Veteran Code Writers: Is it true that early code writers were paid per line of code??

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

If true, this would redefine sphagetti code.

-- Dave (aaa@aaa.com), September 11, 1999

Answers

I have never heard of that happening, and I started programming in 1961.

-- Peter Errington (petere@ricochet.net), September 11, 1999.

I started programming in 1964 and we were not paid by line of code.

Ray

-- Ray (ray@totacc.com), September 11, 1999.


I was a programmer in the old days, and no, that's silly. I am a writer now, however, and some places pay per word. "Bang, bang bang..." (That's three times the bang for your three cents=nine cents.)

-- Mara Wayne (MaraWayne@aol.com), September 11, 1999.

Lines-of-code was certainly used as a productivity measuring tool back in the 1970s; certainly a programmer that produced 30% more code was considered "better", and undoubtedly this was used to influence performance reviews and therefore salary. Note that this it inself lends insight as to Why Things Are with regard to the totally unregulated state of software that has been written, as to put such as "rule of thumb" in place is to encourage as many lines of code as possible (quantity) rather than formulating the best and tightest code needed to do the job (quality).

And by the way, at least where I was working, "lines of code" meant just that -- comments did not count! Talk about a warped sense of priorities....

-- Jack (jsprat@eld.net), September 11, 1999.

Dave,

I began programming in the 1970's. Though I am certain different shops paid people differently, I was never paid per line of code. Payment was always either per job (ie, you would get x rand, lira, dollars, etc. for developing an application that met y specifications); or by the hour. Most of what I did then was paid per job.

Regards,
Andy Ray



-- Andy Ray (andyman633@hotmail.com), September 11, 1999.


I did see some people get raises and promotions based on the space- time product of their code, however. There were two goals for code in the early days -- small and fast. Comprehensibility was not a virtue, nor was modifiability, extensibility, structure, documentation, time to market, portability, re-usability or any of the various synonyms for "good" that have evolved since.

Based on time and space only, 2-digit years were a win-win. They fit in less space, took less time to load and store to and from tape, could be handled by faster instructions and fewer memory accesses. And the ONLY downside was that someday in the far future it would need to be changed. People who could find win-win tradeoffs made the big bucks.

-- Flint (flintc@mindspring.com), September 11, 1999.


If so, I would be a rich man! <:)=

-- Sysman (y2kboard@yahoo.com), September 11, 1999.

I pretty much began my career in a CICS group during the period where CICS macro level and assembler programming ruled. At that time, CICS obtained storage in 2k segments. This meant that if one wrote a program using 2k + 2 bytes, 4k would be obtained. Considered unacceptable by the company at which I worked, I spent most of my time redesigning the code of others to cut their 2k + programs into something that would fit in 2k. It was a fun period in my life, as folks were astonished that I could modify the code to "fit." My answer to their queries at the time were simply, "I just deleted every other line!" [grin]

I trust that you didn't get this theory from a purported expert, but chose to suggest it yourself as a reason for spaghetti code?

-- Anita (spoonera@msn.com), September 11, 1999.


Anita,

My previous job, I was systems programmer where they wrote their own CICS, all assembly, from the ground up! Added features as they were needed. Ultra hi-performance. We had almost 500 remote terminals in the field, all on a 360/40! A guy from IBM visited the site. He said something like "Nice front end, where's the computer?".

We had a max of 16 tasks running at once. Each task was limited to 8K total! Bigger than that, you had to call an overlay. But it's amazing how much you can do with 8K in assembly. We didn't need many overlays.

Also, the "reusable area" where you would save stuff between transmissions during a session, was 128 bytes! That's all you could "remember" while waiting for the next input message.

Yes, those were fun times!

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), September 11, 1999.


I do know that the government estimated contract costs by lines of code. Projects costs were computed at $5000 per line of code to cover all programming, testing and documentation for software-only jobs.

WW

-- Wildweasel (vtmldm@epix.net), September 11, 1999.



There were two goals for code in the early days -- small and fast.

Well in the OLDER days IBM demanded more lines of code. They had some odd idea that the more lines of code the better. At The Sperry Univac computer archatecture and maintenance school in Orange Co. Calif. back in 1979, I was tracing an a short program through the computer and got extremely irritated with the constant inversion of some of the bits. When I asked about it I was told that the computer makers considered more better and it was explained that the more lines of code someone produced the "better" they were thought to be at programming. You'd have thought they would use a little boolean algebra to reduce the redundency of the hardware. I'm not possitive but I believe back then that people did get paid by lines of code produced. No wonder computing got stuck in a rut until Apple and microsoft came along......

-- Cherri (sams@brigadoon.com), September 11, 1999.


Cherri:

My wife and I went to WalMart a while back to buy gallon jugs of drinking water (50 cents each). So we got a few big carts, and loaded up about 35 gallons per cart. When we got to checkout, the clerk started running every single jug past the scanner (water is heavy).

When we asked whether she could just scan one and enter the number of jugs, she said yes, but her performance was evaluated by ringups/hour, and the multiplication trick didn't count as a ringup! So nobody at WalMart (at least the local store) ever used that approach, since it reduced their ratings. Their paycheck, essentially, depended on lines of receipt tape!

-- Flint (flintc@mindspring.com), September 12, 1999.


Well, I've never been *paid* by LOC's, but I did work at a site where LOC's were used a productivity indicator. As each modified program went ito production a compare utility was used to count modified lines. As it was a PL/I shop and PL/I code is "free form" we just used ISPF to shift large blocks of code left or right by 1 column. These came up as "changed" on the reports!

Our section had very high productivity.

RonD

-- Ron Davis (rdavis@ozemail.com.au), September 13, 1999.


Moderation questions? read the FAQ