Off the top of my head, having toggled it in so often (octal!)
16701 26 1272 352 5211 15711EOE after 50 years!
First 10 years as a professional softy were PDP11 assembler on SCADA systems. Most challenging ISTR was a cassette tape driver to run under our own realtime exec.
Yes, I'm aware of that. Wasn't there some difficulty in on-the-fly switching between the two instruction sets? I've the thing described in a DEC sales clossy from 1971.
FORTH is as much Reverse Polish (Lukasciewicz) as the older HP calculators.
-- -TV
You misunderstood my comment.
Let me rephrase ... I was considering a language without using reverse polish along the lines of FORTH's interactiveness.
I worked on 1802 based systems from 1980 to 1987 in Oman. It was all done in assembler, I actually wrote an 1802 macro assember using the very basic assembler that came with the RCA 1802 developement systems.
As you say it was a very odd instruction set. You could use any one of the 16 bit registers as the program counter which could lead to some very confusing (and difficult to debug) code.
-- Chris Green
MPE's VFX Forth family is a Forth compiler that generates optimised native code. It is available for most operating systems and deep embedded use. ARM64 support is coming.
Stephen
-- Stephen Pelc, stephen@vfxforth.com
I was working on SCADA in Oman from 1980 to 1987, the PDP-12 wasn't SCADA though, it was a in a hospital.
Well it wasn't 'difficult' it was just rather odd because the whole addressing structure changed, the PDP-8 had 12-bit addresses (i.e. memory banks were 4k) whereas the Linc machine had 10-bit addresses (i.e. 1k memory banks). It was just a single instruction to change from one mode to the other but the ramifications in addressing etc. could be quite interesting! :-)
It's all a very long time ago now!
-- Chris Green
Den 2021-01-22 kl. 16:03, skrev gareth evans:
I was at a fascinating talk in Madrid in 2015 by a guy at ESA and he said (I can't remember which - or both) that either Rosetta or Philae was coded i Forth, with design/coding started in 1994 and launched 10 years later.
... and can you take the compiled executing code and continue to develop it interactively? (For which I do not mean returning to a source file, editing it and then recompiling)
My thinking is that for any interactive program, there has to be some form of tokenisation to represent the source program, so why should not those tokens be the carefully-chosen unambiguous machine code instructions that execute the code?
I am sure that this is not an original idea and that it has been done many times over the years what with everyman and his aunts, uncles and family pets all having a go at computer programming, but it is an idea with which I dabbled back in 1986 and that now I have the spare brain capacity to continue with!
ie, considerable exhaustion resulted then with intensive programming in the daily round of my job and then attempting even more programming in the evenings.
1986? Half a lifetime ago, I being 70 next month! :-)
Have you looked at John von Neumanns's interactive JOSS language or its derivatives JEAN (found on ICL 1900s) and FOCAL (on DEC boxes).
It used real line numbers (so if the needed a line between lines 1.1 and
1.2 you just gave it 1.15 as its line number) and DO PART 3 ran all lines with with 3 as the integral part of the line number.It was easy to write and debug but conditionals couldn't have an ELSE branch because the condition was the last part of the statement:
2.55 A = B * C + D IF q = "yes"Speed wasn't bad because I think that, like many basics, the 'run' command may gave compiled (or at least tokenised) the program before running it, though I don't remember much gap between startind a run and getting results. I certainly preferred JEAN to BASIC.
-- -- Martin | martin at Gregorie | gregorie dot org
I don't see why not - IIRC FORTH originally came from the astronomy community, who used it to control telescopes.
-- -- Martin | martin at Gregorie | gregorie dot org
The compiled form includes the names and the compiled code. That's all.
Yes, you can keep on developing interactively. The tokens you need are probably the source code. The compiler is very, very fast, on most PC machines over 1 million lines of code per minute.
Stephen
-- Stephen Pelc, stephen@vfxforth.com
Philae used 10 RTX2010 CPUs, an ADSP21020 and two 80C3X devices.
Forth on an RTX2010 was far and away the best choice for performance and ease of test - the CPU was designed for Forth. An RTX2010 is a rad-hard 16 bit dual stack machine made by Harris/Intersil. It clocked at 8 MHz (probably rather less for flight use) and could address 1 Mb of memory.
These days rad-hard development focuses on processors in FPGAs.
Stephen
-- Stephen Pelc, stephen@vfxforth.com
Where is it made? I could not find that. I see "designed in UK", but that sould like Apple's "designed in California".
I don't know enough to comment here.
The documentation I've seen does look first rate. Still, competing against Arduino with the wealth of existing code and experience will be an uphill fight.
Elijah
------ only started using Arduino in the last six months
Around 1980 I was working on a TV character generator based on the DG Nova, another mini of that era. It had a rackmount main frame with a front panel with octal pushbuttons, rather than the imposing row of switches. About half a dozen manually entered codes, than it fired up the paper tape reader.
I wrote a disassembler in machine code, then did some analysis of some of the supplied programs.
There was an optional one or two hard drives, 1.25MB in those big IBM cartridges.
The (TTL-based) Nova actually went out of production while I was with the character generator company, so they made their own using AMD bit-slice microprocessors, AMD2901, if I recall correctly. I believe DG also did this for a later Nova model.
-- Joe
3D printers can definitely benefit from 32-bit microcontrollers, especially if you're going beyond the usual Cartesian or CoreXY kinematics.
_/_ / v \ Scott Alfter (remove the obvious to send mail) (IIGS(
Sony Inazawa, Japan
Well that's nice, I wish they could have done that with their other boards. Rather than all the omissions in the schematics, lack of electrical specifications (except for the Compute Modules), and rare official Broadcom documents that hardly look proof read.
For the Pi Zero you always seem to have to fall back on hearsay and/or a snippet of code from part of the Linux kernel if you can find it.
I get that this thing has some extra peripheral functions, but I'd say a lot of the applications this is aimed at could be achieved with an official and well documented bare-metal programming environment for the Pi Zero. Mind you, maybe the weird business model around the Pi Zero means they don't really want to encourage that.
-- __ __ #_ < |\| |< _#
Thanks.
Elijah
------ obviously did not read the comments on the blog
As I understand, it is not yet set up to use the Arduino IDE but should work fine with an install of micro-python. The Arduino family generally have MUCH less memory to work with.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.