the learning domain but I was glad that Andy Powell picked up its relevance to repositories in his follow-up on eFoundations.

Perhaps the key points that Andy picks up are
“The institutional approach, in my experience, is driven by people who will end up not being the ones doing the actual sharing nor producing what is to be shared. They might have the need, but they are acting on behalf of some larger entity.” and

“because typically the needs for the platform have been defined by the collective’s/collaboration’s needs, and not each of the individual users/institutions, what results is a central “bucket” that people are reluctant to contribute to, that is secondary to their ‘normal’ workflow…”
I’ve written before about fitting into the work flow, and endorse this concern.

I would also point to this quote:
“the whole reason many of us are so attracted to blogs, microblogs, social media, etc., in the first place is that they are SIMPLE to use and don’t require a lot of training.”
I suppose the equivalent for the repository world’s sphere of interest (usually, the scholarly article) is the academic’s web page. Somehow, for many (not all) academics, putting papers up on their web page does seem to fit in their normal work flow. They don’t worry about permissions, about the fine print of agreements they didn’t even read, they just put up their stuff to share. Is that a Good Thing? You betcha! Is it entirely a Good Thing? No, there are problems, for example those that arise in the medium term, when the academic moves, retires, or dies.

Looking at it another way, however, why doesn’t the collective repository infrastructure get credit for “Just Sharing”?

What about data? Well, getting data to be available for re-use is a Good Thing, and if there’s a simple infrastructure (eg the academic’s web page), that’s certainly better than leaving it to languish and fade away on some abandoned hard drive or (worse) CD-ROM. My feelings of concern are certainly elevated for data, however; don’t repositories, data services etc act as domain proxies in encouraging appropriate standardization? Well, I suppose they might, but maybe Leslie is right that the intra-domain pressures that would arise from Just Sharing would be even stronger. “It would be much easier for me to re-use your data (and give you a citation for it) if it were consistent with XXXX” is a stronger motivator than “as repository managers, we believe these data should be encoded with XXXX”.

Data repositories (and they can be at the lab or even personal level) should, of course, encourage the collection of contextual and descriptive information alongside the actual data. And again, their potential longevity brings advantage later, as technology change begins to affect usability of the data.

Now I guess the cheap shot here is to point to UKRDS as an example that is spending years Planning to Share, rather than Just Sharing. But what is the Just Sharing alternative to UKRDS?

Friday, 19 September 2008

A national data mandate? Australian Code for the Responsible Conduct of Research

Andrew Treloar pointed to this Code in a presentation at the e-Science All Hands meeting. All Australian Universities have signed up to the Code, which turns out to have a whole chapter on the management of research data and primary materials. It deals with the responsibilities of both institutions and researchers. Mostly it’s in the style “Each institution must have a policy on…”, but it then gets quite prescriptive on what those policies must cover. Here are some quotes, institutional responsibilities first:
“Each institution must have a policy on the retention of materials and research data.”

“In general, the minimum recommended period for retention of research data is 5 years from the date of publication.”

“Institutions must provide facilities for the safe and secure storage of research data and for maintaining records of where research data are stored.”

“Wherever possible and appropriate, research data should be held in the researcher’s department or other appropriate institutional repository, although researchers should be permitted to hold copies of the research data for their own use.”
On researcher responsibilities:
“Researchers should retain research data and primary materials for sufficient time to allow reference to them by other researchers and interested parties. For published research data, this may be for as long as interest and discussion persist following publication.”

“Research data should be made available for use by other researchers unless this is prevented by ethical, privacy or confidentiality matters.”

“Retain research data, including electronic data, in a durable, indexed and retrievable form.”
To discharge these obligations requires training in good curation practice, significant care, and appropriate infrastructure. Maybe these are regarded as yet more "un-funded mandates", to be treated on a risk assessment basis (will we be found out?). Maybe this does represent "the dead hand of compliance", as a senior colleague once phrased it. But if taken in the spirit as written, it represents a significant mandate for data curation!

I don’t know of an equivalent Code elsewhere that is so specific. The nearest US equivalent may be the Introduction to the Responsible Conduct of Research, from the Office of Research Integrity, Dept of Health & Human Services. It tends to be more a rather bland assembly of good advice, than anything prescriptive.

In the UK, the Research Integrity Office says it is “developing a code”, individual Research Councils have Ethics and related policies, while Research Councils UK is consulting in this area; their consultation closes on 24 October, 2008. It does have a section on Management and preservation of data and primary materials:
“… ensure that relevant primary data and research evidence are preserved and accessible to others for reasonable periods after the completion of the research. This is a shared responsibility between researcher and the research organisation, but individual researchers should always ensure that primary material is available to be checked. … Data should normally be preserved and accessible for not less than 10 years for any projects, and for projects of clinical or major social, environmental or heritage importance, the data should be retained for up to 20 years, and preferably permanently within a national collection, or as required by the funder’s data policy.”
The “normal” period is longer, but otherwise, still not as specific and therefore strong as the Australian Code!

Thursday, 11 September 2008

Many citations flow from data...

I've been at the UK e-Science All Hands Meeting in Edinburgh over the past few days (easy, since it's being held in the bulding in which I work!). Lots of interesting presentations; far too many to go to, let alone blog about. But I can't resist mentioning one short presentation (PPT), from Prof Michael Wilson of STFC. His pitch was simple: publishing data is good for your career, especially now. And he has evidence to back up his claims!

Michael and colleagues put together a Psycholinguistic database many years ago, with funding from the MRC. At first (1981) it was available via postal request, then (1988) via ftp from the Oxford Text Archive, now via web access from STFC and from UWA(since 1994). Not much change over the years, little effort, free data, no promotion.

The database is now publicly available, eg at the link above. You may see that users are requested to cite the relevant paper:
Wilson, M.D. (1988) The MRC Psycholinguistic Database: Machine Readable Dictionary, Version 2. Behavioural Research Methods, Instruments and Computers, 20(1), 6-11.
The vital piece of evidence was a plot of citations over the years (data extracted from Thomson ISI, I hope they don't mind my re-using it):

You'll notice that citations flowed very slowly in the early days, picked up a little after ftp access was available (it takes a few years to get the research done and published), and then really started to climb after web access was provided. Now Michael and his colleagues are getting around 80 citations per year!

To ram home his point, Michael did some quick investigations, and found "At least 7 of the top 20 UK-based Authors of High-Impact Papers, 2003-07 ranked by citations ... published data sets".

There was some questioning on whether citing data through papers (rather than the dataset itself) was the right approach. Michael is clear where he is on this question: paper citations get counted in the critical organs by which he and others will be measured, so citations should be of papers.

Summary of the ptich: data publishing is easy, it's cheap, and it will boost your career when the new Research Evaluation Framework comes into effect.

Wednesday, 20 August 2008

Adding value to data: eScience conference session

If you nearly had a paper ready for the International Digital Curation Conference in December (the closing date was the end of July), you may still have time to get one in to the special session "Adding value to data – Digital Repositories in the e-Science world" at the 4th IEEE International Conference on eScience, whose deadline has just been extended to 31 August. From the call for papers on the DReSNet blog:
"There is a great, untapped potential for synergies between grid/e-science technologies and a cluster of related systems addressing the management of digital assets in digital libraries and repositories. The digital material generated from and used by academic and other research is to an increasing extent being held in formal data management systems; these systems are variously categorized as digital repositories, libraries or archives, although the distinction between them relates more to the sort of data that they contain and the use to which the data is put, rather than to any major difference in functionality. In many cases, these systems are used currently to hold relatively simple objects, for example an institution’s pre-prints and publications, or e-theses. However, some institutions are beginning to use them to manage research data in a variety of disciplines, including physical sciences, social sciences, and the arts and humanities, as well as the output from various digitisation programmes."
This does look very interesting. They suggest several research challenges, which I think tie in pretty much with our own agenda, although the last is novel to me in the way it's framed:
  • Digital preservation and curation in research infrastructures
  • Interoperability
  • Security
  • Provenance
  • Metadata Extraction
  • Workflow Integration
  • Architecture of Participation
There's a good programme committee, so the session could be extremely useful. Go to their entry for the submission details...

Tuesday, 19 August 2008

Credit from citing datasets?

Cameron Neylon has a thought-provoking post on his Science in the Open blog, arising from a discussion he had at Scifoo with Michael Eisen of Berkeley:
"Michael [...] felt people got too much credit for datasets already and that making them more widely citeable would actually devalue the contribution. The example he cited was genome sequences. This is a case where, for historical reasons, the publication of a dataset as a paper in a high ranking journal is considered appropriate.

In a sense I agree with this case. The problem here is that for this specific case it is allowable to push a dataset sized peg into a paper sized hole. This has arguably led to an over valuing of the sequence data itself and an undervaluing of the science it enables. Small molecule crystallography is similar in some regards with the publication of crystal structures in paper form bulking out the publication lists of many scientists. There is a real sense in which having a publication stream for data, making the data itself directly citeable, would lead to a devaluation of these contributions. On the other hand it would lead to a situation where you would cite what you used, rather than the paper in which it was, perhaps peripherally described. I think more broadly that the publication of data will lead to greater efficiency in research generally and more diversity in the streams to which people can contribute."
This seems a strong argument to me. A paper describing a research dataset isn't really a research paper, surely? So it is a round peg in a square hole. But if the alternative is that the dataset can only be cited via a research paper (on some other, probably related topic) that mentions it in passing, then this is likely to be a rather poor proxy. The dataset creator may get little credit, and the research paper authors rather more credit, than either was due.

However, if we move to the situation where more datasets are cited directly, then not only do the dataset creators or providers get the credit that's due to them, but also the credit is the right kind. That is, the citation is recognisably for creating or providing a dataset and not for a specific research contribution. I suppose that papers in the Nucleic Acids Review special issue on datasets are also recognisable as creation/provision papers, not research papers, but few other disciplines have such easily recognisable distinctions. So overall simply using data citations looks like a better bet. As Cameron says:
"So to come back around to the original point, the value of different forms of contribution is not due to the fact that they are non-traditional or because of the medium per se, it is because they are different."
Amen to that!

Wednesday, 16 July 2008

JIF08 Technical Infrastructure session

At the second day of the JISC Innovation Forum, I attended an interesting discussion in the data theme on technical infrastructures. This post derives from that discussion, but is neither a complete reflection of what was said, nor at all in the order of discussion; it reflects some bits I found interesting. My thanks to Matthew Dovey who chaired, and to all who contributed; I’ve identified none rather than only some! The session is also being blogged here…

Although early on we were talking about preservation, we came back to a discussion on immediate re-use and sharing; it seems to me that these uses are more powerful drivers than long term preservation. If we organise things right, long term preservation could be a cheap consequence of making things available for re-use and sharing. Motivations quoted here include data as evidence, the validation aspects of the scientific method. We were cautioned that validation from the data is hard; you will need the full data and analysis chain to be effective (more data, providing better provenance…).

Contextual metadata might include (parts of) the original proposal (disclosing purpose and motivation). Funders keep these proposals, as part of their records management systems, but they might not be easily accessible and are likely eventually to be deleted. An action for JISC might be to find ways of making parts of these proposals more appropriately available to support data.

There are issues of the amount of disclosure or discoverability that people want, or maybe should offer whether they want or not; we touched tangentially on open notebook science. Scary, but so is sharing data, for some! Extending re-usability through re-interpreting to integration may be big steps.

To what extent does standardisation help? Well, quite a lot, obviously, although science is slow and science is innovative, so many researchers will either have no appropriate standards to se, or not know about the standards, or have poor or no tools to implement/use the standards, or maybe find the standards less than adequate and overload them with their own private extensions or encodings. And to some extent re-use could be about the process as much as the data (someone spoke of preserving workflows, and someone of preserving protocols, although “process” may be much less formal than either…).

We had a recurring discussion about the extent this feeds back to being a methodological question. There was a question about the efficacy of research practice; how do we maximise positive outcomes? Should we aim change science to achieve better curation? While in the large scale this sounds exceedingly unlikely, research being exceedingly resistant to external change, successes were reported. BADC have managed to get scientists to change their methods to collect better metadata, and maybe better data. Maybe some impact could be made via research training curricula? Another opportunity for JISC?

We talked about the selection of useful data; essential, but how can it be achieved? Some projects, such LHC have this built into the design as the volumes are orders of magnitudes too large to deal with otherwise. While low-retention selection might be attractive, others were arguing elsewhere for retaining more data that are currently lost, to improve provenance (ie successive refinement stages).

The idea of institutional data audit was mentioned; there is a pilot data audit framework being developed now, and at least one institution in the room was participating in the pilot process. This might be a way of bringing issues to management attention, so that institutions can understand their own needs. Extending this more widely may be useful.

We talked about what a research library could do, or maybe has to start thinking of doing. On the small scale, success in digital thesis deposit is bringing problems in managing associated data files (often of widely varied types: audio, video, surveys, spreadsheets, databases, instrumentation outputs, etc). At the moment the answer is often to ZIP up the supplementary files and cross your fingers (as close to benign neglect as possible!). This is very close to the approach in use for years with paper theses, where supplementary materials were often in portable media in a pocket inside the back cover… close the volume and put it on the shelves. This might be acceptable with a trickle, but not a deluge… What would be a better approach in dealing with this growing problem (exacerbated by issues related to the well-known neologism problems of theses, ie that young researchers will likely not use applicable standards in quite the right way)? It was clear that dealing with this stuff is beyond the expertise of most librarians, and requires some sort of partnership with domain scientists. This whole area could represent an opportunity for JISC to help.

Since we were thinking of institutional repositories and larger scale subject repositories, we had a skirmish on the extent to which we need a full-blown OAIS, as it were, or a minimal, just-enough effort. It seemed in a sense the answer was: both! It does make sense to think carefully about OAIS, but also to make sure you do just enough, to ensure that high preservation costs in one area are not preventing collection in other areas (selection anyway being so fallible). A good infrastructure should help keep it cheap. There is a question on how you would know if you were paying too much in effort, or too little? Perhaps repository audit and certification, or perhaps risk assessment might help.

This brought us on to the issue of tool development. Most existing, large scale data centres use home-built, filestore-based systems, not very suitable for small-scale use. Existing repository software platforms are not well suited to data (square peg in round hole was quoted!). Funding developments to improve the fit for data was seen as a possible role for JISC. Adding some better OAIS capabilities at the same time might also be useful, as might linking to Virtual research Environments such as Sakai, or to Current Research Information Systems (CRIS’s). Is the CERIF standard developed by EuroCRIS helpful, or not?

Overall, it was a useful session, and if JISC takes up some of the opportunities for development suggested, it should prove doubly, trebly or even more useful!

Wednesday, 9 July 2008

Semantic web on the Today programme

Turning on the radio this morning, I was surprised to hear someone discussing data on the Today programme on BBC Radio 4. It turned out to be Sir Tim Berners Lee talking about the semantic web; he even managed to mention RDF and HTML without confusing the interviewers too much. The interesting 8 and a half minute discussion is available via the BBC iPlayer.

I'm quite keen on understanding better how the semantic web might relate to science data. When people talk about data in relation to the semantic web, they often seem to be thinking the sort of relatively unitary facts or simple relationships that RDF triples are quite good at expressing, such as the population of China is 1,330,044,605. It's certainly not clear to me how this generalises to express the changing population, let alone how it could express the data from a spectrometer or
a crystal structure determination or remote-sensing data. If anyone can point me to a good resource discussing this, I would be grateful!

Wednesday, 25 June 2008

4th International Digital Curation Conference - Radical Sharing: Transforming Science?

This is a reminder that closing date for submission of full papers, posters and demos for the conference (details below) is 25 July 2008 . We invite submissons from individuals, organisations and institutions across all disciplines and domains engaged in the creation, use and management of digital data, especially those involved in the challenge of curating data in e-Science and e-Research. Templates for submissions and other details are available at http://www.dcc.ac.uk/events/dcc-2008/ .

The 4th International Digital Curation Conference will be held on 1-3 December 2008 at the Hilton Grosvenor Hotel, Edinburgh, Scotland, UK. The conference will be held in partnership with the National e-Science Centre and supported by the Coalition for Networked Information (CNI).

We are sure there are plenty of you out there with great ideas to share with colleagues; still time to get writing!

Sunday, 4 May 2008

Tranche

Tranche sounds interesting. This possibly over-ambitious sound-bite from its web page:
"This project's goal is to solve the problems commonly associated with sharing scientific data, letting you and your collaborators focus on using the data."
In effect it is an encrypted, highly distributed file sharing system, independent of any central authority, and suitable for "any size of file", but maybe there are still some sharing problems left for the scientists (;-)?

The Science Commons blog reports:
"The National Cancer Institute will soon be using Tranche to store and share mouse proteomic data from its Mouse Proteomic Technologies Initiative (MPTI). Tranche, a free and open source file sharing tool for scientific data, was one of the earliest testers of CC0. Many thanks to Tranche for providing us with such valuable early feedback on CC0."
Tranche uploads are known as projects, and there are apparently 5399 of them so far. The largest number of replicas appears to be 3, and 44 projects are reported as having missing chunks.

Tuesday, 1 April 2008

National Statistics no joke

Today the UK's new independent National Statistics Authority began work, replacing the Office for National Statistics. This is clearly an important agency; in the Today Programme this morning they were claiming that they would make a clear distinction between the statistics and political interpretation. I thought it was worth a look at their web site.

Two things initially discouraging: first, a Google search for National Statistics Agency [sic] produces a web page for the old ONS at http://www.statistics.gov.uk/default.asp; this is ALMOST but not quite the same URL as the new authority at http://www.statistics.gov.uk/... but what about the apparently similar but possibly different http://www.statisticsauthority.gov.uk/? Well, teething troubles no doubt. Second there's a prominent link on the (first of the above) home page that "ONS independence comes into effect on 1 April", and the link is broken. More teething...

A quick explore led me to the UK snapshot, with lots of interesting web pages summarising data. As an example, there is a page Headed "Acid Rain" under the "Environment" section; you get a graph and a few paragraphs of text, eg "Emissions of chemicals that can cause acid rain fell by 53.8 per cent between 1990 and 2005, from 6.9 million tonnes to 3.2 million tonnes." In fact, the page doesn't tell us whether rain became less acidic during this period, but again that's a quibble.

But I was looking for data, not summaries. I found some under "Time Series data", but had to go through a complicated sequence of selections to find an actual dataset. I selected Share Ownership, then "Total market value by sector of beneficial owner: end-2006", then "DEYQ, SRS: Ungrossed: Total Market Value:Individuals" before I got a download button. The 3 download options were:
  • View on Screen
  • Download CSV
  • Download Navidata
Where Navidata is a tool they make available. View on screen gives me a nicely formatted web page displaying a two-column table. Downloading the CSV gave me a cryptically named CSV file, which looks as follows (Crown Copyright, from the national Statistics Authority, Reproduced under the terms of the Click-Use Licence):
,"DEYQ",
" 1998",154.8,
" 1999",163.3,
" 2000",181.0,
" 2001",148.4,
" 2002",104.0,
" 2003",136.0,
" 2004",122.3,
" 2005",..,
" 2006",155.8,

"DEYQ","SRS: Ungrossed: Total Market Value:Individuals"
,"Not seasonally adjusted"
,"Updated on 8/ 6/2007"
All fine, I think... except I could not find any way to do this automatically. Maybe they have an API I haven't found, maybe they have plans not yet come to fruition. Anyway, some good stuff here, but perhaps room for improvement?

Monday, 31 March 2008

Repositories for scientists

Nico Adams in a post to Staudinger's Semantic Molecules has added to the scenarios for repositories for scientists:
"Now today it struck me: a repository should be a place where information is (a) collected, preserved and disseminated (b) semantically enriched, (c) on the basis of the semantic enrichment put into a relationship with other repository content to enable knowledge discovery, collaboration and, ultimately, the creation of knowledge spaces. "
The linkage seems to be the important thing, which current repositories don't do well. He goes on:
"Let’s take a concrete example [CR: scenario, I think] to illustrate what I mean: my institutional repository provides me with a workspace, which I can use in my scientific work everyday. In that workspace, I have my collection of literature (scientific papers, other people’s theses etc.), my scientific data (spectra, chromatograms etc) as well as drafts of my papers that I am working on at the moment. Furthermore, I have the ability to share some of this stuff with my colleagues and also to permanently archive data and information that I don’t require for projects in the future."
He then develops this argument further, making a lot of sense, and concludes:
"Now all of the technologies for this are, in principle, in place: we have DSpace, for example, for storage and dissemination, natural language processing systems such as OSCAR3 or parts of speech taggers for entity recognition, RDF to hold the data, OWL and SWRL for reasoning. And, although the example here, was chemistry specific, the same thing should be doable for any other discipline. As an outsider, it seems to me that what needs to happen now, is for these technologies to converge and integrate. Yes it needs to be done in a subject specific manner. Yes, every department should have an embedded informatician to take care of the data structures that are specific for a particular discipline. But the important thing is to just make the damn data do work!"
I have heard it suggested elsewhere that DSpace may not be up to this kind of linkage, and also the FEDORA is a suitable candidate, but I don't have sufficient experience of the workings of either to be sure. Anyone comment?

Tuesday, 25 March 2008

Pulsating white dwarf stars!

I really like the physics arXiv blog, which has quirky posts on interesting stuff recently posted to the arXiv. This post (New type of pulsating star discovered) in particular interests me, in this case from a digital curation rather than a general interest point of view. The post starts:
"New types of stars aren’t found very often but last year, Patrick Dufour and pals discovered several white dwarfs with carbon atmospheres. Before then white dwarfs were thought to come in two flavours: with atmospheres dominated by either hydrogen or helium. Astronomers suddenly had a new toy to play with.

"Dufour found nine examples of his carbon dwarfs in the data regurgitated by the Sloan Digital Sky Survey and more are likely to be found as the skies continue to be searched."
What's more they worked out that some of these new stars should be pulsating. An update suggests they found one shortly afterwards. A nice example of science emerging from the data; no doubt there are many more in astronomy, with these huge sky surveys and virtual observatories!

Tuesday, 18 March 2008

Novartis/Broad Institute Diabetes data

Graham Pryor spotted an item on the CARMEN blog, pointing to a Business Week article (from 2007, we later realised) about a commercial pharma (Novartis) making research data from its Type 2 Diabetes studies available on the web. This seemed to me an interesting thing to explore (as a data person, not a genomics scientist), both for what it was, and for how they did it.

I could not find a reference to these data on the Novartis site, but I did find a reference to a similar claim dating back to 2004, made in the Boston Globe and then in some press releases from the Broad Institute in Cambridge, MA, referring to their joint work with Novartis (eg initial announcement, first results and further results). The first press release identified David Altshuler as the PI, and he was kind enough to respond to my emails and point me to their pages that link to the studies and to the results they are making available.

Why make the data available? The Boston Globe article said "Commercially, the open approach adopted by Novartis represents a calculated gamble that it will be better able to capitalize on the identification of specific genes that play a role in Type 2 diabetes. The firm already has a core expertise in diabetes. Collaborating on the research will give its scientists intimate knowledge of the results."

The Business Week article said "...the research conducted by Novartis and its university partners at MIT and Lund University in Sweden merely sets the stage for the more complex and costly drug identification and development process. According to researchers, there are far more leads than any one lab could possibly follow up alone. So by placing its data in the public domain, Novartis hopes to leverage the talents and insights of a global research community to dramatically scale and speed up its early-stage R&D activities."

Thus far, so good. Making data available for un-realised value to be exploited by others is at the heart of the digital curation concept. There are other comments on these announcements that cynically claim that the data will have already been plundered before being made accessible; certainly the PIs will have first advantage, but there is nothing wrong with that. The data availability itself is a splendid move. It would be very interesting to know if others have drawn conclusions from the data (I did not see any licence terms, conditions, or even requests such as attribution, although maybe this is assumed as scientific good practice in this area).

Business Week go on to draw wider conclusions:
"The Novartis collaboration is just one example of a deep transformation in science and invention. Just as the Enlightenment ushered in a new organizational model of knowledge creation, the same technological and demographic forces that are turning the Web into a massive collaborative work space are helping to transform the realm of science into an increasingly open and collaborative endeavor. Yes, the Web was, in fact, invented as a way for scientists to share information. But advances in storage, bandwidth, software, and computing power are pushing collaboration to the next level. Call it Science 2.0."
I have to say I'm not totally convinced about the latter phrase. Magazines like Business Week do like buzz-words like Science 2.0, but so far comparatively little science is affected by this kind of "radicalsharing". Genomics is definitely one of the poster children in this respect, but the vast majority of science continues to be lab or small group based with an orientation towards publishing results as papers, not data.

So what have they made available? There are 3 diabetes projects listed:
  1. Whole Genome Scan for Type 2 Diabetes in a Scandinavian Cohort
  2. Family-based linkage scan in three pedigrees with extreme diabetes phenotypes
  3. A Whole Genome Admixture Scan for Type 2 Diabetes in African Americans
The second of these does not appear to have data available online. The 3rd project has results data in the form of an Excel spreadsheet, with 20 columns and 1294 rows; the data appear relatively simple (a single sheet, with no obvious formulae or Excel-specific issues that I could see), and could probably have been presented just as easily as CSV or another text variant. There's a small amount of header text in row 2 that spans columns, plus some colour coding, that may have justified the use of Excel. Short to medium term access to these data should be simple.

The first project shows two different types of results, with a lot more data: Type 2 Diabetes results and Related Traits results. The Type 2 Diabetes results comprise a figure in JPEG or PDF, plus data in two forms: a HTML table of "top single-marker and multi-marker results", and a tab-delimited text file (suitable for analysis with Haploview 4.0) of "all single-marker and multi-marker results". These data are made available both as the initial release of February 2007, and an updated release from March 2007. There is a link to Instructions for using the results files, effectively short-hand instructions for feeding the data into Haploview and doing some analyses on them. The HTML table is just that; data in individual cells are numbers or strings, without any XML or other encoding. There are links to entries in NCBI, HapMap and Ensembl, however.

The Related Traits results also come in an initial release (also February 2007) and an updated release from September 2007. The results again have a summary, a table this time but still in JPEG or PDF form. The detailed results are more complex; there is a HTML table of traits in 4 groups (Glucose, Obesity, Lipid and Blood Pressure), and for each trait (eg Fasting Glucose) up to 4 columns of data. The first column is a description of the trait as a PDF, the next is a link to a HTML Table of Top Single Marker Results for Association, the next is a link to a text Table of All Single Marker Results for Association, and the last is a link to a text table of Phenotype summary statistics by genotype (both these have the same format as above, although the latter has different columns).

It seems clear that there is a lot of data here; how useful they are to other scientists is not for me to judge. Certainly a scientist looking through these pages could form judgments on the usefuleness and relevance of these data to his or her work. There's not much to help a robot looking for science data from the Internet. I'm not sure what form such information might take, although there are examples in Chemistry. Perhaps the data cells should be automatically encoded according to a relevant ontology, so that the significance of the data travels with them. Possibly microformats or RDFa could have (or come to have) some relevance. However, both the HTML and text formats are very durable, (more so than the Excel format for project 3) and should be easily accessible (or transformed into later forms) at least as long as the Broad Institute wishes to continue to make them available.

Friday, 7 March 2008

Data, repositories and Google

In a post last year, Peter Murray Rust criticised DSpace as a place to keep data:
"The search engines locate content. Try searching for NSC383501 (the entry for a molecule from the NCI) and you’ll find: DSpace at Cambridge: NSC383501

"But the actual data itself (some of which is textual metadata) is not accessible to search engines so isn’t indexed. So if you know how to look for it through the ID, fine. If you don’t you won’t. [...]

"So (unless I’m wrong and please correct me), deposition in DSpace does NOT allow Google to index the text that it would expose on normal web pages. [...]

"If this is true, then repositing at the moment may archive the data but it hides it from public view except to diligent humans. So people are simply not seeing the benefit of repositing - they don’t discover material though simple searches."
Peter isn't often wrong, but in this case it was clear from comments to his post that Google does normally index DSpace content, not just the metadata. There were a couple of reasons for the effects Peter saw, but the key one related to the nature of the data. Jim Downing wrote, for example:
"Not sure what to tell you about your ChemML files. Possibly Google doesn’t know what to do with them and doesn’t try?

"That’s my understanding - interestingly, if you lie about the MIME type, Google does index CML (here, for example)."
The data Peter refers to is Chemical Markup Language data in a file with extension .cml. My Mac does not know what it is, and I guess no more does Google… unless perhaps you tell Google that it’s text, as Jim Downing seemed to be suggesting in his comment (I’m not sure this constitutes lying, more selective use of the truth). I can open CML files in my text editor, fine, although of course to process them into something chemically interesting, I would need some additional software or plugins… Here's a chunk of that file [sorry, tried to include some XML here but Blogger swallowed it up]...

There's real data here [trust me: INCHI and SMILE at least, plus bond strengths etc] that could be indexed but isn't. The point is, surely, that this would be just as much a problem if the repository was simply a filestore full of CML files, which is how data is often made available. But unlike the filestore, there is usually some useful metadata in the repository which can assist data users (ie people, in this case); in a filestore, this is either absent, encoded in filenames, or in some conventional place such as README.TXT where it's relation to the actual data file is problematic).

So: in the first place, Google et al are unlikely to index data, particularly unusual data types. And in the second place, repositories encourage metadata, which does get indexed. So from this point of view at least, a repository may provide better exposure for your data (and hence more data re-use) than simply making the files web-accessible.

This doesn't mean that current, library-oriented repositories are yet fit for purpose for science data! Far from it...