root%bostonu.csnet@csnet-relay.arpa (BostonU SysMgr) (08/21/85)
Ok, I for one have noticed quite a rash of letters (some from myself!) from people who seem to be otherwise quite competent with software and manuals. All of these seem to be a general inability to figure out new features on UNIX systems, so far: Semaphores, sxt's, bitmap (me, 7300, caused panics), tty ioctls etc. The only way *I* figured out sxt's was to read shl.c, the trick with the pipe I wouldn't have figured out so easily. Suggestion: what are the chances of seeing sources to running, reasonably useful programs using (especially) new features being delivered with systems? It's nice when you have a source distribution, though some things aren't very much used or take a while to track down (quick, which piece of software in my SYSV.2 source distribution has a good example of semaphores in it?...buzz..time's up, I dunno either, grep semop */*/*/*). I think we have entered a very peculiar state of the haves and the have nots: Either you get source distribution or you don't get one C program.* I have mentioned before on this list that I really don't think it would send AT&T into title 11 to always include the sources to login.c and a few other key programs with binary distributions to give a site a chance to adapt to their needs (ya know, like different password schemes or 'can s/he login now' or other things that developers may not have thought of.) Here's another need. Developers? * Ok, maybe one or two, like twinkle in the curses manual, and thanks! -Barry Shein, Boston University