Old Computers – why do people bother

[Alan JacksonParticipant @alanjackson47790]

Apologies for double posting this on the wrong thread but I have been very happy with the operation of my old Dell Inspiron laptop driving my Stepperhead lathe via DOS and TurboCNC. It operarates on WIN 98 converted to DOS and can perform many operations, threading, profiling, dividing, etc. I do not see why it this system is not used further on converted lathes. It is simple and an inexpensive conversion method for a more electronic operated lathe.

Alan

[old martParticipant @oldmart]

If you stopped wearing the emery impregnated gloves while word processing, the symbols would last longer.

[Neil WyattModerator @neilwyatt]

Old computers are great for relatively straightforward, linear operations like CNC.

3D CAD is far more demanding, especially if you want to move your parts in real time.

If I want to stack 60-odd 32MB images or extract data from a half-gigabyte video I need a bit of muscle!

Neil

[Robin GrahamParticipant @robingraham42208]

A bit away from the original question I know, but I've often wondered about software bloat. When I were a lad (here we go!) I coded a a basic X-Window server which ran within the 640K 'conventional memory' of an IBM PC running MS DOS and still left plenty of room for data. I did it in Borland C – I think the most powerful graphics library function might have been 'line'. It was tight and focused code. That was before the time of OO programming.

Nowadays of course there must be library 'objects' which can do all sorts of things, some of which might be redundant for the programmer's purpose. I wonder if in the same way as designers of IC's find it more economical to produce multi-function off-the-peg chips rather than 'tailor made', programmers might prefer to use tried and tested 'objects' – which might have redundant capabilities – which still get compiled into code space. Maybe?

Robin.

Edited By Robin Graham on 06/02/2020 03:16:06

[Danny M2ZParticipant @dannym2z]

G'day. I have been holding back but following this thread with interest as the subject is dear to my heart.

The first PC that I built was from a kit in 1983, an Australian Microbee Microbee Wiki Linky

This little Z80 based machine inspired me to participate in a programming course for the mighty Z80 processor at Wangaratta TAFE. Just a weekly 150km round trip but well worth the effort,

No mucking about on this course, the teacher knew his stuff, Straight into the the architecture, the registers, interrupts and addressing the number of clock cycles to cycle a traffic light.. I wish that a modern microprocessor had as much access to the core and registers,

Built an expansion board soon after and moved most cassette programs into EPROMS after building an EPROM burner.

The Assembly Code tutorials were well worth driving through the fog from the Ovens River and so the Microbee still has a place in my workshop.

Currently monitors and graphs battery charge cycles via a home made A/D converter on the old green screen. Sure beats watching an ESV with a stopwatch.

I have a BBC Master Compact packed away btw, Never used as it came without a boot disc (demo model)

* Danny M *

[Mick B1Participant @mickb1]

Posted by Danny M2Z on 06/02/2020 08:47:12:

G'day. I have been holding back but following this thread with interest as the subject is dear to my heart.

The first PC that I built was from a kit in 1983, an Australian Microbee Microbee Wiki Linky

This little Z80 based machine inspired me to participate in a programming course for the mighty Z80 processor at Wangaratta TAFE. Just a weekly 150km round trip but well worth the effort,

No mucking about on this course, the teacher knew his stuff, Straight into the the architecture, the registers, interrupts and addressing the number of clock cycles to cycle a traffic light.. I wish that a modern microprocessor had as much access to the core and registers,

…

* Danny M *

'83? Yes, I remember that, and the Z80 processor. I was on a HTC course in Computer-Assisted Engineering at Brighton Poly. We had to take in a sine wave from a signal generator, A to D it into a Z80, add in a third harmonic and output the fundamental and resultant through a D to A into a scope.

Interesting exercise completed successfully, but I can remember nothing of the assembly code now…

[Nick Clarke 3Participant @nickclarke3]

Posted by Danny M2Z on 06/02/2020 08:47:12:

………I wish that a modern microprocessor had as much access to the core and registers,

 

The micro computer world has moved away from that with CP/M, Unix and MsDos/Windows from the 1960s onwards.

Once you accept the concept of an operating system you then cannot access the processor as it would mean that your programs were no longer processor independent and multi threading would become difficult, if not impossible.

If programmers took advantage of this opportunity (and many would IMHO) you could be in the nightmare scenario of many, many different versions of programs for different hardware setups, and needing to pay for a new one when anything changed. A program that only worked on certain versions of Intel processor, Last years Apple devices and not on any AMD chip except one version that is not the one fitted to your particular laptop? No thank you! – Having an OS may be bloated, but it is practical.

If you want direct access I believe PICs and similar devices will allow this, and many of these devices of today are in fact far in advance of the processors of 40 years ago – When you get to computers (laptop, Apple, *nix or Windows) it is just a different universe.

Mind you I am glad I am not the only one still alive and able to remember 6502, 6809 and Z80 chips – even if not the code!

Take care

Nick

 

Edited By Nick Clarke 3 on 06/02/2020 10:40:47

[TrickyParticipant @tricky]

Posted by Nick Clarke 3 on 06/02/2020 10:39:31:

Posted by Danny M2Z on 06/02/2020 08:47:12:

………I wish that a modern microprocessor had as much access to the core and registers,

The micro computer world has moved away from that with CP/M, Unix and MsDos/Windows from the 1960s onwards.

Once you accept the concept of an operating system you then cannot access the processor as it would mean that your programs were no longer processor independent and multi threading would become difficult, if not impossible.

If programmers took advantage of this opportunity (and many would IMHO) you could be in the nightmare scenario of many, many different versions of programs for different hardware setups, and needing to pay for a new one when anything changed. A program that only worked on certain versions of Intel processor, Last years Apple devices and not on any AMD chip except one version that is not the one fitted to your particular laptop? No thank you! – Having an OS may be bloated, but it is practical.

If you want direct access I believe PICs and similar devices will allow this, and many of these devices of today are in fact far in advance of the processors of 40 years ago – When you get to computers (laptop, Apple, *nix or Windows) it is just a different universe.

Mind you I am glad I am not the only one still alive and able to remember 6502, 6809 and Z80 chips – even if not the code!

Take care

Nick

I do not understand your logic here. As an OS such as Windows accesses the hardware you are saying you need a different OS for each version of Intel Chip. This is incorrect if you are talking about IBM PC architecture systems which were all developed from the Intel 8086/8088 chipset. Subsequent processors have added instructions and extended the addressing range but they belong to the same family. MS-DOS did not stop you accessing the processor.

User programs now run in a protected mode and are unable to access all the memory to stop them crashing the whole system if they fail. I first encountered this on IBM/370 mainframes where each program was given a virtual memory of 16MB which was the maximum size of partition and this included the OS!

In terms of operating systems on Floppy disks you are mistaken. MS-DOS was on 3 x 31/2" floppies, Windows 3.1 on 6, Windows for Workgroups 8 and Windows 95 0n 13 (I think, I can't find my copies). All NT based systems from NT/4 onwards (Win 2000, XP, Vista, 7, 10) were on CD only although sometimes a floppy was required to load drivers before installation could start.

Richard

[Nick Clarke 3Participant @nickclarke3]

Posted by Tricky on 06/02/2020 12:52:38:

As an OS such as Windows accesses the hardware you are saying you need a different OS for each version of Intel Chip.

Any processor features are handled by the OS. The whole point to the OS is that you don't need a separate version of your application depending upon which features are present in your particular processor and system – the OS handles that. You also only need hardware drivers for the OS and not for each application. Most systems have chipset drivers available to help with the differences between systems. There is not total backwards compatibility across all x86 family chips as features have been superceded and are no longer needed – for example the default startup in real mode found on the 80286 and the 80186 was just unusual!

MS-DOS did not stop you accessing the processor.

No but the only authorised way of doing so in the Dos API was to issue a software interrupt (frequently INT33 (INT21h). Any other way of accessing the hardware was unapproved and in theory could have changed with subsequent versions of the OS

All NT based systems from NT/4 onwards (Win 2000, XP, Vista, 7, 10) were on CD only although sometimes a floppy was required to load drivers before installation could start.

That is all very well if there is an operating system running and the CD can boot. This was not always the case with CD drives. If it was an empty machine with a non bootable CD you created 6 Boot Disks using a routine provided by Microsoft and used those to make the CD readable to load the rest of the OS.

Making XP install floppies

Apologies for the bold text, nothing personal, it is only to make it stand out when responding to stuff in the indent.

Take care,

Nick

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by Tricky on 06/02/2020 12:52:38:

In terms of operating systems on Floppy disks you are mistaken. MS-DOS was on 3 x 31/2" floppies,

Earlier versions such as MS-DOS 1.25 (also known as PC-DOS 1.1) came on a single 5 1/4 in floppy disk.

The OS on the first computer I used, an Elliot 803, came on a punched tape.  On second thoughts it was probably a boot loader which we followed by an ALGOL compiler also on paper tape.  Modern PCs are definately an improvement.

Russell

 

Edited By Russell Eberhardt on 06/02/2020 16:35:56

[SillyOldDufferModerator @sillyoldduffer]

Microcontrollers are the way to go when there's a need to program hardware. Arduinos are far better than a PC because they don't have an operating system! There's nothing between the programmer and the hardware. The IDE supports a C++ compiler and has many libraries to simplify control functions and device interfaces. PIC, STM, TI and other chips with similar capabilities available to taste. Also, boards such Raspberry, Beaglebone et al offer interesting possibilities, essentially because Linux can be stripped down to bare essentials and made 'real time'.

The Personal Computer parallel printer interface once provided a simple way of bit twiddling individual input-output pins. Used with a single-user/single-task Operating System like CP/M or MS-DOS, it was a reliable way of controlling external devices. Unfortunately the parallel interface ceased to be predictable for time critical switching from Windows 3.0 on, and access to the pins became much harder in Windows 95. By then, the parallel interface was fading fast and now only ancient kit has the original simple to use parallel port. Later versions of the interface are likely to be an emulation with unpredictable time performance.

Sadly, since Windows 95 each operating system and interface upgrade has made it easier to do tricky complicated things like high-speed data-transfer and ever harder to do basics like binary signalling. The modern world strikes again!

Dave

[Nick Clarke 3Participant @nickclarke3]

Posted by SillyOldDuffer on 06/02/2020 18:02:29:

Microcontrollers are the way to go when there's a need to program hardware. Arduinos are far better than a PC because they don't have an operating system! There's nothing between the programmer and the hardware. The IDE supports a C++ compiler and has many libraries to simplify control functions and device interfaces. PIC, STM, TI and other chips with similar capabilities available to taste. Also, boards such Raspberry, Beaglebone et al offer interesting possibilities, essentially because Linux can be stripped down to bare essentials and made 'real time'.

Dave

My own preference is for Arduino, in several different flavours, even though the concept of a 'sketch' as a pretty picture is a little toytown at times.

A friend who is an electronic engineer of over 60 years experience prefers to do his own I/O so likes the vanilla PICs – but while we have often had spirited conversations regarding their relative merits over a glass of 'something stronger than water' we both agree that while an amazing device that in its latest incarnation is a valid alternative to many low end PCs, the Raspberry Pi is really too much of a computer to be a true microcontroller – but we both still use them!

Take care,

Nick

Edited By Nick Clarke 3 on 06/02/2020 18:39:30

[Enough!Participant @enough]

Posted by Nick Clarke 3 on 06/02/2020 10:39:31:

Mind you I am glad I am not the only one still alive and able to remember 6502, 6809 and Z80 chips – even if not the code!

 

8080, 4040 ?

Incidentally, anyone dating back to 8080/Z80 days might find some memories here:

http://www.hartetechnologies.com/manuals/Processor%20Technology/

(the ALS-8 in particular did it for me)

and here:

http://www.hartetechnologies.com/manuals/

Edited By Bandersnatch on 06/02/2020 18:54:49

[Another JohnSParticipant @anotherjohns]

For those of us raised on 8-bit computers, Bill Godbout, a name from the past for the S-100 bus and CompuPro, died in a forest fire in California late 2018.

https://en.wikipedia.org/wiki/Bill_Godbout

(nicer story – back in high-school (pre University) I was making 1802-based computers (only diagnostic tool was a volt meter; one could single-step the 1802 and verify my wire-wrap wiring) I got a newsletter called Ipso-Facto, and it ends up that one of the executive of that newsletter is a model engineer, and ended up purchasing My Centec mill from him – it's only later I realized that we had coresponded by letter decades previously. Bert deKat is his name)

[Neil WyattModerator @neilwyatt]

Anyone wishing to use assembler on their PC should look at BBC BASIC for Windows.

As well as Pentium plus assembler (I think that's the name) and a very advanced basic that produces executables, it also includes full access to the Windows API.

So you can get right under the hood, if you want to.

Neil

[Neil WyattModerator @neilwyatt]

I haven't done AVR (the chips in Arduinos) for a while. I used to program them in assembler which annoyed most AVR users who love using C, but I just enjoy assembler – it's like solving crosswords.

Neil

[SillyOldDufferModerator @sillyoldduffer]

Posted by Neil Wyatt on 06/02/2020 21:04:06:

I haven't done AVR (the chips in Arduinos) for a while. I used to program them in assembler which annoyed most AVR users who love using C, but I just enjoy assembler – it's like solving crosswords.

Neil

Not sure why writing assembler for pleasure should annoy anyone, unless perhaps they distrust masochism! AVR vs PIC rows are another non-event in my book. I'm with Nick and other members of the Arduino club, though I admit Joe Noci is absolutely right about STM Nucleo.

Only one unforgivable heresy in computing: BASIC!

Dave

[Nick Clarke 3Participant @nickclarke3]

Posted by SillyOldDuffer on 06/02/2020 21:41:59:

Posted by Neil Wyatt on 06/02/2020 21:04:06:

I haven't done AVR (the chips in Arduinos) for a while. I used to program them in assembler which annoyed most AVR users who love using C, but I just enjoy assembler – it's like solving crosswords.

Neil

Not sure why writing assembler for pleasure should annoy anyone, unless perhaps they distrust masochism! AVR vs PIC rows are another non-event in my book. I'm with Nick and other members of the Arduino club, though I admit Joe Noci is absolutely right about STM Nucleo.

Only one unforgivable heresy in computing: BASIC!

Dave

Beats me why people object to masochism – beats me…. – Oh never mind!

As a sometime C programmer I still believe that it gives close to the speed and low down access but without the pain – however lets agree to differ,

But can you write the Hamlet line to be or not to be etc as a C program???

Regarding BASIC I can imagine the meeting at Dartmouth College where the IT department is describing, for the first time their Beginners All purpose Symbolic Instruction Code – Oh look everyone – the acronym is BASIC – don't tell me that wasn't forced!!!

[Frances IoMParticipant @francesiom58905]

I still have a 4040 chip set (you needed several chips to make a system) – should still work but has been sitting in its conductive foam for over 40years – bought as a spare in case a student project wrote one off.

Re S100 my home computer(or parts thereof) dating from late 70s is in the Science Museum reserve collection as it was one of a small number of 68000 based UNIX look alikes (IDRIS written by Ken? Plauger ?) which used the S100 bus (not a good combination!) – built in Newcastle by a small startup company and bought by another startup software company for which I wrote software as a consultant – we split the cost between us – we were asked by the Science Museum to let them have a system which was cannibalised from the working kit (I got an early 286-based system as a replacement – needed as I was distant from the company base)

[Danny M2ZParticipant @dannym2z]

Posted by Neil Wyatt on 06/02/2020 21:04:06:

I haven't done AVR (the chips in Arduinos) for a while. I used to program them in assembler which annoyed most AVR users who love using C, but I just enjoy assembler – it's like solving crosswords.

Neil

Neil, you are a bloke after my own heart.

Writing Assembly Language code, even for a lowly Z80 or 6502 processor is like a crossword puzzle, but when it all comes together how sweet is the feeling?

I used to belong to a PC geek club in Melbourne and the member's used to push each other with a monthly challenge. It was most educational.

Friend David once won a comp for a screen saver, it was about 80 bytes. He remarked that to include a variable time delay( default was about 60 seconds) would blow out the length of the program. I still have it on a floppy.

Actually my ancient Microbee has an EPROM containing a Z80 Assembler but also a Z-80 dis-assembler which came in very handy for studying/tweaking the code.

* Danny M *

[SillyOldDufferModerator @sillyoldduffer]

Posted by Danny M2Z on 07/02/2020 06:35:33:

Posted by Neil Wyatt on 06/02/2020 21:04:06:

I haven't done AVR (the chips in Arduinos) for a while. I used to program them in assembler which annoyed most AVR users who love using C, but I just enjoy assembler – it's like solving crosswords.

Neil

Neil, you are a bloke after my own heart.

Writing Assembly Language code, even for a lowly Z80 or 6502 processor is like a crossword puzzle, but when it all comes together how sweet is the feeling?

I used to belong to a PC geek club in Melbourne and the member's used to push each other with a monthly challenge. It was most educational.

…

* Danny M *

Nothing like an example to show why Neil and Danny enjoy pain.

This 'C' program outputs: abcdefghijklmnopqrstuvwxyz

Neil and Danny prefer to write:

To be fair, the Assembler instructions are for a modern Intel processor and these are unfriendly compared with the good old 8080, Zilog's souped up Z80, or the 6502. One reason is modern CPUs are much more complicated than early microprocessors, another is chip designers develop instructions to suit compilers rather than humans because 99% of programs are developed with interpreters, compilers and generators, not by people writing low-level code.

G-code is similar: although it can be read and written by humans, it's really meant to be generated by a CAD/CAM tool and read by a CNC machine.

Used to be quite easy for humans to write better code than a compiler, nowadays the opposite is more likely. 70 years of development have made compiler optimisations very clever indeed. As it needs an exceptionally good programmer to beat an ordinary compiler and she takes forever to do it, writing assembly is reserved for special occasions, one of which is doing it for fun. Otherwise, don't bother!

Dave

Dave

[TrickyParticipant @tricky]

Posted by Nick Clarke 3 on 06/02/2020 16:09:31:

All NT based systems from NT/4 onwards (Win 2000, XP, Vista, 7, 10) were on CD only although sometimes a floppy was required to load drivers before installation could start.

That is all very well if there is an operating system running and the CD can boot. This was not always the case with CD drives. If it was an empty machine with a non bootable CD you created 6 Boot Disks using a routine provided by Microsoft and used those to make the CD readable to load the rest of the OS.

Making XP install floppies

The system requirements for All NT systems from NT 4 onwards included a CD drive as the OS was supplied on a CD. Most PC bios support boot from CD so there is no need to use floppies to install on a clean PC. There are other ways for installing the OS but these mainly apply to organisations with a large number of PCs. See Deployment here,

Apologies for the bold text, nothing personal, it is only to make it stand out when responding to stuff in the indent.

My original post was trying to show the growth in size of the MS operating systems rather than be pedantic as to how you could load them. Certainly from Vista onwards a DVD drive was required if loading from optical media.

As an aside, when I was working at the Coal Boards Computer HQ in Cannock a friend of mine developed a system called Planning Information and Scheduling System with Optional Forecasting Features which was actually a good description of the system but management did not spot the acronym.

Richard

[Paul H 1Participant @paulh1]

It is quite refreshing to hear about the Z80 again, but does anybody remember MP/M the multi user version of CP/M. We ran in the 80s Altos microcomputers with 4 terminals and 2 printers on 1 Z80 powered machine. Running Wordstar, a spreadsheet and accounts. Worked really well.The Z80 is not dead at all. Look on Digikey for instance and you will find a range going from the original 40pin DIP at 6Mhz to square SMD packages at 20Mhz. There are also some sites offering Z80 kits to learn on with assembler.

Assembly language does seem to be getting talked about on the forums orientated to electronics engineers in a positive fashion at the moment. Personally I kept several NT2000 pcs and XP for legacy applications and to reuse if I manage to get to do a CNC conversion. Otherwise I stick to Ubuntu on recent hardware.

If anybody wants a really tight coding challenge have a look at the Padauk 3 cent microcontroller on Dave Jones' EEVblog Youtube channel and forum.

[Geoff TheasbyParticipant @geofftheasby]

Bandersnatch' mention of obsolete computers reminded me of my time with a NCR 500 visible record computer in 1968. When it broke, we would go to their office on the Leeds ring road to deal with that day's transactions. Ideal in theory, but the temptation to manually correct the entries on the card rather than ask my staff to do it on the machine was impossible for some to resist. Also, the abrasive iron dust from our foundry next door didn't help.

I also remember a statuesque young programmer called Marion de Vries.

[Roger ClarkParticipant @rogerclark]

I remember way back coding a flight sim on the ZX81 in binary, that was fun!!! , got it working though , since then gone through Pascal, Cobol for the 2,000 rollover, C#, Delphi, C++ et al, only bits left in the memory cells now

Roger

[Author]

Posts