On Thu, 13 Feb 2003, Adrian Midgley wrote: ... Adrian,
I think we need to make clear distinctions between free design (OpenEHR/OIO), free tools (OIO), and free metadata (e.g. some archetypes/OIO forms). You said: > The Open Source approach to commonalising is different from what has > gone before, and is that one _offers_ elements such as archetypes, > without attempting to maintain complete control over who may use them > or for what, but requiring that changes and im provements which are > redistributed are made available for the pool. My response was : > > This statement is utterly un-true. None of the free software > > licenses require this. In fact this is one of the FUD campaign > > that Microsoft has been propagating. Be warned! My rationale: Yes, one can use OpenEHR or OIO to produce "open source" archetypes/forms, but that is not mandatory per OpenEHR or OIO's free license. As far as I know, neither OpenEHR nor OIO comes with a fixed/static set of archetypes/forms. This means additional archetypes/forms can be produced by others - some of which will not derive from existing archetypes or forms! Since we do not "attempt to maintain complete control", how can we require that all archetypes/OIO forms produced by everyone be made available under a free license? Even if the archetypes/forms are created or modified via a GPL-licensed archetype/forms editor, the author(s) have the option to share their work under any license they wish. A related issue - What are we commonalising? The design (=metamodel), the tools (editors/kernel), and/or the metadata (archetypes/forms)? My understanding is that OpenEHR was a bit ambivalent about whether to also commonalise the metadata layer. On the other hand, the OIO project aims to commonalise just the design/metamodel layer - not the metadata layer nor the tools layer! Do we share the same understanding? Commonalising the design does not require commonalising the tools or the metatdata. For example, OIO is currently implemented using Zope - but it can also be implemented using JAVA, Eiffel, .NET etc. ... > The requirement to release the source code into the commons of > any version of the program that you redistribute is core and > fundamental to the Free/Libre and OSI software philosophy. Sure, but only for derived work. ... > 1. OK under GPL, but it may well be that under a licence ideal > for healthcare and specific to coding and classification systems > disclosure of any changes to a classification library (eg the > Read codes) should be required regardless of redistribution. I don't think that is going to work. A carrot always works better than a stick. Can't you come up with a carrot? :-) I think there are sufficient incentives for publishing the *first* or *most useful* free archetype/OIO form for <you put in the purpose here> that this may not be a problem. In any case, if I am wrong, then someone may get some financial compensation out of publishing non-Free archetypes/OIO forms. That will work too. What doesn't work is for a cartel or monopoly to control the production and distribution of archetypes or OIO forms. This is the lesson that I have learned from what you kindly shared on this list. That is also why I concluded that no controlled vocabulary can be fully sufficient. :-) > To derive one, we might subclass the GPL or FDL to produce a > more restrictive, copyleft required licence. I am unsure of the > merits of doing so. This approach is unwise and unnecessary. GPL does not take away freedom, it gives freedom. GPL is a more restrictive license only for those who seek to attack and take away other people's freedom. Do you disagree? > 2. The permission to do what you like with it but not > redistribute the result is also a fundamental freedom, but in a > healthcare ecology it is perverse and profitless I agree, it is not necessary to concern ourselves with this. If the improvement is useful, someone else will eventually produce the same improvement. If your "improvement" remains secret(=unpublished) or non-Free, it will soon be irrelevant when the freely published version arrives. Best regards, Andrew --- Andrew P. Ho, M.D. OIO: Open Infrastructure for Outcomes www.TxOutcome.Org (Hosting OIO Library #1 and OSHCA Mirror #1)
