[open-science] Planning for the cost of free

Peter Murray-Rust pm286 at cam.ac.uk
Sat Dec 17 17:29:04 UTC 2011

On Sat, Dec 17, 2011 at 2:56 PM, Puneet Kishor <punkish at eidesis.org> wrote:

> Like all of us on this list, I too have been thinking about OA for a long
> time, but the recent question about the link between OA, data mining,
> sinking servers, and hence, a possible need for an exception in the legal
> obligations has brought to mind once again the mechanism for enabling OA.
> The desire to be open has to be supported by the capability to be so.
> Every time I put out a link to one of my data applications on any one of
> the various programming lists, I can see the process monitor spike up as my
> server is hit by a barrage of queries. I have unlimited bandwidth, and a
> pretty capable, top of the specs computer, but it does start sweating and
> breaking stride. So, it makes sense that if I were using and creating
> hundreds of GB of data, and serving processes that were very
> compute-intensive, I wouldn't be hiding all the data behind a pokey dial-up
> line and a 5 year old computer with an incapable operating system.

I think it's worth distinguishing OA for documents and document-sized data
and OpenData with hundreds of gigabytes per project. Also distinguish
between "final snapshots" and ongoing dynamic data.

I doubt that publishing manuscripts, diagrams, excel spreadsheets etc. is a
major hit. And I suspect that most institutional repositories would be
delighted to have something to put in their pot. An *institution* that
canot manage shouldn't be in the game. I agree there is a potential problem
for people without institutions but I doubt this is a problem.

It's more difficult when there are zillions of files or they are huge.
Again if these are in anyway institution-related you should try them.

Having said that I don't think data should generally be at institutions and
this is a problem that the domain or the country should solve. Dryad,
Tranche, etc. are doing a good job of looking after data. I think there has
been talk of a modest charge. Data costs money, but given the emphasis by
research funders solutions should be emerging

> But, what if I didn't build into my project the cost of making OA
> possible? As I like to say, free is very expensive. If all of the project
> funds are devoted to science, there will be no money to make OA possible,
> and conversely, if all the funds are devoted to OA, there will be no money
> for research. There is a sweet spot somewhere that balances the funds
> between doing research and making its data and results available freely to
> everyone. Some of us call it the Warnick Curve (another story).
> Many funders are asking for data management plans at time of grant
submission and so this can be costed out as a proportion. I agree it's hard
at present because there is no clear economy for this but it will emerge.
(I do not see conventional publishers as being the solution - generally
they have no expertise in data management).

> So, the question -- can we genuinely plead inability to make OA possible
> because we have inadequate capacity?

You sound as if you *want* to plead incapacity!

> Or, we should have built in the cost of OA into the research proposal in
> the first place so we couldn't hide behind such an excuse?


> In other words, we have no business doing complicated research with lots
> of data using public monies unless we also think of the continuing costs of
> making it all available.
> Yes. It will be hard at first. This is an area where IMO the funders and
the institutions need to be ahead of the researchers and make it easy for
them. Personally I think national libraries are a better place to develop
this than universities

> Of course, we have to define "available," but that is another rant.

If it's institutional or national then it's their responsibility to provide

> --
> Puneet Kishor http://punkish.org
> science http://earth-base.org
> advocacy http://creativecommons.org
> _______________________________________________
> open-science mailing list
> open-science at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/open-science

Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-science/attachments/20111217/2de15b70/attachment-0001.html>

More information about the open-science mailing list