Guest
In article <esurfc$ds3$6@blue.rahul.net>,
kensmith@green.rahul.net (Ken Smith) wrote:
was the file on the tape.
of the file. IOW, you implemented something that wasn't in the spec.
In fact, your implementation is 100% contrary to the spec.
directory of the tape, untouched by editing hands.
that were on the tape when we made the master. If the tape
they received didn't match the ones listed in the file, then
they knew they got a bum tape and could alert us that our
manufacturing process has a problem. It was quite possible
that SDC (software distribution center) had an order of tapes
that were a lot less than 2400'. If the master was longer than
the physical lenght of those tapes, then customers would get
a distribution that didn't have the end of the save set.
Not putting a directory of the tape onto the tape would create
risks.
side effects. Not shipping the file would have caused expensive
side effects.
was OK to have one 36-bit word incorrect on that tape. I've thought
about the CATCH-22 problem often since then because those
are neat puzzles.
of the problems in a manufacturing environment when the products
produced were OSes and the computers they were installed on.
That is very different from what it really is like.
The whole point was to document what was on the tape when we made
it.
/BAH
kensmith@green.rahul.net (Ken Smith) wrote:
No.In article <esu74a$8qk_001@s861.apx1.sbo.ma.dialup.rcn.com>,
jmfbahciv@aol.com> wrote:
In article <esrtcj$qj4$1@blue.rahul.net>,
kensmith@green.rahul.net (Ken Smith) wrote:
[....]
In other words, you wrote the TAPE.DIR after all the other files were
written. This means you had to write a file of that size and then the
data and then open TAPE.DIR for writing.
If this is what you did then my method of putting a correct checksum still
works. If you wrote the TAPE.DIR before the other files and never changed
it, my method still works.
[....]
Anybody who has done any grocery shopping would know the difference
between the two. Just because dog food is on your shopping list
never guarantees that dog food will be in your car when you get home.
The only listing that shows you were successful in putting
dog food in the shopping cart is the cash register receipt.
Does this mean you wrote TAPE.DIR after the other files were all written?
Yes. TAPE.DIR was created after the files were written to the tape.
That is how you get a directory listing of the tape onto the tape.
The first cut of the tape had a zero block TAPE.DIR for a place holder
on the tape. Then a DIRECT DSK:TAPE.DIR=TAPE:/CHECKSUM was done.
Then another save was done; this time TAPE.DIR that was on the
tape contained a checksummed directory of the tape.
In other words, you "editted" the tape.
No. I wrote the whole save set. In magtape terms, the savesetYou wrote one file and then wrote
something different in its place.
was the file on the tape.
I know you've explained ad nauseum. You keep ignoring the pointGiven this, the method of making the checksum of TAPE.DIR is done very
easily. I have already explain how several times.
of the file. IOW, you implemented something that wasn't in the spec.
In fact, your implementation is 100% contrary to the spec.
But you don't fulfill the requirement that the file is aI don't think so. You objected to opening it for writing. This means
that you only could have put it onto the tape before the others. You must
therefor have known what you intended to write. You also would have
checked when you were done that you really had written what you intended.
If what you actually wrote did not match what you intened to write, you
would have taken some action. Most likely you would make a new tape.
A second save exercise had to be done. Howver, the checksum of
the file TAPE.DIR in the file TAPE.DIR would never be correct.
We could live with that. It would not generate any SPRs nor
create problems in the field.
As I explained, the claim that it can't be correct it wrong. The method I
explained makes it correct. There is no problem.
directory of the tape, untouched by editing hands.
Nobody assembled the file called TAPE.DIR. It listed the filesNo, what I suggested solves the problem you had cleanly and without adding
any problems. You just seem not to be able to grasp what it does.
I know what it does. And the solution would cause more problems than
it fixed. If your solution was the only one, I'd punt the idea.
You don't seem to be able to understand the idea so yes, I guess you would
punt the idea. leaving the tape incorrect and your customers at an extra
risk.
that were on the tape when we made the master. If the tape
they received didn't match the ones listed in the file, then
they knew they got a bum tape and could alert us that our
manufacturing process has a problem. It was quite possible
that SDC (software distribution center) had an order of tapes
that were a lot less than 2400'. If the master was longer than
the physical lenght of those tapes, then customers would get
a distribution that didn't have the end of the save set.
Not putting a directory of the tape onto the tape would create
risks.
No, there is no gap.[....]
Nothing I suggested would result in that. The net result of what I have
suggested is one line of text in the TAPE.DIR file if it is an ASCII text
file and an extra byte if it is binary.
Since the medium is a magtape, inserting one tiny bit would make a
mess of the save set and cause it to not be restorable.
No, it would not, You just need to think about what I really said. When
you first made the tape, you left a gap for the file.
You do not understand how tapes and savesets worked.You then write this
file after. Before you write this TAPE.DIR, you think for about 3 seconds
and then write a TAPE.DIR that is completely correct instead of one with
an error in it.
Yes. It was an error that was acceptable. The error had noThis
was the master cut of the tapes that would soon be made to ship
to all customers. It had to be perfect and copyable.
..... and yet you shipped it with a mistake on it.
side effects. Not shipping the file would have caused expensive
side effects.
I thought about this problem for months before I decided that itThis means it was not
perfect at the instant you created it. The checksum was wrong because you
didn't think for about 3 seconds.
was OK to have one 36-bit word incorrect on that tape. I've thought
about the CATCH-22 problem often since then because those
are neat puzzles.
We had to do this on a master tape. You simply have no conceptAnother trick that had to done was to save files to tapes that
had been physically logically "shortened". That was done by
moving the silver strip to make a "short tape".
This has nothing to do with the question at hand. The strips on the ends
of tapes were moved on ones that were being used as scratch. You would
never do this on a production tape.
of the problems in a manufacturing environment when the products
produced were OSes and the computers they were installed on.
No, it doesn't. Your method says what we think it should be like.[..........]
Once again what you are saying contradicts what you suggested earlier.
You had to check that you really did put onto the tape what you intended
to put onto the tape. My method makes no change to that requirement.
Your method does not put a director _OF_ the tape onto the tape.
Yes it does! You can do exactly all the same steps as what you did plus a
very small amount of thinking and get it right.
That is very different from what it really is like.
I didn't know because the magtape had not been made yet.[....]
WRONG! You have always been able to edit magnetic tape. You can't cause
the lengths of things to change but you can overwrite a file with one of
the same size.
But it isn't the same size. You cannot predict the size of any
record on a magtape.
You wrote something onto the tape called TAPE.DIR. I have not suggested
anything that makes its size change in anyway that you did not have.
Either you knew before hand its size or you didn't.
The first TAPE.DIR on the tape was a zero block file.I assume you did
because you had to write something to take up that much space. Nothing in
what I suggested makes that any sort of a problem.
Humans made everything. Humans told the computers what to do.[....]
Hardly new. Reel to reel tape isn't used much anymore. Even very stone
aged reel-to-reel drives like the Kennedy 9700 could do it. You being
unaware of this contradicts statements you made earlier.
I was never aware that the purpose of IRGs was so a human could
edit directly to the tape.
We are not talking about humans doing things we are talking about what
computers could do.
YOur method does not provide a directory of what is on the tape.Yes! THat is the only way to get a directory of the tape.
Then as I have explained, the method works the checksum did not need to be
wrong.
The whole point was to document what was on the tape when we made
it.
/BAH