[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
(In reply to your message)
- To: DLA at MIT-AI
- Subject: (In reply to your message)
- From: Richard M. Stallman <RMS at MIT-AI>
- Date: Thu ,7 Jan 82 02:41:00 EDT
- Cc: BUG-LISPM at MIT-AI, BUG-ZWEI at MIT-AI
The utility of :DIRECTORY-LIST-STREAM is debatable too, and if you had
taken the time to propose something like it, you would have been
informed of the issues involved. It HAS been pondered before.
I was happy with the speedup. Aside from that, the introduction of
:DIRECTORY-LIST-STREAM is unrelated to the change to how random
properties are printed, and it doesn't incompatibilize anything.
Because :DIRECTORY-LIST-STREAM uses a
:DIRECTORY-STREAM, you are forcing me to write either :DIRECTORY-STREAM
or :DIRECTORY-LIST-STREAM for my file system.
My changes don't REQUIRE your file system to support either of these
two. LMFILE supports :DIRECTORY-LIST-STREAM but actually that was
needed only for testing.
I don't THINK I use :DIRECTORY-STREAM as my entry point,
but perhaps I don't understand you.
The fact that you implemented it such that others must implement it,
makes it an entry point.
***** I did not do so, and you don't have to implement it, *****
and I don't think I deserve it when people get angry at me over things
which did not even happen.
Don't YOU get upset when other people break what you are doing without
consulting you??
Actually, I don't. You probably don't know about the times that this
happened to me, unless it was you who did it, and I came to you and
discussed calmly how to fix it without screwing you back. I didn't
yell at anyone after I spent four hours tracking down the microcode
bugs that garbage was left in VMA by writing the map, and when nobody
else bothered to fix it, I ultimately did that too. The only message
I sent was a calm four line description of the technical details of
the bug I had found.
I accept that I will have to fix bugs introduced by others that
first manifest themselves in interaction with correct code that I
write, because I can't see how this could possibly be prevented.
I expect people to try to think about not breaking what I'm working
on, which I DID DO in this case, but I know that other people can't
anticipate everything that would break my code, and when they do I just
want the other person to be willing to look for a solution that does
what he wants AND what I want.
Also, the fact that I tend to give compatibility less priority and
improvement more priority than perhaps you do, makes me more accepting
of having to change code I maintain to go with an incompatible change
that someone else makes, if it is an improvement. I don't feel that
that is wasted work; it is helping to make the improvement.
Wrt :DIRECTORY-STREAM, I thought I had worked out that existing uses
would not break, because the properties used by QFILE are explicitly
known and their parsing was not affected by the change I made.
RMS, when dealing in a group of people, you win some and you lose some.
You should accept losses and work for the common good of the group, win
or lose. Your input is invaluable in the cases where you win, for
otherwise the group may have not reached such an optimal solution. But
work with the group first in mind, accept your losses and you'll find
out that the world isn't out to get you.
The forms of courtesy that you are urging seem reasonable
but I don't think it is fair to urge me to follow them and
ignore the question of whether others do so with me.
You may think I sound paranoid, but even paranoids have enemies.
I myself thought I was paranoid when I thought people were sneaking
around to avoid me for dinner expeditions, until someone told me it
was true.
I could accept losing some if I felt that people would be open-minded
about considering the idea, and not say no without listening.
But in the times other people did ultimately agree with
me, I often met with an initial condemnation in the absence of
understanding, which I was able to change only through a humiliating
process of pleading for people to pay attention to the details of what
I was saying. This is painful even if it succeeds.
I am much happier when I feel that my suggestions are being listened
to, because then I feel comfortable bringing them up.
Also, the other Lisp machine hackers are often unwilling to discuss
the projects that they are planning with me. That I don't discuss my
plans with others much reflects the fact that people refuse to talk
with me. I often feel that I have to beg to get an audience.
If people WANT me to discuss my plans, they should act as though they
want it.