laura (06/12/82)
Okay. now, everybody get ready with their delete key. I warn
you now that this is going to be a long message. Part A deals with
"Even my sister" -- and Part B is (since I was asked) the long, (and
I hope final) treatment of the "rm *" question.
-------PART A --------
Nobody is going to get shot for saying "Even my sister...."
For one thing, meaning depends upon context. If you were
describing the fact that your family is exceptionally strong, and
mentioned that "you can lift 180 lbs comfortably, and even my sister
can lift sofas and chesterfields" you wouldnt get a rise out of me.
I know women who can singlely move chesterfields. (One runs a
one-family Moving Service). They are exeptionally strong people, who
are all the more remarkable given that, in general, women are less physically
strong then men. (Though tests have shown that they can, on the average,
sustain pain better than men can). However, in the context of eis.341
there is the assumption that while it is strange that Mr. Anonymous and his
brothers became hackers, it is even more strange that his sister did. Why?
Werent the same family conditions which were claimed to have attracted
his brothers to computing present for his sister?? ("Even my
sister, who was raised by my aunt ...." would be one thing,
but seing as this was not mentioned, one would have to assume that the
very fact that she was a woman makes her all the more unusual.
This topic probably will develop into a full-fledged flame
discussion of its very own, but, here goes anyway:
No one can deny that the percentage of women in science careers,
including computer science, is low. Even given studies which
show that, on the average, women are better at verbal skills while men
tend to do better at tasks of reasoning, there is still a disportionate
sex-ratio among scientists. After all, there are many males with verbal
excellence...what happened to their female counterparts? There are
many intelligent women who do better in verbal skills than in logic
and reasoning but still do better than 80 or 90 percent of the men in
logic. (And then there is me -- ever since I was misdiagnosed as mentally
retarded as a young child, I have had a flock of Phd students poking
into my head to try and use me as a well documented control against
other pre-mature babies. I do substantially better in logic than in
verbal skills every time around.) So given this large pool of potential
female scientists ... where do they all go? It is my firm belief that
the reason you rarely see them is *that society does not yet approve of
females in positions where their chief concern is ideas, not people*.
Nursing and Teaching, traditional female careers, thus meet with the
standard, while physics and math do not.
Thus, the statement made in eiss.341 is doubly dangerous -- first
it is a subtle crack at those women already in the sciences, but worse
it supports the prevailing attitude which is the cause of the disportionality.
I apologize to those offended by the fact that I lost my temper.
Anyone wanting references to the studies I mentioned can get them
from me if they mail. If sufficient interest ensues, I will post to the
net.
postscript I dont think that I have suffered any damage from *my* brother --
we get along quite well ... of course the time I accepted his double dare to
drink a glass of vinegar made me rather ill, but thats about it.
-------Part B -----
Every three months somebody comes up with another version of
rm(1) for the naive user. The following is a list of every strategy
I have ever heard of, its virtues and flaws. The next time someone
starts this up, I am going to mail it to him or her. (Unless they are
on the arpanet and I cant get to them, maybe someone who is on both can
save a copy to mail arpanauts). Anyone who knows of something which
I havent mentioned, or *heavens!* points out a mistake that I have made
(no typos, please!) can add to this. But it seems a waste of time and
(in the case of those of us with long distance bills to pay) money to
do this one so often.
1. Do nothing. After the user loses a file or 3 they will learn.
Advantages: Its easy to implement.
Disadvantages:40 irate users can inflict considerable
damage on one poor system manager. Arguably
it is your responsibility to serve these
people better.
2. Get rid of rm altogether.
A:Nobody can find rm by mistake.
D:Nobody can find it on purpose, either! This idea is
probably no good...better to have a rm, but make a
separate command (del for the purposes of this note)
for naive users to use.
3. Okay -- let del be synonomous with rm -i; a user must confirm
the deletion of a file before it is done.
A:Its easy to implement.
D:It is awfully wordy for the user who doesnt make mistakes,
any if you give more experienced users the "rm" command it
will enter into user community knowledge ... and the problem
may reoccur. It doesnt help the user who is twice wrong.
4. Let del appear to remove files but actually place them in an
invisible directory somewhere where they are not actually
deleted until the user logs out. "getback file" then
could be used if the user discovered his mistake.
A: Most mistakes are noticed quickly. It isnt verbose
D: Normal collision problems. It requires disk space.
If you want to use this scheme for editing as well as
deleting (and perhaps moving) perhaps too much disk space.
What do you do about a deletion in the background? It
doesnt help the user who deletes files, logs off, goes to
lunch, and comes back and finds that he has made a mistake.
It would require a lot of local twiddling to reflect a
local enviornment (number of free blocks, disk drives, etc)
and so would end up as a local product even if you got
Bell or Berkeley to distribute it.
5. Forget about del -- just do backups.
5a) on disk
A: Disks can be accessed fast, and by several users at once.
D: They are expensive. They occupy a lot of space. Given
that some users will want a backup of everything you will
have to install disk quotas or some sort of policing. On
VMS systems where every file edit produces a new file and
does not delete the old, without disk quotas you could find
800 versions of somebody's thesis, and 900 fortran programs
(870 dont compile, 15 have execution mistakes, and the
last 15 have one line additions, comments, and changes in
output format).
5b) on tape
A: using incremental daily backups and weekly fulls you can
provide reasonable service -- if this is too excessive you
can modify it...this is what we do here. Users can be
encouraged to do their own, as well.
D: Somebody has to do them. On a one disk system, the
incrementals arent too bad--but on a four or five
disk system you would spend hours (of course you
probably also have several tape drives which you
could use concurrently, but still). You still cant
help the user who works for four hours between backups
and makes mistakes. Tapes too are bulky -- eventually
some must be disgarded. Broken tape drives, bad tapes...
Doing backups monopolizes a valuable system resource
a (the?) tape drive. If you shut the system down for
backups it means there will be a resulting loss in service.
(Actually this can be quite useful ... if users already
know that the system is going to be down for a while every
night they dont get as upset when you extend things to
check hardware, or test a new kernel modification you have
made Single-user before unleasing it on the whole, compared
to the wrath of users who get an apb "shutdown in 15 minutes
for an indefinite period to install such-and-such" when they
expect continuous service).
5c) on videodisk
MY FAVOURITE
Some people somewhere (Ive forgotten -- CMU??) are working
on a system where files are backed onto tertiary storage --
videodisks. Accessing them is slow, but they hold a lot of
stuff and are easy to store, are relatively cheap and with
them it is quite possible to backup absolutely everything
forever! The disadvantage -- they arent available yet.
Now everyone interested in the question can go consiser the system they
are using and use some or all of these to personally solve the problem
of naive users deleting files accidentally.
Thank you for wading through all this -- I realize that it is verbose.
laura
decvax!utzoo!laura