New Pico2

Indeed. Also, picking on RPi would be particularly bad publicity.

It's nice how they've integrated more stuff into the SiP, so you can make things smaller. The trouble with the original Pico was that it took up more board space than say an ESP32.

Now, if somebody would do a module pin-compatible with the ESP32 modules that includes wifi, then things will be Very Interesting. Looks possible - there's a board with the RP2350 and an ESP32-WROOM side by side which shows the comparison:

What's nice also is that it's a Proper Arm Chip, rather than the weird Tensilica Xtensa CPU architecture in the base ESP32s. Some programming languages like Rust are better supported on Arm than on oddball CPUs. ESP32 have gone to RISC-V for some of the newer ones (but not all, so you can't get all the features with a RISC-V CPU) which is better, but with the RP2350 you get all the boxes ticked in one package. Aside from wifi, of course.

(apparently the wireless version of the Pico 2 is due end of the year - meanwhile people are putting RP2350s with ESP32 modules on boards like the above. Maybe RPi have do some certification work before releasing).

Theo

You sound like a young'n.

I suppose you'll now tell us about the joys of using punched cards. I don't envy you the joys of sorting a pile of punched cards that's been dropped on the floor ...

Well I could - and mention the joys of using a card saw to unblock the 1442.

If you prepared the cards with sequence numbers in columns 73-80 then you just dropped them into a sorter (if the site had a sorter) - otherwise it was less easy. It's the paper tape that someone trod on while you were feeding it into the spooler that was less fun.

No I do not miss those days, nor the overflowing box of waste paper beside the printer under the notice ...

TOWARDS THE PAPERLESS OFFICE

That adorned one place where I worked on developing machines with

That is why you put a sequence number in col 73-80. A few passes on the sorter and no problem.

Sometimes you got lucky, and the dropped deck would smear itself across the floor in such a way that, if you were careful, you could gather the cards up in sequence.

The sequence numbers in columns 73-80 were primarily for source code. Data decks were usually sorted on various other columns (e.g. customer number and/or date).

I soon gave up on punching sequence numbers in source decks. Our programs were subject to frequent modifications that required moving large blocks of code around, which meant that the sequence numbers became meaningless. You would have to re-punch the deck using a utility that re-generated the sequence numbers - and then you'd need to run the new deck through an interpreter so you could read what was on the cards. It usually wasn't worth the effort.

My first real computer was a PDP-8 with 4K of 12-bit core memory and a teletype machine with a 10 char/sec paper tape reader-punch.

It was actually useful, running the Focal interpretive language. I modeled a steamship propulsion system and got us some megabucks of business. Actually, I was just a kid at the time.

Cards were a great improvement over paper tape. Draw a diagonal stripe (color coded!) across each subroutine in case you ever droppped them, but don't drop them.

You could look at a 2-foot long card deck and *see* the program structure.

Computers used to boot a lot faster in those days. And had less bugs.

On 14/08/2024 17:55, john larkin wrote: <snip>

Not sure about that - in the late 80's our Vax8530 took about 30 mins for a reboot.

My X86 VMS instance in a VM on modest hardware, takes less than a minute to reboot

My first exposure was paper tape. I typed in my program one week, left it, and it came back the next with syntax errors. I treble checked it and could see none. I resubmitted it and the next week it ran,. Well sort of. there was a software bug , I gave up completely on computers until they turned up with keyboards screens and floppy disks.

MS-DOS was a lot faster, on a 2 MHz 8088.

The longest part was doing the startup beep.

---druck

Even quicker with a 8MHz NEC V30 (8086 equivalent)

One of these chips is sitting on my bookshelf. Maybe I'll put it on a breadboard and do some wirewrapping with some support chips and have a play. :-D

Boot times are always a somewhat interesting subjest - Yes, the Beeb and Apple II spend more time doing the Beep at startup than anything else. the Beeb had an advantage over the Apple II in that the filing system was in ROM where on the Apple II it had to boot from a small PROM on the disk controller card then load up DOS which took extra seconds...

Booting a Pi can appear to take a long time, but most of that is the stuff post-kernel being loaded. I have my own little bare-metal OS for the Pi and it loads and can be running in well under a second - unless I want USB and that, for some reason takes another second to initialise (it's not my code, I stole it from elsewhere - there's only so much you can do on your own on something as complex as a Pi these days)

I like the idea of the Pico 2 and my even look at it - I wrote my little OS initially on the 65816 then migrated it to RISC-V and this is a very nice RISC-V platform for it - decent RAM, nice peripherals and a 2nd CPU too.

I have also ported it to ARM (on the Pi v1) but it's ARM32 assembler and I'm not sure I fancy re-writing it (again) in Thumb32 instructions...

-Gordon

If you didn't care about not being in DOS, you could switch it on, hit Ctrl-Reset, and get dropped into a BASIC prompt right away. If a disk controller wasn't installed at all (it was about a year from the introduction of the Apple II to the introduction of the Disk II floppy drive), it'd boot straight to BASIC.

After I added a hard drive to my IIe, the longest delay in starting the computer was waiting for the drive to spin up. :)

I have/had a EEEpc that'd boot to a cutdown linux & open a browser (FF2.0) in about 10s from pressing "on" button. Of course very little on the internets these days will work with a basic html browser; loads of webshites seem to have to download a heap of javascript and large pictures before toggling to telling you your browser is no longer supported.

Remember, the Raspberry Pi’s whole raison d’être is about experimenting and having fun.

What happened to the idea of “silicon compilers”? As I recall, they would automate the process of turning a high-level schematic into a low-level transistor network.

Because lawyers.