Thing I've larned from working on ah-tty
- Setting up a new session or tty is a very odd, poorly
documented, and partly unportable tasked. This is a weakness in UNIX
- The GNU libc has a great deal of information on the tty and how to set
them up. But that information isn't neccessarily portable (what if there is
no glibc?)
- VT100 emulators are stuck with some non-deterministic parts,
and the protocol is kinda hard to find good documentation on.
- VT100 is kinda like an RPC mechanism
- VT100 there are too many ways to do the same thing in VT100, the ones
that aren't any more efficient shouldn't have been created.
- If you create a set of relative cursor motion schemes, you should never
create a cursor addressing scheme that is absolute or relative depending on
some mode flag. Just make it absolute.
- Not all termios, terminfo, termcap or curses implemenations are the same.
- There are many ok sources of information on the net
- VT100 specs are seldom followed (and heavily extended)
- Termcap and terminfo are usually the more definitive source of terminal
emulation information
- My $TERM variable is not honored very much.
- `screen' is a great program, but not the answer to everything;
plus it has lots of LISP like stuff. But it is truly great.
- Bash's screen managementt has a few bugs of its own (especially near
the bottom of the display)
- I don't like CVS. It requires a lot of man-power for simple things.
- Trouble-tickets, or problem reporting software can allows techies to
close problems too quickly. Sometimes before the problem is solved, and
confirmed by the person needing help.
- A problem is solved only when the solution meets the acceptance criteria
- Tech support people are overworked and get stuck dealing with a lot of
cranky folks. Okay, we all knew that. Hug a techie today.