March 18th, 2009

A business selling a dollar for fifty cents..that sounds like a popular idea to me.

Popular ideas aren't always..sustainable ideas.   And while fundamentally bad but highly populist memes may find easy purchase in the minds of politicians who have the authority to tax and spend their way into an implementation... private for-profit and non-profit leadership alike must be much more discriminating about their activities and how they go about marshalling resources for their work so that its a sustainable effort over time. 

Not only do private sector leaders have to avoid fundamentally bad ideas, they also have to make the sometimes hard decision to forego from implementing popular, good ideas, because the resource burn would be unsustainable.  Now, private for-profit businesses have a lot of ways to get a handle on the potential and ongoing loss associated with resource allocation when developing a new product or service. And  they have tools to keep an eye on how efficient they are across your day-to-day operations.  Even for small companies, measurements like revenue per employee can give you some relative sense of how you are doing relative to competitors.

But how do non-profits or highly volunteer centric efforts get a handle on what is and is not sustainable activity? What does sustainability really mean when you dont have things like a revenue per employee measurement to guide you as you are relying on a significant investment of volunteer labor? How do you price in the value of the labor that is gifted? How do you price in the cost associated with managing that gifted labor in a way that respects the interests of existing individual volunteers as well as the long term project sustainability? And even more importantly, once you are self-aware that you might not be engaging in an activity that is sustainable as a non-profit or a volunteer-centric project, how do you efficiently change course?

-jef

My most important Fedora blog post ever.

I'm not exaggerating with that title.

Next week I am attending an NSF sponsored workshop on Sustainable Cyberinfrastructure:
http://cisoftwaresustainability.iu-pti.org/

There are a number of position papers you can read as well:
http://cisoftwaresustainability.iu-pti.org/positionpapers

I think this is a pretty big deal and I think everyone should think its a big deal.... especially for people who believe in either the function of basic science research as a catalyst for technical and social progress or people who believe strongly in open development methodologies as a catalyst for deeper and more impactful collaborations.  Even more so if you happen to be in the union of those groups and a US citizen and care about how the NSF as a Federal agency goes about funding research and education.

The workshop agenda is packed with some very interesting questions about how to how to build and manage a sustainable ecosystem of software resources for science research.   Code reuse, long term maintenance, development funding, educational outreach...all the hard questions are on the table.

And thanks to the graces of the Fedora Board and Red Hat I'm going to be there making the case for an open development model patterned on my personal experience with the Fedora Project.

For those who really don't know me.  I am a research scientist. I take data, I analyze data, ultimately I produce capital S science as a product. All along the way in that process I rely heavily computer hardware and software as tools. Gone are the days when we scientists are reading analog gauges and taking Polaroids of oscilloscope  traces. Well at least for some of us those days are gone.  For those of us who work with digital telemetry of both laboratory and real world phenomena, computer software is are essential tool that often times we have to build ourselves to meet are own particular needs. 

Yep, we write a lot of code, even though we are not primarily trained as computer scientists.  And we don't necessarily do the best job of it either, especially in terms of long term  maintenance and code re-usability.  And even those we build some of those tools, they aren't necessarily open. Science codebases are riddled with legacy licensing or proprietary  dependencies which can hurt the ability to effectively collaborate. Especially when that collaboration needs to bridge across the  academic and private sector because of the difference in pricing of some of the software tools for academic versus private users.  A sustainable science platform which stresses openness could knock some of those barriers down.

Having the NSF champion and organize an open development ecosystem for science related software could help address a number of issues.  I think its the sort of possibility worth exploring something the larger open source ecosystem that is already out there might want to actively support by helping out where they can.  Stay tuned.

-jef