Hi Kusti, On 2013.05.18 05:21, Kustaa Nyholm wrote: > Regardless, this paper talks about 'in kernel interfaces' whereas > libusb(x) falls logivally into the 'kernel to userspace interface' > category of things IMO which according to the paper is very stable > over time and will not break. And again IMO that is the way it > should be.
Remember when I said "first step", and that some more where needed to reach "enlightenment"? What you quoted was exactly why. Allow me to get slightly philosophical then, and help you towards a second step, with the following: Whenever anybody indicates that there exists two very different sets of rules, one of them being a lot less restrictive than the other, and press on justifying why they are in a special position and entitled to have only the less restrictive set to apply to them, you should question the reality of these alleged differences, and whether, if you construe them as legitimate, some or all may be abolished. In the case of Greg's note above, this may bring you to further questions such as: what are the elements that are supposedly making kernel development special? And what, for non-kernel software, would be the obstacles in making it satisfy the requirements where the less restrictive set of rules could apply. Following that train of thought, you may go through topics such as historical software development methods (which, as with any traditional way of doing things, you should question the continuing relevance), proprietary software (which sadly, Greg only alludes there, despite being a most crucial part), scarcity of development resources, close-knit communities, more-or-less centralized development trees, FLOSS licensing, security concerns, and so on. If you spend some time doing that, you may be able to draw some conclusions with regards to whether the development models we seem to abide to are the fairest we can achieve, for both developers and users, or if some of them could see some improvements... > In any case I would vote to keep 1.x API/ABI compatible, be it > nonsense or not. OK. If historically POSIX has been using the major for ABI breakage, as Hans pointed out, then 2.0 may be more suitable. Regards, /Pete ------------------------------------------------------------------------------ AlienVault Unified Security Management (USM) platform delivers complete security visibility with the essential security capabilities. Easily and efficiently configure, manage, and operate all of your security controls from a single console and one unified framework. Download a free trial. http://p.sf.net/sfu/alienvault_d2d _______________________________________________ libusbx-devel mailing list libusbx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libusbx-devel
