Re: ANN: DataBasin 0.4

[Scott Christley]

On Nov 5, 2011, at 7:02 AM, Richard Stallman <rms@gnu.org> wrote:

> 
> We define SaaS as doing your own computing in someone else's server.
> That is always bad, because you lose control of your computer that way.
> We must develop local replacements to do the computing people want to do.
> 
> See http://www.gnu.org/philosophy/who-does-that-server-really-serve.html.

This is an interesting essay, but it poses a conundrum for some academic 
research. Biological research is undergoing a significant change whereby large 
quantities of biological observations can be made quickly and cheaply, but 
extracting useful knowledge from this data requires significant computation, 
for example assembling a genome from millions of sequence reads. Is free 
software available to do this computation? Yes, debian-med is an example of a 
group that collects and packages such software. Does every research lab have 
the computing resources to do this computation on their own machine? Typically 
no, and requiring every lab to have their own locally sufficient computation 
means a much greater amount of limited research dollars would be spent on 
computer hardware versus for other scientific activities, thus to be cost 
effective they use SaaS.

It can become even more fuzzy as the instruments used for observation combine 
both measurement and computation. These devices are almost exclusively 
proprietary and expensive, so labs must send genetic material to a company 
which performs the measurement and computation, thus SaaS, for them. Doesn't 
sound very much like freedom does it?

An argument might be made that this scientific data should be free anyways, 
except it can put a person's individual liberty at risk. What if that DNA was 
your own? An insurance company might deny you coverage because of a genetic 
mutation. A company might not hire you because you have a mutation in the 
"works well with others" gene. More covertly, somebody may make a change which 
leads a doctor to make an incorrect diagnosis and prescribe the wrong medicine. 
These decisions might be made without you ever being able to look at the 
underlying source code used in the computation.

So while I agree with some of the basic tenants of your SaaS essay, I feel the 
broad statements of "Don't use SaaS!" and "this is always bad" inhibits 
scientific progress.

Scott