10 updates today

[Ed Cryer]

The PC has always been byte-addressed, with variable length instructions
and data "words" (fields).

Addressing was via a 20 bit base address in a base register plus a 16
bit offset. The low order four bits of the base address were by
definition zero, so that address was held in a 16 bit register.

In the hardware, the base register value was shifted left four bits and
added to the unshifted offset to get the 20-bit effective address.

There was also relative addressing, where the address value in the
instruction was added algebraically to the current program counter to
get the effective address.

All the above is still true, with changes to the lengths of fields and
the addition of mapped memory addressing.

Assembly language programming was a lot of fun in that system.

While I was earning my bread programming ICL 1900 mainframes in the day
time, I used to amuse myself at night with a ZX Spectrum. I would often
use an assembler for the Zilog chip; mostly for disassembling games,
giving them infinite lives, and then giving the saved program to friends.
I once got an early space invaders program to let the missiles go right
through the aliens without any damage.

Ed

 

[R. C. White]

?Hi, Paul.

Thanks for the link. That article sure brings back memories! (No pun
intended. <g>)

My first computer was the original TRS-80 in December 1977. I skipped the
minimal 4 KB RAM model and went for the biggie with 16 KB. As an
accountant, not a techie at all, it took me years to get a rudimentary
understanding of the architecture. Even Adam Osborne's "Microcomputers:
Volume 0" was way over my head. But I did pick up enough binary/hexadecimal
arithmetic to understand that a kilobyte was 1024 bytes, not an even decimal
1000 - and why. But if some of my comments below are not technically
correct, maybe you'll cut me some slack.

That TRS-80 had a Zilog Z80 CPU and could address 64 KB of memory. Of that,
1 KB was dedicated to the monitor screen (16 lines of 64 characters, 64 * 16
= 1024 bytes). LEVEL I BASIC was on a 1 KB (?) ROM; LEVEL II BASIC arrived
on a ~3 KB ROM a few months later; with keyboard and other support
functions, these consumed the first 4 KB of the address space. We could
write BASIC programs as large as 3,284 bytes in a 4 KB LEVEL II machine, or
15,572 byte in a 16 KB TRS-80 (per the "Proof Copy" of the LEVEL II BASIC
Reference Manual, which I still have). (I wrote a 13 KB income tax
estimating program in January 1978; my CPA firm was still using an updated
version of that after I left the firm in 1980.)

So, when I got my first IBM PC-style computer in the early 1980's, 640 KB
did seem like "more memory that we would ever need". My first MS-DOS
machine was the Tandy 2000, which used the less-popular 80186 chip, rather
than the first PC's 8088 or the later 80286. The 8088 held a 16-bit
processor, but it was throttled down by 8-bit address lines; the 186 was
16-bit all the way, like the 286 and later x86 CPUs. That Tandy T2K used a
different upper-memory architecture and I was able to install a whopping 768
KB of RAM in it! ;<)

In the later '80s and the '90s, I was building my own computers and fighting
the expanded/extended memory battles with everybody else. We could hardly
even imagine how we might possibly use more than a GB of RAM - or HD space.

Ah, the nostalgia of a computer nut...

RC
--
R. C. White, CPA
San Marcos, TX
(e-mail address removed)
Microsoft Windows MVP (2002-9/30/10)
Windows Live Mail Version 2011 (Build 15.4.3504.1109) in Win7 Ultimate x64
SP1 RC


"Paul" wrote in message

How would 640K relate to any IT hardware architecture limitation?

Ed

http://en.wikipedia.org/wiki/640K

Paul

 

[DGDevin]

"Char Jackson" wrote in message

I can't see complaining about drive space these days.

The price is nice, but I miss the old reliability. My first hard drive was
a 20meg monster, it never failed and neither did the next few. But in the
past couple of years I've had two drives die, and neither one had lived very
hard lives, they just expired with no warning. So now I run multiple
outboard backups which are at least cheap.

 

[Char Jackson]

"Char Jackson" wrote in message



The price is nice, but I miss the old reliability. My first hard drive was
a 20meg monster, it never failed and neither did the next few. But in the
past couple of years I've had two drives die, and neither one had lived very
hard lives, they just expired with no warning. So now I run multiple
outboard backups which are at least cheap.

I've had pretty good luck, personally. I had two Maxtor's fail in the
past year, but both failed slowly and provided plenty of advance
warning. One was nearly 7 years old and the other was just over 4
years old. I think I've mostly been replacing drives because they're
too small and not because they failed.

 

[Ken Blake]

I've had pretty good luck, personally. I had two Maxtor's fail in the
past year, but both failed slowly and provided plenty of advance
warning. One was nearly 7 years old and the other was just over 4
years old. I think I've mostly been replacing drives because they're
too small and not because they failed.

I've owned PCs and hard drives since 1987. I've lost count, but
somewhere around 15 in total, and several of them have had multiple
drives. And a bunch of external drives too. So approximately 25-30
drives over those years. Of all those drives, I've replaced only one
that failed, and it was in a laptop. And it's failure was probably not
the drive's fault; it failed because the laptop was used in a
situation that overheated it.

 

[KCB]

?

How would 640K relate to any IT hardware architecture limitation?

Ed

Ever use MS-DOS?

 

[Ed Cryer]

?


Ever use MS-DOS?

Not until quite recently. I earned my living programming in high-level
languages, mostly COBOL and FORTRAN; ICL OSs were George and VME. Also
with NCR on midicomputers.

On a ZX Spectrum I used Sinclair BASIC as well as the low-level
assemblers available. That was the most challenging of all. I had to
work in little gaps in the RAM memory, between the screen memory, system
variables and whatever was loaded. I feel most proud of having succeeded
there than all the major central and local govt. systems that I implemented.

Ed

 

[KCB]

?

Not until quite recently. I earned my living programming in high-level
languages, mostly COBOL and FORTRAN; ICL OSs were George and VME. Also
with NCR on midicomputers.

On a ZX Spectrum I used Sinclair BASIC as well as the low-level assemblers
available. That was the most challenging of all. I had to work in little
gaps in the RAM memory, between the screen memory, system variables and
whatever was loaded. I feel most proud of having succeeded there than all
the major central and local govt. systems that I implemented.

Ed

To power users(?) of MS-DOS, editing the config.sys and autoexec.bat to
maximize free memory (under 640K) was considered a fun challenge. Or at
least it was with my friends and I. <Alt> <F> <S>, <Alt> <F> <X> anyone?