The concept of free/open software needs some rethinking when software is used as a web service, rather than run on one’s own computer. I’ve posted on this before, as have many others. Important though the issue is, I am dismayed at the thought of new free/open software licenses to address it.
The thought of an open services definition sounds more interesting. So I was pleased to see Luis Villa’s think-post on the subject (via Matt Asay).
My first thoughts are:
- Would an open services definition differ from a free services definition?
- Is it best to start with a definition? Or is it better to start thinking about what licenses or other agreements might be useful?
- Might there end up being a family of licenses (cf Creative Commons)? Each member of the family would represent a combination of points on the various “sliding scales” that Lius discusses to characterize the extent to which a service is free/open.
Q: Would an open services definition differ from a free services definition?
A: Almost certainly; an open definition might (for example) not require source access, as long as data access was available. The FSF would never settle for that.
Q: Is it best to start with a definition? Or is it better to start thinking about what licenses or other agreements might be useful?
A: I expect it’ll be iterative, but starting with implementations is never the right solution. Understand the problem, then implement, then iterate. ‘Plan to throw one away’ doesn’t mean ‘make the first one up as you go along.’
Q: Might there end up being a family of licenses (cf Creative Commons)?
A: Maybe. There will certainly be a variety of licenses; the question is to what extent they are related to each other- tightly, like CC, or loosely, like the various OSI-approved licenses? I’d love to see CC-style, but OSI-style seems much more likely.