dean@violet.berkeley.edu.UUCP (01/24/87)
I keep reading (with interest and excitement) about MINIX and its proposed extensions. We may wish to address (now, before things start happening) the problem of keeping track of all the good modifications people make. It's one thing to subscribe to *.sources and try out the interesting looking programs - it's a bit tougher to try out different kernal (kernel?) modifications, each of which may depend on the presence or absence of others. I have a sneaking suspicion that Andy Tanenbaum is not going to want to spend the rest of his waking hours administrating the maintenance and extension of MINIX. Just as a pre-proposal to think about, I could forsee the following developments. i) The official Prentice-Hall distribution. This would form the core of the system and would remain largely unchanged except for bug-fixes. Very few, if any, extensions would be made to this version (subject to the author's and publisher's discretion, of course). ii) A semi-official collection of tested, documented, and coordinated extensions, maintained somewhere on the net by a group of Good People who foolishly volunteer their time for this project. These extensions would come from: iii) Unofficial postings of extensions in this group, which people would be free to try out, improve, rant and rave about, etc. Once an extension reached some stability and consensus of usefulness, the Good People could pick it up, pester the poster for adequate documentation, and add it to the canonical list of extensions. This system would give those of us who want a somewhat stable system a chance to benefit from improvements, while those of us who want to live on the dangerous (but fun) cutting edge could try out all the new gizmos. Of course, it all hinges on finding a group of Good People who are bright enough to do the job, but dumb enough to volunteer. As my 7th grade social studies teacher used to say: "Questions-problems- comments-answers??" -Dean Pentcheff (dean@violet.berkeley.edu)
ee161aba@sdcc18.UUCP (01/27/87)
How about setting up mod.minix.patches?
I'll volunteer myself, contingent on how well I like Minix when I get
the book, for moderator. Anybody else want to volunteer to test patches
out?
Cheers,
David L. Smith
Send here --> {ucbvax, ihnp4}!sdcsvax!man!sdeggo!dave
{ucbvax, ihnp4}!sdcsvax!sdcc18!ee161aba
"Consistency is the hobgoblin of little minds"ebh@cord.UUCP (01/29/87)
Let's not forget Beta Test sites. If/when we start hacking Minix (especially the big hacks like shared libraries), the new versions should go through pretty extensive test phases before release to the general public. Also, this is the time when portability problems can get weeded out. Different beta sites can run different PC/XT/AT clones. Also, has anyone thought about obtaining/developing a validation test suite? -Ed Horch