late night linkage

7 thoughts on “late night linkage”

  1. Linus’ questions are mostly trash-talking – if Sun does go GPLv3, they’ll want to dual-license with CDDL, so they couldn’t steal v2 or later code. Or if they don’t require CDDL for new code, there will have to be replacements for all the encumbered bits, which project emancipation is working on. This combined with Indiana’s goals of being fully open source will show Sun is committed to software freedom, if the GPLing of Java wasn’t enough.

    Somewhat apropos to the above and your gcc link is this story of RMS and the MIT AI lab:

  2. Does anyone recognize that elevator product? Looks like some smaller no-name brand to me, and I can’t match it with the products I have seen from the 4 largest players..

  3. Note that if you read carefully on the elevator thing, it looks like an embedded flash system where the HW failed for some reason, not necessarily a software failure.

  4. Luis, although the drivers in Linux are always brought up as something OpenSolaris wants, I don’t believe that’s a particularly strong reason to relicense OpenSolaris under GPLv3. For one, the Solaris DDI (device driver interface) is different from that of Linux. Also, the design center around the APIs exposed by the DDI is different between the two systems. For example, the OpenSolaris DDI works to ensure interface compatibility between kernel revisions (such that drivers from 10 years ago or more still work on Solaris 10 and beyond.)

    While it may be possible to “port” a driver from Linux to OpenSolaris, the work involved to do it properly would be substantial and then you would effectively have a fork from the original source. Updating that driver in the future would be a constant maintence cost.

  5. David: yeah, I should have been more clear- I’m not sure it really makes sense strategically for Sun to do that, just that it is legally possible, which is not the case the other way around.

    (I do think that it is a problem that Sun badly needs to fix, but that is a different discussion and only tangential to whether or not licensing opens a possible mechanism for that.)

    James: given the Apache folks’ unanswered and (as far as I can see valid) concerns with the JCP, I’ll refrain from giving Sun full credit on Java just yet. But yes, Sun has come a long way, and I’m beginning to swing around towards ‘give Sun the benefit of the doubt’ rather than ‘always assume the worst.’

Comments are closed.