What We Did Before The Pentium!
Last December, just before the tax year ran out, I
purchased one of the new "productivity machines," a microcomputer. In spite of
the lateness in the year, this was no impulse purchase. My wife and I had surveyed our needs
carefully and spent the previous two years trying to find a set of packages,
both hardware and software, that would satisfy our needs.
http://www.slideshare.net/johnwannamaker/how-can-i-tap-my-husbands-mobile-phone
We primarily wanted something that would do word
processing. My wife is a writer and needed something
that would be easy to use for short stories, articles, and perhaps a novel. I
wanted something that would allow me to write textbooks, for which I needed
automatic sectioning (for new chapters, sections, subsections, etc.), sub- and
superscripts for formulas and footnotes, margin justification, and a number of
other somewhat sophisticated functions. Besides this, we wanted something that
our children could use for BASIC programming, some computer games, and so on.
It took us 18 months to find a suitable software package, and then the hardware
decision was almost automatic.
In spite of this leap into new technology, I am not
a "believer" in computers. They do not replace thinking, they certainly do not
replace managing, and they are not esy to learn or use. I also question their
purchase and use in the great majority of cases. I think the computer should be
a last resort. It was for me. Let me explain.
I am a professor at a rather large university; we
have something like 47,000 students enrolled at our local campuses. When I came
here in 1976, there was a computer center with two large Amdahl computers, some
smaller units, and a host of peripherals. Few of my faculty colleagues took
advantage of the center. The facilities were mostly used by administration for
record-keeping and similar tasks. Access to the computer was simple and easy,
as was access to computer personnel the same.
Since then, demand for computer time and assistance
from computer personnel has multiplied many times over. Consequently, the
center has grown substantially, as have most computer centers in the last
decade. With this demand behind it, our center has done its best to satisfy its
users, provide adequate computer time, and still keep up with the latest
developments. Computer personnel pride themselves on having the latest version
of every package and system. But I don't.
Last term I attempted to run one of my examination
programs--the kind that automatically prepares stencils for final exams. I
hadn't used it since the previous term. When I tried to get on the computer
from my terminal I kept getting the message "max users; session terminated."
Never having received that interesting message before, and being forcibly
ejected from the system anyway, I decided to go to the (possibly considerable)
trouble of finding someone who might know what this message meant. I was lucky.
On my first call to a computer consultnat at the university, I reached one and
he said that the message, s all students knew, meant there were already too
many terminals connected to the computer for me to get service. The answer was
to try again later, especially when the students were't so apt to be on it,
such as the dinner hour, or 1 a.m., or before 9 a.m.
Okay. I tried at dinner time and, sure enough, I was
allowed to sign on--almost. When I entered my password, the computer informed
me my account monies from last term were no valid and I would hve to get a new
account. This was a bit baffling since we operate on a "funny money" basis,
anyway. It was like saying I needed new counterfeit money. This procedure
regularly takes a week, but expedited through one's secretary, can be done in a
day. So now I owe the secretary another favor. Finally, at dinner time the next
day, I was allowed to sign on the computer. As I tried to call in my
examination program, the computer told me my library was empty. As a matter
fact, I had not library at all. At this, I seemed to recall in the dim recesses
of my memory a proposal policy to "roll out" any user's library that was not
accessed for 90 days. As I counted back, and included Christmas vacation, it
was indeed 92 days sicne I had last even been on the computer. Apparently, I
was already using the computer less and less without realizing it. (I use it
even less now.)
So, back to the computer consultants. This time I
was not so lucky. One would return my call, I was told, if one ever became
free. Promptly the next day my call was returned. I explained my dilemma: I was
trying to write an exam off the computer so the secretaries wouldn't have to
type it. But time to run if off as dittos was slipping away, never mind trying
to get the secretaries to type one for me on such short notice. The consultant
would check on my library, I was told, and roll it back in if it had been
rolled out. "But you should use it more often if you want to access it
regularly." Ah yes. Even if I didn't need it I should call it up, jsut to keep
it there. Kind of like using your air conditioner in the winter to keep the
seals from drying out and losing the Freon in the System, I imagine.
Two days later my library was available and there I
was, on the computer, entering data into my exam program. Goddy! I told it to
"run" and then I waited. The results came back within minutes. Isn't the power
of a computer wonderful? But the output wasn't quite what I had been expecting.
No exam questions in nice rows down the page. Instead, I saw a bunch of
computerese and number/letter combinations (e.g., 6A2). I needed a consultant.
When I went in to the univeristy a couple of days
later. I found the right line and waited for a consultant to help intepret my
output. After some time, one saw me. He dignosed my problem as using obsolete
JCL. They had updated their version of Wylbur to version 6.2, as everyone had
been informed in their newsletter, and version 6.0, which I have been using for
some time, didn't work any longer. Of course, they had replaced version 6.0
with 6.1 some time ago and these versions were compatible, but all users were
supposed to switch over to version 6.2 in the meantime.
It was really my own fault for not converting when I
was supposed to, but the thought intrigued me that, since the program worked
the last time I ran it and didn't now, there must have been one day when the
program would run and then the next day when it wouldn't. For that matter,
there was one minute, even one second, when my program would run and the next
second it wouldn't. I found the thought fascinating.
Back at my terminal the next day, I corrected the
JCL and reran the program. Again, I fetched the output and looked for my now
desperately needed test. Instead I saw the words "User performed an abend at
location 7X0." Now, I didn't know exactly what that meant, but I did know an
"abend" was a very bad thing to do and that the computer wouldn't let my
program run. Somewhat ashamed of myself, though not sure for what, I vowed to
reinput the data without doing one of those nasty abends again.
But darned if I didn't get another abend. I decided
to spend some time at this and very carefully review my inputs one final time
to see if it wasn't just some careless error I was making. After all, the
program worked the previous term so the error must be in my input data.
After dinner, I spent some time going over the data,
and sure enough, there was an error in my input. I was ecstatic! I called up
the computer again, input the data stream correctly this time, and told it to
run. It responded, as it always does, with my position in the input queue. I
was in position 71! I had never seen a position beyond 15 and, for that matter,
didn't know the computer could count that high (or at least hold that many jobs
at once). AS I asked the computer to keep "locating" my job over the next 10
minutes. I found it didn't seem to move from position 71. Someone was hogging
the whole computer and no one else could get on! To pass the time I decided to
take a shower. When I came back I was flabbergasted to find I was now in
position 75! How could that happen? Can some users get priority to move ahead
of others? Apparently so. I decided to go out for a beer with the gang. I don't
drink beer, but I had to do something to chew up the time and, simultaneously,
calm my anxiety. The secretaries were going to be awfully mad at me for handing
them stenciles so late in finals week. I would be in debt for quite a few
favors now.
After the beer, I sopped by my office to gather the
results and leave them on my secretary's desk to run copies in the morning.
Indeed, my job was finished, but it still hadn't run--more abends! I couldn't
stand it. It was now 1 a.m., 31 hours before my 8 a.m. exam was scheduled to be
handed out to 250 students. I gave up. I sat down and started manually writing
up the final exam. In an hour I had completed a test almost as long as the one
I had contemplated giving off the computer. I left it on the secretary's desk
with an apologetic note of explanation and went home to bed. Driving home, I
couldn't help thinking about the 15 or more hours I'd spent over the previous
two weeks trying to use a computer to do pherhaps one and a half hours' worth
of manual effort, not to mention all the frustration and anxiety.
The next day I decided to look very carefully at the
computing center's newsletter to see if I could learn something of value.
Perhaps it contained something that would help me avoid another nightmre in the
future. Here is a sampling and some of my thoughts as I read.
FORTRAN H will be removed from the system on Jan. 1.
Users must change their JCL to reference S475. VFORTLIB.
What in the world is S475? VFORTLIB?
IBM did carck down on some sloppy coding practices
that were previously acceptable.
So now users won't be able to run at all unless
their JCL is absolutely perfect.
Check the schedule for a short course on the
differences between these two compilers.
I need to take a coruse because they updated a
package?
As of Jan. 1, any datasets with invalid project
numbers will be archived to tape.
There goes my library again.
Data sets without a valid high-level qualifier will
be scratched, regardless of when they were last accessed. So if you create a
data set with a name of BOY or GIRL, don't expect to have it stay around.
I wonder what a valid high-level qualifier is; maybe
SMART BOY?
Beginning Jan. 2, a new disk space management
facility will be introduced.
Will this affect me? Am I on a disk?
Users should note that unless the DSBS option is
added to the PARM field, the BLKSIZE option in the DCB field will be ignored.
Huh?
We have received new versions of the RTI procedures
for Statistical Analysis System (SAS).
Won't be able to use SAS anymore.
The installation of the new version of the Job Entry
System (JES) has been postponed again.
Reprieve!
Job cards that do not have the project number coded
on the same card as the job name will not be acceptable when the new JES is
installed.
Expect trouble soon.
A new release of SHAZAM has been installed. The old
version will be removed Jan. 3.
Captain Marvel, where are you?
From this I concluded that all those people at the
computer center aimed to keep the systems as unused as possible by changing
them so often that no one could ever learn them. This reminded me of the early
days of libraries and other service organizations whose staff hated to see
patrons come to use their nice neat facilities because they messed thme up so
and were such a bother.
Anotehr user disservice of many computer centers in
the name of progress is the elimination of all those noisy, troublesome
keypunch machines. Our center's goal is to have all clients transfer their
programs onto tape. When I inquired how to do this, I was sent to the tape
consultants, where I spoke to four people who disagreed about teh two dozen
statements and formats that were required. They did, however, agree that I had
to supply my own tape, which, I was told, could be bought "anywhere," such as
our bookstore. As it turns out, a reel of tape costs $25. My programs would
require about five feet or so, I imagined, but the bookstore seemed indignant
when I asked them to cut off about seven feet of tape for me. Local computer
supply outlets told me I could indeed get tape in smaller rolls, but only if I
bought $100 worth at a time.
One of my graduate students was doing a research
project with the dp center of a local firm and wanted to have the data put on
tape to transfer them to our facility and anlyze them. I suggested he use cards
instead, but he insisted that tape was "the way to go." A month later when I
inquired how his project was coming, he stated that he couldn't seem to
communicate correctly the tape format requirements to the firm's dp people, and
that he was trying to get our dp people to talk directly with their people
about the necessary format. At the end of the term I found out he had finally
gotten a tape from the firm but was never successful in getting it read. He
therefore had gathered his family togetehr and they were transcribing the data
by hand.
I wondered if I were the only fculty member to have
such difficulty with the center. Other people didn't seem to complain. Since
that time, I have found that everyone, even the computer people themselves,
have the same problems I had. (I'm not sure if I'm relieved or troubled!) My
experience taught me one thing, though. Computers, at least the mainframe
dinosaurs, are not "user friendly." I have concluded that any computer system
that requires experts to intervene will never be suitble for me. The answer is
either to have someone always available who can talk to both you and the
computer (which I cannot afford), or else to find a system you can have
complete control (and can afford). Now, with the advent of the micro, I have
found an acceptable last resort.
The micro was not easy to learn to use. Even before
starting to read the word processing manual I made myself read the computer
guide to operations book. And then the operating system book. And then the
BASIC manual. Then two books on my micro "made easy" (one was and one wasn't).
Then even the printer manual, a large tome in itself. (And they all have
different functions associated with the same keys.)
The computer system was expensive--almost $8,000.
But at least no one will be changing the JCL on me. What worked yesterday, or a
minute go, will work now. And when the system "max users" me, I can toss mykids
off it and get on myself. I still occasionally get some "abends" types of
messages when I try to do something different, but this is always the case when
learning or trying something new. Besides, if I get an abend when I'm doing
something old, I know it's my fault, so I look for and find my error.
I am convinced that the individually controlled
micros will sweep through the organizations of America and the world; and the
big dinosaurs that frustrate us at every turn, whether mainframe, mini,
distributed, or timeshared, will disappear. The micros will become more
standardized, have more internal and external storage, be cheaper and faster,
and, most important, become easier for users to use. In the near futrue, the
only people using large computers will be those simulating nuclear explosions
or the dynamics of the universe. Businesses will go totally micro--we can't
afford the luxury of having to wait to get on the system, of getting abends
when we need results, of going through keepers of the brain to translate our
problem into computerese, and of charging us funny money while wasting our real
money.
purchased one of the new "productivity machines," a microcomputer. In spite of
the lateness in the year, this was no impulse purchase. My wife and I had surveyed our needs
carefully and spent the previous two years trying to find a set of packages,
both hardware and software, that would satisfy our needs.
http://www.slideshare.net/johnwannamaker/how-can-i-tap-my-husbands-mobile-phone
We primarily wanted something that would do word
processing. My wife is a writer and needed something
that would be easy to use for short stories, articles, and perhaps a novel. I
wanted something that would allow me to write textbooks, for which I needed
automatic sectioning (for new chapters, sections, subsections, etc.), sub- and
superscripts for formulas and footnotes, margin justification, and a number of
other somewhat sophisticated functions. Besides this, we wanted something that
our children could use for BASIC programming, some computer games, and so on.
It took us 18 months to find a suitable software package, and then the hardware
decision was almost automatic.
In spite of this leap into new technology, I am not
a "believer" in computers. They do not replace thinking, they certainly do not
replace managing, and they are not esy to learn or use. I also question their
purchase and use in the great majority of cases. I think the computer should be
a last resort. It was for me. Let me explain.
I am a professor at a rather large university; we
have something like 47,000 students enrolled at our local campuses. When I came
here in 1976, there was a computer center with two large Amdahl computers, some
smaller units, and a host of peripherals. Few of my faculty colleagues took
advantage of the center. The facilities were mostly used by administration for
record-keeping and similar tasks. Access to the computer was simple and easy,
as was access to computer personnel the same.
Since then, demand for computer time and assistance
from computer personnel has multiplied many times over. Consequently, the
center has grown substantially, as have most computer centers in the last
decade. With this demand behind it, our center has done its best to satisfy its
users, provide adequate computer time, and still keep up with the latest
developments. Computer personnel pride themselves on having the latest version
of every package and system. But I don't.
Last term I attempted to run one of my examination
programs--the kind that automatically prepares stencils for final exams. I
hadn't used it since the previous term. When I tried to get on the computer
from my terminal I kept getting the message "max users; session terminated."
Never having received that interesting message before, and being forcibly
ejected from the system anyway, I decided to go to the (possibly considerable)
trouble of finding someone who might know what this message meant. I was lucky.
On my first call to a computer consultnat at the university, I reached one and
he said that the message, s all students knew, meant there were already too
many terminals connected to the computer for me to get service. The answer was
to try again later, especially when the students were't so apt to be on it,
such as the dinner hour, or 1 a.m., or before 9 a.m.
Okay. I tried at dinner time and, sure enough, I was
allowed to sign on--almost. When I entered my password, the computer informed
me my account monies from last term were no valid and I would hve to get a new
account. This was a bit baffling since we operate on a "funny money" basis,
anyway. It was like saying I needed new counterfeit money. This procedure
regularly takes a week, but expedited through one's secretary, can be done in a
day. So now I owe the secretary another favor. Finally, at dinner time the next
day, I was allowed to sign on the computer. As I tried to call in my
examination program, the computer told me my library was empty. As a matter
fact, I had not library at all. At this, I seemed to recall in the dim recesses
of my memory a proposal policy to "roll out" any user's library that was not
accessed for 90 days. As I counted back, and included Christmas vacation, it
was indeed 92 days sicne I had last even been on the computer. Apparently, I
was already using the computer less and less without realizing it. (I use it
even less now.)
So, back to the computer consultants. This time I
was not so lucky. One would return my call, I was told, if one ever became
free. Promptly the next day my call was returned. I explained my dilemma: I was
trying to write an exam off the computer so the secretaries wouldn't have to
type it. But time to run if off as dittos was slipping away, never mind trying
to get the secretaries to type one for me on such short notice. The consultant
would check on my library, I was told, and roll it back in if it had been
rolled out. "But you should use it more often if you want to access it
regularly." Ah yes. Even if I didn't need it I should call it up, jsut to keep
it there. Kind of like using your air conditioner in the winter to keep the
seals from drying out and losing the Freon in the System, I imagine.
Two days later my library was available and there I
was, on the computer, entering data into my exam program. Goddy! I told it to
"run" and then I waited. The results came back within minutes. Isn't the power
of a computer wonderful? But the output wasn't quite what I had been expecting.
No exam questions in nice rows down the page. Instead, I saw a bunch of
computerese and number/letter combinations (e.g., 6A2). I needed a consultant.
When I went in to the univeristy a couple of days
later. I found the right line and waited for a consultant to help intepret my
output. After some time, one saw me. He dignosed my problem as using obsolete
JCL. They had updated their version of Wylbur to version 6.2, as everyone had
been informed in their newsletter, and version 6.0, which I have been using for
some time, didn't work any longer. Of course, they had replaced version 6.0
with 6.1 some time ago and these versions were compatible, but all users were
supposed to switch over to version 6.2 in the meantime.
It was really my own fault for not converting when I
was supposed to, but the thought intrigued me that, since the program worked
the last time I ran it and didn't now, there must have been one day when the
program would run and then the next day when it wouldn't. For that matter,
there was one minute, even one second, when my program would run and the next
second it wouldn't. I found the thought fascinating.
Back at my terminal the next day, I corrected the
JCL and reran the program. Again, I fetched the output and looked for my now
desperately needed test. Instead I saw the words "User performed an abend at
location 7X0." Now, I didn't know exactly what that meant, but I did know an
"abend" was a very bad thing to do and that the computer wouldn't let my
program run. Somewhat ashamed of myself, though not sure for what, I vowed to
reinput the data without doing one of those nasty abends again.
But darned if I didn't get another abend. I decided
to spend some time at this and very carefully review my inputs one final time
to see if it wasn't just some careless error I was making. After all, the
program worked the previous term so the error must be in my input data.
After dinner, I spent some time going over the data,
and sure enough, there was an error in my input. I was ecstatic! I called up
the computer again, input the data stream correctly this time, and told it to
run. It responded, as it always does, with my position in the input queue. I
was in position 71! I had never seen a position beyond 15 and, for that matter,
didn't know the computer could count that high (or at least hold that many jobs
at once). AS I asked the computer to keep "locating" my job over the next 10
minutes. I found it didn't seem to move from position 71. Someone was hogging
the whole computer and no one else could get on! To pass the time I decided to
take a shower. When I came back I was flabbergasted to find I was now in
position 75! How could that happen? Can some users get priority to move ahead
of others? Apparently so. I decided to go out for a beer with the gang. I don't
drink beer, but I had to do something to chew up the time and, simultaneously,
calm my anxiety. The secretaries were going to be awfully mad at me for handing
them stenciles so late in finals week. I would be in debt for quite a few
favors now.
After the beer, I sopped by my office to gather the
results and leave them on my secretary's desk to run copies in the morning.
Indeed, my job was finished, but it still hadn't run--more abends! I couldn't
stand it. It was now 1 a.m., 31 hours before my 8 a.m. exam was scheduled to be
handed out to 250 students. I gave up. I sat down and started manually writing
up the final exam. In an hour I had completed a test almost as long as the one
I had contemplated giving off the computer. I left it on the secretary's desk
with an apologetic note of explanation and went home to bed. Driving home, I
couldn't help thinking about the 15 or more hours I'd spent over the previous
two weeks trying to use a computer to do pherhaps one and a half hours' worth
of manual effort, not to mention all the frustration and anxiety.
The next day I decided to look very carefully at the
computing center's newsletter to see if I could learn something of value.
Perhaps it contained something that would help me avoid another nightmre in the
future. Here is a sampling and some of my thoughts as I read.
FORTRAN H will be removed from the system on Jan. 1.
Users must change their JCL to reference S475. VFORTLIB.
What in the world is S475? VFORTLIB?
IBM did carck down on some sloppy coding practices
that were previously acceptable.
So now users won't be able to run at all unless
their JCL is absolutely perfect.
Check the schedule for a short course on the
differences between these two compilers.
I need to take a coruse because they updated a
package?
As of Jan. 1, any datasets with invalid project
numbers will be archived to tape.
There goes my library again.
Data sets without a valid high-level qualifier will
be scratched, regardless of when they were last accessed. So if you create a
data set with a name of BOY or GIRL, don't expect to have it stay around.
I wonder what a valid high-level qualifier is; maybe
SMART BOY?
Beginning Jan. 2, a new disk space management
facility will be introduced.
Will this affect me? Am I on a disk?
Users should note that unless the DSBS option is
added to the PARM field, the BLKSIZE option in the DCB field will be ignored.
Huh?
We have received new versions of the RTI procedures
for Statistical Analysis System (SAS).
Won't be able to use SAS anymore.
The installation of the new version of the Job Entry
System (JES) has been postponed again.
Reprieve!
Job cards that do not have the project number coded
on the same card as the job name will not be acceptable when the new JES is
installed.
Expect trouble soon.
A new release of SHAZAM has been installed. The old
version will be removed Jan. 3.
Captain Marvel, where are you?
From this I concluded that all those people at the
computer center aimed to keep the systems as unused as possible by changing
them so often that no one could ever learn them. This reminded me of the early
days of libraries and other service organizations whose staff hated to see
patrons come to use their nice neat facilities because they messed thme up so
and were such a bother.
Anotehr user disservice of many computer centers in
the name of progress is the elimination of all those noisy, troublesome
keypunch machines. Our center's goal is to have all clients transfer their
programs onto tape. When I inquired how to do this, I was sent to the tape
consultants, where I spoke to four people who disagreed about teh two dozen
statements and formats that were required. They did, however, agree that I had
to supply my own tape, which, I was told, could be bought "anywhere," such as
our bookstore. As it turns out, a reel of tape costs $25. My programs would
require about five feet or so, I imagined, but the bookstore seemed indignant
when I asked them to cut off about seven feet of tape for me. Local computer
supply outlets told me I could indeed get tape in smaller rolls, but only if I
bought $100 worth at a time.
One of my graduate students was doing a research
project with the dp center of a local firm and wanted to have the data put on
tape to transfer them to our facility and anlyze them. I suggested he use cards
instead, but he insisted that tape was "the way to go." A month later when I
inquired how his project was coming, he stated that he couldn't seem to
communicate correctly the tape format requirements to the firm's dp people, and
that he was trying to get our dp people to talk directly with their people
about the necessary format. At the end of the term I found out he had finally
gotten a tape from the firm but was never successful in getting it read. He
therefore had gathered his family togetehr and they were transcribing the data
by hand.
I wondered if I were the only fculty member to have
such difficulty with the center. Other people didn't seem to complain. Since
that time, I have found that everyone, even the computer people themselves,
have the same problems I had. (I'm not sure if I'm relieved or troubled!) My
experience taught me one thing, though. Computers, at least the mainframe
dinosaurs, are not "user friendly." I have concluded that any computer system
that requires experts to intervene will never be suitble for me. The answer is
either to have someone always available who can talk to both you and the
computer (which I cannot afford), or else to find a system you can have
complete control (and can afford). Now, with the advent of the micro, I have
found an acceptable last resort.
The micro was not easy to learn to use. Even before
starting to read the word processing manual I made myself read the computer
guide to operations book. And then the operating system book. And then the
BASIC manual. Then two books on my micro "made easy" (one was and one wasn't).
Then even the printer manual, a large tome in itself. (And they all have
different functions associated with the same keys.)
The computer system was expensive--almost $8,000.
But at least no one will be changing the JCL on me. What worked yesterday, or a
minute go, will work now. And when the system "max users" me, I can toss mykids
off it and get on myself. I still occasionally get some "abends" types of
messages when I try to do something different, but this is always the case when
learning or trying something new. Besides, if I get an abend when I'm doing
something old, I know it's my fault, so I look for and find my error.
I am convinced that the individually controlled
micros will sweep through the organizations of America and the world; and the
big dinosaurs that frustrate us at every turn, whether mainframe, mini,
distributed, or timeshared, will disappear. The micros will become more
standardized, have more internal and external storage, be cheaper and faster,
and, most important, become easier for users to use. In the near futrue, the
only people using large computers will be those simulating nuclear explosions
or the dynamics of the universe. Businesses will go totally micro--we can't
afford the luxury of having to wait to get on the system, of getting abends
when we need results, of going through keepers of the brain to translate our
problem into computerese, and of charging us funny money while wasting our real
money.
