Why is EDIT so hard ?

IsaacKuo · 05-22-2007, 02:27 PM

Quote:

Originally Posted by coward77

Yes, I have always used vi, because it will always be everywhere. It is harder to learn, but once you know it you can always edit on any machine. The real question is why small DOS programs have seldom been ported to unix (or linux).

Maybe because there weren't very many small DOS programs and they sucked? Maybe because they tend to use DOS graphics characters and DOS style PC clone text coloring, whereas most *nix systems had different display hardware?

Quote:

I think it has to do with the keyboard buffer and screen buffer. In BASIC for example you can easily write $INKEY to capture a key, this is not that easy in unix. Also DOS machines wre born as single user machines, hence the nicer single user programs, unix was born as multiuser, terminal, high end, not with single user in mind.

Umm...it's best if you just erase the above thoughts from your head. They are so wrong on so many levels, that it's not worth going point-by-point to gather up the few bits that aren't entirely wrong.

It shouldn't be hard to Google up a history of Unix to read for yourself how it came to be.

osor · 05-22-2007, 02:33 PM

Quote:

Originally Posted by coward77

I think it has to do with the keyboard buffer and screen buffer. In BASIC for example you can easily write $INKEY to capture a key, this is not that easy in unix.

First, I don’t think DOS EDIT was written in BASIC. I think originally, it must have been written in assembly, but the “current” versions are probably written in C(++). In either platform, using assembler to access the terminal in raw mode is similar. For low-level (i.e., non-buffered) keyboard use in C, on Unix platforms you use curses (for keystroke capturing use getch() preceeded by cbreak() and noecho() if so desired); on Windows/DOS, you would use conio (which incidentally has a similar getch() function).

osor · 05-22-2007, 02:34 PM

Quote:

Originally Posted by Alien_Hominid

Maybe all these editors long time ago looked liked EDIT but their devs didn't have what to do and so they added and added features until there was no more EDIT.

Actually, a timeline of editors goes something like this:

Code:

                       TECO (1963)
 qed (1966)             |
  |                     |
  ed (1968/1969)        |
  |                     |
  em (1975)           EMACS (1975)
  |                     |
  en (1976)             |
  |                     |
  ex (1976)             |
  |                     |
  vi (1976/1977)        |
  |               Gosling Emacs (1981)
 sam (1983)             |
  |                 GNU Emacs (1984)
  |                     |
acme (199x)           XEmacs (1991)

Of course the years aren’t very accurate (pieces of software do not just begin to exist at a certain time, they are created over a timespan — many of the above are still being developed today), but I tried to put the year the first usable version of each noted editor came out. I didn’t put influences on other software (e.g., ed influenced grep and sed, but since they aren’t interactive editors, they don’t make the list). I also don’t have too much history from the MIT-lineage.

The point is that most of the quirks of vi are inherited from line-based editors.

Sources:
An incomplete history of the QED Text Editor
George Coulouris: Bits of History
Bill Joy’s greatest gift to man – the vi editor
Wikipedia