21 Jun
2004
21 Jun
'04
5:07 p.m.
From: "Tom Hudson" <tomhudson(a)execpc.com>
I remember one program where I, honest to god, had to
optimize the code
because I was something like 7 bytes too big, and had to go through the
code and figure out whether I could move some data to the 6502's "page
zero" RAM or use some other technique to shave off the extra bytes!
Another huge project I did on the Atari 8-bit machine required my own
home-made overlay mamager. God, was that fun.
I recognize the situation.
Back in my Philips (one ell, not two) days, I was working on a 32K (8 bit)
mini system for local authorities. Everything was devided in 2K/8bit pages.
The program was so big, that (1) in order to get a 10-20 byte routine, I had
to split up a procedure in two parts (A and B) , where B could then be
called from a procedure C. The reason to call B from C, was that I otherwise
would not have room to end C with a RETURN instruction. Disgusting, but it
teaches you about effective (in the sens of compact) programming.
Another project in the system, had been "undersold" by the representative.
He tought 32K was more then sufficient for RJE, 4-5 tellers, and 1x 3270
simulation. It wasnt. It ran efficiently on 64K, but the customer didnt want
to buy 32 more K. The seller didnt want to admit loose of face, so what we
did, was to implement a disk overlay routine on 8" SSSD floppies. God, it
stank ! The traffic on the floppy was so intense, that we had to replace the
floppy at least once a week. Totally worn out. However, the seller was
happy, as he got his provision from a 32K sale
Nico
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.701 / Virus Database: 458 - Release Date: 07-06-2004