A vote for terminal paging in the kernel

[mzal at pegasus.UUCP]

I'd like to cast my vote in favor of terminal paging in the kernel.
Let us look at some of the objections raised and my counterpoints:

1. A new user will be confused when the terminal stops printing.
Answer: Make paging an option which is initially off.  The TOPS-20
   systems which I have used always did this.

2. The problem with most terminal pagers is they stop every 24 lines.
Answer: Under TOPS-20, the terminal stops 23 lines after the users
   last input.  The person who wrote this objection must have used an
   extremely crude pager.

3. How does the system know what the user's terminal and its attributes are?
Answer: Tell it, just like I do now under both TOPS-20 and UNIX.

4. Terminal paging is better handled:
   a) In the programs that need it.
   Answer: So every program I might want to page output from (ls, mail,
      cat, who) should have paging code?  This seems a bit redundant.
   b) Via a filter (like more).
   Answer: This involves extra typing, something extra for the user to
      remember, and extra processes to slow down machines.
   c) A program to log a user's entire terminal session.
   Answer: On a heavily used system, this would be the death of all disk
      space.
   d) In the user's terminal.
   Answer: This "solution" does nothing for those people whose terminals
      are not so equipped.  This solution also seems to me to smack of
      a kind of primitive computer science: "things aren't right - we
      better rewire the machine".

5. There is no need for terminal output paging (for various reasons).
Answer: Obviously this is a matter of taste.  I never thought about
   paging until I used a system with it.  Now, systems without it seem
   somewhat cruder.  Remember, this can be put in as a user option.

6. Various problems are raised (wraparound, etc).
Answer: So it may not be perfect, it would still be better than nothing.

7. It is, in some sense, "bad", to change the kernel for all kinds of
little things.
Answer: This argument strikes me as a religous argument, more than a
   factual one.  What it suggests is that users should be inconvenienced
   because of the lofty goal of a "minimal" kernel.  To my way of looking
   at things, starting and stopping output at page boundries is as much
   a part of the kernel's job as all the other device controling it does.

-- Mike^Z   (Zaleski at Rutgers.Arpa, allegra!pegasus!mzal)