Maker Pro
Maker Pro

why do computer scientists say 1KB=1024 bytes?!!

D

David R Tribble

Jan 1, 1970
0
Presumably base 10 BCD was used when applications programs were coded in
assembler.

And COBOL, primarily. PL/1 and COBOL are two of the very few
languages that directly support decimal data types (a.k.a. BCD,
packed decimal, and zoned decimal).

BCD instructions or at least fixed point instructions would be great for
accounting and currency formats in spreadsheets!

Which of course explains why IBM used it to begin with, because it
does not suffer from the rounding errors that arise when converting
from binary. Actually, the origin of BCD and its cousin EBCDIC can
be traced to Hollerith punched cards, which were decimal-based.
 
K

Keith Williams

Jan 1, 1970
0
You don't understand the probate system.

Oh, I do. You don't understand what others understand. ;-)
Then there was no property to distribute. The lawyer had to have
generated a report that the estate had an evaluation below
the magic number for both state and federal (I don't know about
cities--I'm sure certain ones feed off the dead. Note that this
was the tax situation 10 years ago. Congress is trying to
bring it back. You need a signoff
from probate courts to transfer titles.

Each state is different. We had no problem transferring title of
her car with just the death certificate (DMV was most helpful
believe it or not). Selling it was no problem at all. The only
reason we hired (wife actually) was for advice.
If there isn't any
estate taxes to be paid, the transfer involves a judge to sign
a piece of paper. If there are estate taxes to be paid, no
transfers are allowed until both the state and federal accountants
sign off. My state ensures this happens by attaching a lien to
all property which prevents all property transfers.

Selling her house was also no problem. Bank accounts weren't a
problem since my wife's name was already on them. The estate was
under the estate tax limit and under the state limit where it had
to be probated (and she was an only heir). There was a period
where people could submit debts that had to be paid (the hospitals
tried some shenanigans on that one) and that was pretty much it.
Not all states are as corrupt as MA, though I was surprised that IL
wasn't one.
If it weren't for the lawyer, JMF's testosterone producing vehicle
would have become a very expensive boat anchor in my garage.

We sold her car at auction a month after her death. No problem, my
wife just signed the title as executrix of the account. She sold
her house the same way.
And then there was the little Democrat technique of squeezing the
last drop of blood from the dead by insisting that all property
be reevaluated four years after the death date which would have
increased the property valuations by 50% because the real estate
prices had almost doubled. If this had been allowed, the probate
period would have taken another 4 years and the estate taxes would
have acquired most of the assets. The lawyer definitely earned
his money.

I guess MA is even more corrupt than even I had imagined! ...and
that's going some! Ouch!
 
M

Mike Young

Jan 1, 1970
0
Shmuel (Seymour J.) Metz said:
It's also a sequence of 2 bits, 4 bits, 7 bits or whatever size you
would like it to be. For an example of correct usage see the
instruction set of the DEC PDP-6 or the IBM 7030.

So, how many bits in a kilobyte in your funny world?
 
R

Rich Grise

Jan 1, 1970
0
And COBOL, primarily. PL/1 and COBOL are two of the very few
languages that directly support decimal data types (a.k.a. BCD,
packed decimal, and zoned decimal).


Which of course explains why IBM used it to begin with, because it
does not suffer from the rounding errors that arise when converting
from binary. Actually, the origin of BCD and its cousin EBCDIC can
be traced to Hollerith punched cards, which were decimal-based.

There are no rounding errors when converting from binary, if everything
is an integer. ;-) With a 32-bit word, you can do integer arithmetic
up to whatever 2^31 - 1 is in decimal, divided by 100, since it'll be
in pennies.

Computing interest on a bank account can be interesting - I heard a UL
about some programmer for the bank's system which, every time there
was a roundoff of 0.00 < R < 0.01, he just transferred that amount to
his own account, and truncated the customer's account amount. Who's
going to notice less than a penny? ;-)

Out of every customer's account(s).

I heard that in just a few months, he withdrew several millions and left
the country. ;-)

Cheers!
Rich
 
G

George Ellison

Jan 1, 1970
0
Rich Grise said:
Computing interest on a bank account can be interesting - I heard a UL
about some programmer for the bank's system which, every time there
was a roundoff of 0.00 < R < 0.01, he just transferred that amount to
his own account, and truncated the customer's account amount. Who's
going to notice less than a penny? ;-)

I heard an urban legend that there was a really funny movie once
called "Office Space" where something like this was done. In the UL,
the programmer responsible fucked up and xferred some ridiculous
amount of money into his account.
 
J

Jeff_Relf

Jan 1, 1970
0
Hi Rich_Grise, David_Tribble, Rex_Ballard and Peter,

Grise wrote: ...he withdrew several millions and left the country. ;-)

You can commit much more interesting crimes
if you're willing to sully your name like that,
a disposable identity is very powerful... hence urban crime.
 
D

Dave L. Renfro

Jan 1, 1970
0
George said:
I heard an urban legend that there was a really funny
movie once called "Office Space" where something like
this was done. In the UL, the programmer responsible
fucked up and xferred some ridiculous amount of money
into his account.

How is this an urban legend, or am I missing some kind of joke
here? This is a fairly well known movie that came out a few
years ago, and it's on cable every couple of weeks it seems.

Dave L. Renfro
 
G

George Ellison

Jan 1, 1970
0
Dave L. Renfro said:
How is this an urban legend, or am I missing some kind of joke
here? This is a fairly well known movie that came out a few
years ago, and it's on cable every couple of weeks it seems.
It's a joke. Relax.
 
T

The Ghost In The Machine

Jan 1, 1970
0
In comp.os.linux.advocacy, Rich Grise
<[email protected]>
wrote
There are no rounding errors when converting from binary, if everything
is an integer. ;-) With a 32-bit word, you can do integer arithmetic
up to whatever 2^31 - 1 is in decimal, divided by 100, since it'll be
in pennies.

You haven't memorized 2147483647 yet? :) In any event, $21.5M,
give or take, would be 2^31 - 1 pennies.
Computing interest on a bank account can be interesting - I heard a UL
about some programmer for the bank's system which, every time there
was a roundoff of 0.00 < R < 0.01, he just transferred that amount to
his own account, and truncated the customer's account amount. Who's
going to notice less than a penny? ;-)

Out of every customer's account(s).

I heard that in just a few months, he withdrew several millions and left
the country. ;-)

I remember at least one film having that as a subplot. (IIRC the
actor was Richard Pryor.)
 
Oh, I do. You don't understand what others understand. ;-)


Each state is different.

I know that. I was also talking about federal which is not
different.
We had no problem transferring title of
her car with just the death certificate (DMV was most helpful
believe it or not). Selling it was no problem at all. The only
reason we hired (wife actually) was for advice.

How did you prove that you didn't owe estate taxes to the IRS?
There had to be some form with an approval form coming back.
Selling her house was also no problem. Bank accounts weren't a
problem since my wife's name was already on them. The estate was
under the estate tax limit and under the state limit where it had
to be probated (and she was an only heir).

You had to have submitted a form will all the evaluations to
prove that the estate was below tax cap. Who did that?
There was a period
where people could submit debts that had to be paid (the hospitals
tried some shenanigans on that one)

Yup. In my case it was the nursing home and somebody else who
I never figured out. I also received a bill for the Newsweek
magazine which JMF wouldn't have been caught dead with.
and that was pretty much it.
Not all states are as corrupt as MA, though I was surprised that IL
wasn't one.


We sold her car at auction a month after her death. No problem, my
wife just signed the title as executrix of the account. She sold
her house the same way.


I guess MA is even more corrupt than even I had imagined! ...and
that's going some! Ouch!

You want to hear the one they tried this month? The fucking
idiots tried to increase the taxes for 2002 and expected to
collect them. Yes, this is not a typo, the tax year two thousand
and two.

/BAH
 
K

Keith Williams

Jan 1, 1970
0
I know that. I was also talking about federal which is not
different.

Title transfers aren't a federal issue. Taxes are, but that's a
matter of filing out the tax forms. ...pretty easy for a simple
estate. IIRC she didn't even owe income tax, but there could have
been a small tax bill due (don't remember). As far as the estate
tax, that was an easy one. There wasn't anything close to the
federal trip point and IL law makes it pretty painless for estates
below a certain threshold.
How did you prove that you didn't owe estate taxes to the IRS?
There had to be some form with an approval form coming back.

IIRC we simply filed the tax forms showing that nothing was owed.
I didn't do it (my wife even does our taxes), so I don't know the
details.
You had to have submitted a form will all the evaluations to
prove that the estate was below tax cap. Who did that?

My wife, with the advice of a lawyer.
Yup. In my case it was the nursing home and somebody else who
I never figured out. I also received a bill for the Newsweek
magazine which JMF wouldn't have been caught dead with.

Yep, the nursing home tried to play a bunch of games. We told them
to FO and they eventually went away. Double-billing was a game
everyone seemed to play.

You want to hear the one they tried this month? The fucking
idiots tried to increase the taxes for 2002 and expected to
collect them. Yes, this is not a typo, the tax year two thousand
and two.

Sheesh. I thought Clinton's tax increase (retroactive to the
beginning of the year) was bad enough.
 
Sheesh. I thought Clinton's tax increase (retroactive to the
beginning of the year) was bad enough.

But nobody bitched about that, did they? You should thank us,
taxpayers of Massachusetts, for stopping this fucking nonsense.
We seem to be the first guinea pigs for politicians' new methods of
extracting money from pockets. You may also thank us for
stopping diminishing tax rates on capital gains based on
length of ownership. This one had me calling up our DoR
weeping every year. Those blessed ladies who manned the phones were
wonderful; "ignore the instructions and do what comes
naturally" was the bottom line answer.

Caveat: I'm pretty sure it wasn't our pain that changed
their minds; I'm pretty sure that a politician discovered
that he couldn't use the >6-year losses against <6-year gains.

/BAH
 
K

Keith Williams

Jan 1, 1970
0
But nobody bitched about that, did they?

"They" did, but the politicians didn't listen.
You should thank us,
taxpayers of Massachusetts, for stopping this fucking nonsense.
We seem to be the first guinea pigs for politicians' new methods of
extracting money from pockets. You may also thank us for
stopping diminishing tax rates on capital gains based on
length of ownership. This one had me calling up our DoR
weeping every year. Those blessed ladies who manned the phones were
wonderful; "ignore the instructions and do what comes
naturally" was the bottom line answer.

"stopping diminishing tax rates on capital gains based on length of
ownership"? I always thought long/short term
(investment/speculation) differences were reasonable.
Caveat: I'm pretty sure it wasn't our pain that changed
their minds; I'm pretty sure that a politician discovered
that he couldn't use the >6-year losses against <6-year gains.

Quite likely, but they're usually better about writing themselves
loopholes than that.
 
"They" did, but the politicians didn't listen.

No, "they" did not. There wasn't a whimper in the
collective. I always thought that this was due to
the fact that payments were spread out over the next
three years so nobody noticed when writing the checks.
"stopping diminishing tax rates on capital gains based on length of
ownership"? I always thought long/short term
(investment/speculation) differences were reasonable.

There were six classes of long term capital gains plus short
term capitals gains. If you held the capital for longer than
six years you would not have to pay taxes on the gains. Now
before you react with "that's a wonderful idea", this also
applies to capital sold at a loss so you cannot apply those
losses against long nor short term capital gains nor,
in the case of this state, some interest income.

The instructions of which gains offset which losses were
at 10 pages long...and wrong. This was further complicated
by the fact that before one item on the long form could be
calculated, you had to fill out Schedule B, but..before
you could finish Sched. B, you had to fill out Sched D, but...
before you could finish Sched. D you had to have a figure
from Sched. B which occurs later than the number that got
you in the middle of Sched. D in the first place.

Whenever I do this state income tax, I am always reminded
of physics' multi-body problem. By the time I'm done,
I feel like I've whacked a million moles and never hit
one head.

/BAH
 
K

Keith

Jan 1, 1970
0
No, "they" did not. There wasn't a whimper in the
collective. I always thought that this was due to
the fact that payments were spread out over the next
three years so nobody noticed when writing the checks.

You live in a different world than I. It even made the major (s)nooz
networks. Many were royally PO'd. ...and it didn't take three years to
notice. The next paycheck showed the difference.
There were six classes of long term capital gains plus short term
capitals gains. If you held the capital for longer than six years you
would not have to pay taxes on the gains. Now before you react with
"that's a wonderful idea", this also applies to capital sold at a loss
so you cannot apply those losses against long nor short term capital
gains nor, in the case of this state, some interest income.

I see your point, but losses are something I prefer to not carry.
The instructions of which gains offset which losses were at 10 pages
long...and wrong. This was further complicated by the fact that before
one item on the long form could be calculated, you had to fill out
Schedule B, but..before you could finish Sched. B, you had to fill out
Sched D, but... before you could finish Sched. D you had to have a
figure from Sched. B which occurs later than the number that got you in
the middle of Sched. D in the first place.

As long as the series converges... ;-)
Whenever I do this state income tax, I am always reminded of physics'
multi-body problem. By the time I'm done, I feel like I've whacked a
million moles and never hit one head.

;-). Our state income tax is much simpler, at least for me (multiply the
Fed by .24 and send it in). Amazingly, for those making less, the forms
are much more difficult (property tax munged into the income tax). In
fact, it's almost impossible for a low-middle wage earner to do his own.
 
You live in a different world than I. It even made the major (s)nooz
networks.

So? That isn't the collective.
Many were royally PO'd. ...and it didn't take three years to
notice. The next paycheck showed the difference.

I'm talking about that. I'm talking about the retroactive.
I don't see how back taxes owed could have been taken out of
paychecks spread across three years.

You are confusing grumbling about tax increase and then
acceptance with honest outrage of the many. Since this
increase had been advertised by those same news reports
as a tax on the rich, the middle class didn't peep.

Story time: Around here, when Reagan's tax increase was
about to be implemented, the banks ran seminars about
the increase. I went to one. The enthusiastic gal
who gave the presentation asked me, as I was walking out,
about my views of the subject. I told her that I only
came to find out how much my taxes were going to increase.
She was appalled and tried to tell me that it was a decrease.
I told her to do her last year's taxes with the new laws
and she'll see her taxes were going to increase. I told
her that any time a politician says that a tax increase is
only going to apply to the "rich", she can be sure that
it will be her paycheck affected.

Another thing I don't understand about these PC people
is their insistence that screwing the rich is a winner
because it is the middle class who pays these bills.
I see your point, but losses are something I prefer to not carry.

You don't have a choice a lot of times. With a cash buy-out,
the price and date of sale is set for you. If you've been
reinvested dividends, all those little purchases over the
years have to be split up into all seven categories. It was
actually 7 for the state and two for the Federal because
end points of each category didn't quite match. When these
idiots started it, there was also a before-date period and
an after-date period.

Now I could have cheated successfully and lumped all of these
into fewer categories, assuming that a purchase of .12387
shares wouldn't be traceable w.r.t. long/short-term holding
periods. But I was pissed off so I listed each category
and the data enterers had to do the same. If you want
to change policy, you show the data entry people and/or
secretaries why the change is a good idea.
As long as the series converges... ;-)

A miracle happens. I have yet to figure out how this
miracle works. Note that I enjoyed doing taxes because
it was a simpler exercise of sketching a flow chart than
I did at work. Each "event" block also gave hints of
what the politicians intended in the future so I could
extrapolate and plan. I can still do this with the US
Federal forms with the exception of all those fucking
worksheets. Those are beginning to look like a clone
of Mass. tax forms.

The best way I can describe the flow chart of a Mass.
income tax form is a PUSHJ to a co-routine which calls
itself, then JRSTs out of a DO loop at location x.
Eventually, one of the instructions in the coroutine
(which you havent' exited yet and you're still in the
PUSHJ call) jumps back to the line at location x-1/2.
Yes, that is an x minus one-half. When you've finished
with the form, you still have a dangling call on the
stack that hasn't been POPed.
;-). Our state income tax is much simpler, at least for me (multiply the
Fed by .24 and send it in). Amazingly, for those making less, the forms
are much more difficult (property tax munged into the income tax). In
fact, it's almost impossible for a low-middle wage earner to do his own.

Of course. That is intentional. Think about the work flows (that's
what I call them). It creates more work, and costs more, to
collect payroll taxes that will be refunded if only the taxpayer
submits a form. Most won't or they will have somebody else do
the work for a price. All of these laws are passed under the
banner of "getting the rich to pay more taxes". The goal,
my hypothesis, is to remove monetary control from the "poor"
and have "The Governement" do it for them.
Gradually, the "poor" threshold will creep up until the
threshold is middle class, perhaps a little higher.

Another tax-soaking technique is to pass a tax law that
supposedly is for the high-income people. However, this
threshold is placed such that, after a couple of years,
the people who are paying the tax are the middle to upper-
middle- class. Example is the AMT. The people who
do have a lot of money, have enough to hire full time
accountants to avoid this tax. The only ones who don't
have this kind of money and are not trained to be
bosses, are the people who are in the middle class range.
They don't know how to be a boss nor how to hire nor
supervise accountants, lawyers, and all the infrastructure
needed to support their work.

/BAH
 
J

JosephKK

Jan 1, 1970
0
It has nothing to do with marketing but with bookkeeping. The file
information takes up space. This information tells where the file
is, which directory, how big, all the date-time stamps that
are needed to bookkeep the files, etc. We called this storage
overhead. In our scheme, for each file, there were about 100
words (36bits/word) that were needed to just maintain the file
system. There was another set of files maintained by the OS
that managed the physical bits of each disk.

Note also that a directory is also a file from the point of view
of the operating system. So, if you have 10,000 files on your
disk at installation time, you have to multiply 10,000 by
the number of bytes it takes to retain information about each
file.


EVen for teensy little storage media, there had to be data
for bookkeeping purposes.

For instance, in our scheme, each block had a bit reserved in
a file. Thus, if there were 80,000 blocks on the disk,
there was file that contained 80,000 bits. Whenever a block
went "bad" (cosmic rays were always a convenient explanation),
the operating system would mark the corresponding bit in this
BADBLK.SYS file and the OS used this info to avoid trying
to write good data over a block that couldn't be written.

This is another task that we called bookkeeping.

/BAH
No. Eric has told you the actual correct explanation. The lying
mercantilist advertisers have redefined 100 GB drives to store (sans file
system overhead) to store 100,000,000,000 bytes, instead of 100*2^30 bytes.
As a direct result the drive capacities (before file system overhead) are
about 6% overstated for immature computer professionals expectations. File
system overhead takes an additional 2% to 5%.
 
D

David R Tribble

Jan 1, 1970
0
There are no rounding errors when converting from binary, if everything
You haven't memorized 2147483647 yet? :) In any event, $21.5M,
give or take, would be 2^31 - 1 pennies.

That's a pretty small number if you're programming for a financial
institution.

IEEE double-precision floating point numbers give you 52 bits of
precision, but they are binary, so they suffer from rounding errors
when converted into decimal. Unless you scale it by 2^52 to be
an integer, but it's a lot of trouble to go through to do it correctly.

COBOL and PL/1 give you 64-bit binary or 18-digit decimal (BCD)
fixed-point numbers to work with, which is very convenient for
money calculations.
 
J

John Larkin

Jan 1, 1970
0
COBOL and PL/1 give you 64-bit binary or 18-digit decimal (BCD)
fixed-point numbers to work with, which is very convenient for
money calculations.

So does PowerBasic. Plus 80-bit floats.

John
 
Top