"Clank" wrote in message
...
"tim..." Wrote in message:
Engineers didn't like creating designs for these ever smaller SIMs. It
was a real PITA.
But it was what Marketing wanted

Nonsense! We wanted to create smaller, better, cooler handsets
just as much as "marketing" - and the ridiculous credit-card
sized SIM was a major barrier to that.

well yes

but I was referring to the move from standard to micro to nano SIMs

whereupon inserting the SIM was changed to require removing the
IIRC for the the phone that I had that took a full credit card size SIM
you still had to fit it in under the battery

Indeed, and this was always a feature rather than a bug - it meant
we could confidently design the software stack to assume the SIM
it booted up with would never change (for as long as it was
running.) This mattered when you were coding for a 68k
derivative with memory measured in peanuts, and every byte
counted...

I don't recall working on "terminals" where memory was measured in peanuts

we had enough of it.

The problem was it wasn't very developer "friendly".

we still worked with PROMs and had to physically reprogram them each time we
changed the code.

tim

In message , at 15:03:06 on Sun, 14 Jul
2019, tim... remarked:

Nonsense! We wanted to create smaller, better, cooler handsets
just as much as "marketing" - and the ridiculous credit-card
sized SIM was a major barrier to that.

well yes

but I was referring to the move from standard to micro to nano SIMs

I wondered if you were, despite you replying in a subthread about the
CC-sized SIMs.

we still worked with PROMs and had to physically reprogram them each
time we changed the code.

Wow! Even back in the mid 80's we'd advanced to electrically
re-programming them, where I worked. Cutting those little links on the
PROM chip must have been really hard work for you.

In case you think I'm being facetious, I have seen ULA chips where a
small amount of [re]programming was done with a micro-scalpel.
--
Roland Perry

"Roland Perry" wrote in message
...
In message , at 15:03:06 on Sun, 14 Jul 2019,
tim... remarked:

Nonsense! We wanted to create smaller, better, cooler handsets
just as much as "marketing" - and the ridiculous credit-card
sized SIM was a major barrier to that.

well yes

but I was referring to the move from standard to micro to nano SIMs

I wondered if you were, despite you replying in a subthread about the
CC-sized SIMs.

we still worked with PROMs and had to physically reprogram them each time
we changed the code.

Wow! Even back in the mid 80's we'd advanced to electrically
re-programming them, where I worked. Cutting those little links on the
PROM chip must have been really hard work for you.

you know that I didn't mean that

I meant that we had to take them off the board to reprogram them

none of this downloading into in situ flash, lark



tim

"tim..." Wrote in message:
counted...I don't recall working on "terminals" where memory was measured in peanuts
we had enough of it.

Ahh, POCSAG+ and 8051 microcontrollers with 256 bytes of RAM, how
I miss thee; and yes, while the GSM days were better - much less
incredibly ugly reusing-the-same-buffer-a-dozen
-times-in-different-places, we even had something approximating
malloc/free - wasting good memory on being able to handle a
completely unnecessary feature like changing SIM with the power
on would mean memory not going on something useful. I wrote the
first WAP/WML browser outside the original Unwired Planet
reference implementation (it was still called HDML at the time,
in fact), and fighting against memory constraints was a constant
battle...

The problem was it wasn't very developer "friendly".
we still worked with PROMs and had to physically reprogram them each time we changed the code.

We could at least afford EEPROMs and In-Circuit Emulators. But
they were horrendously unreliable pieces of kit (not least the
flimsy ribbon cables that connected the ICE to where the chip
would have been) that stopped working if someone in the next room
sneezed, so one of my first gigs was building a test framework
that massively improved development productivity. I didn't
emulate the CPU, so native assembly couldn't be tested in it -
fortunately there wasn't much of that about even then - but built
a set of libraries that would allow the entire phone to be
recompiled and run on a Sun Sparc workstation, with all the
hardware devices simulated by mocks. As I recall - and it is
25-odd years ago - I had fun getting even the DMA-accessed
peripherals to emulate right, with no code changes to the phone
source, even if it was bit-banging them - using Sys-V shared
memory segments... (Interrupts were emulated using Unix
signals...)

Writing the mock instances of things like the LCD controller chip
(which I rendered to the workstation screen using X) bug-for-bug
compatible with the hardware ones was genuinely great
fun...

Gloriously happy days.


--

"Clank" wrote in message
...
"tim..." Wrote in message:
counted...I don't recall working on "terminals" where memory was measured
in peanuts
we had enough of it.

Ahh, POCSAG+ and 8051 microcontrollers with 256 bytes of RAM, how

well of course such devices were still available

and if you worked on a "cheap" or simple consumer product they would still
be used

but no-one used these for mobile (or cordless) phones

tim