VOGONS


What hasn’t been done?

Topic actions

Reply 140 of 164, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Yeah maybe, was shooting from the hip. I find I am so out of patience with the major distros these days and if I want to do something I'm using bionic puppy or some old knoppix

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 141 of 164, by Tiido

User metadata
Rank l33t
Rank
l33t
appiah4 wrote on 2024-03-15, 08:27:

I'd be more interested in flat panels with built in high quality upscalers than actual CRTs to be honest. There is probably a market for this, and it is not tech that does not exist - most good TV sets from late 2000s had this feature.

It would be good if there was choice of which algo to use but nobody offers such an option... There are things like game mode but it won't still let you pick an actual algo and balance any tradeoffs...

But as far as the stuff in TVs goes, it usually comes with a cost in latency (and depending on game it can be most unacceptable). It is very hard to scale when you have only seen one part of the image, and computer monitors with emphasis on low latency usually have pretty poor scaling because of it. When you get to see all of the image and then work with it you can do a lot better but then you have 1 frame of latency, and when you want to make sure stuff looks ok in motion too you need to see several frames... and then you really get into problems with interactive content and in particular action games (which are the only type of game I play personally)...

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 142 of 164, by InTheStudy

User metadata
Rank Member
Rank
Member
BitWrangler wrote on 2024-03-15, 14:24:

Yeah maybe, was shooting from the hip. I find I am so out of patience with the major distros these days and if I want to do something I'm using bionic puppy or some old knoppix

Right with you. On the plus side, KDE has finally come good again after the v4 debacle - I'm running Ubuntu Studio for audio work and they are Plasma based. On the flip, Gnome is just a disaster assembled from bad UX decisions. And Ubuntu as a whole - snap is just grim. Between aptitude, npm, pypi and snap I've got four parallel packaging systems in play on a machine - substitute anaconda for pypi if that's more your mood. Then there's IBM Red Hat wrong-footing CentOS users by pulling the plug on support mid-lifecycle.

Though I can't complain too much. Actual RHEL8 is great on my daily driver laptop, and Ubuntu Studio is doing an amazing job in the recording studio.

Reply 143 of 164, by appiah4

User metadata
Rank l33t++
Rank
l33t++
Tiido wrote on 2024-03-15, 18:01:
appiah4 wrote on 2024-03-15, 08:27:

I'd be more interested in flat panels with built in high quality upscalers than actual CRTs to be honest. There is probably a market for this, and it is not tech that does not exist - most good TV sets from late 2000s had this feature.

It would be good if there was choice of which algo to use but nobody offers such an option... There are things like game mode but it won't still let you pick an actual algo and balance any tradeoffs...

But as far as the stuff in TVs goes, it usually comes with a cost in latency (and depending on game it can be most unacceptable). It is very hard to scale when you have only seen one part of the image, and computer monitors with emphasis on low latency usually have pretty poor scaling because of it. When you get to see all of the image and then work with it you can do a lot better but then you have 1 frame of latency, and when you want to make sure stuff looks ok in motion too you need to see several frames... and then you really get into problems with interactive content and in particular action games (which are the only type of game I play personally)...

I get where you are coming from but there are a ton of <1fps Amiga HDMI converters out there these days, some use FPGAs, some even use SBCs like the Pi - so this should not be impossible to achieve without latency..

Reply 144 of 164, by Tiido

User metadata
Rank l33t
Rank
l33t

Those converters only work because they have to work with one single pixel clock domain and few very specfic resolutions, which dramatically simplifies what needs to be done plus they all have access to stuff like pixel clock and other support signals which are the real game changers.

A standalone monitor that has to work with all sorts of inputs is not going to get such a performance, except via digital inputs where pixel clock is available... and then it still comes down to whatever particular algo is used for the scaling operation and hence my comment about at least offering various algos and adjsutments, and perhaps comprehensive input formatting options to get things work with all the bazillion "off-spec" sources out there.

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 145 of 164, by darry

User metadata
Rank l33t++
Rank
l33t++
InTheStudy wrote on 2024-03-11, 02:22:

Well, I have an obscure one - and only because I suddenly finding myself needing it and unable to get it: An audio interface (USB or PCIe) that takes multiple SPDIF inputs and digitises them independently (clocks and all). This is less retrocomputing than it is synth/music, but since this is all for retro computer-music synths, I don't feel overly guilty.

It'd be nice if there was something like a four-input version of the Cubilux E52-Z10-19BK. That said, for my actual current use-case I'm just going to get a single one of those and call it done; and if my primary audio interface could handle 48k, even that wouldn't be necessary. Still annoying though. Means I can't be tempted to buy more 30 year old synths. Sad.

Have a look here Not so crazy idea : using a Raspberry Pi 4 with jackd , Zita A2J bridge and jack_mixer to make a software S/PDIF mixer

EDIT: some assembly required, batteries not included. 😉

Reply 146 of 164, by nd22

User metadata
Rank Oldbie
Rank
Oldbie

Windows 9X driver for Geforce 8 and later graphics cards. It has become impossible to find any card that is Windows 9X compatible at a reasonable price! Even for geforce 5200/radeon 9200 prices are going up and up! I was lucky to start collecting 9 years ago and now I have a lot of cards - today prices are completely insane!

Reply 147 of 164, by Dhigan

User metadata
Rank Newbie
Rank
Newbie

Windows XP driver for GMA 950 and its variations (GMA 3100, GMA 3150) to support for OpenGL 2.1
Still using the intel GMA 950 under windows XP ? Just think about all those indie games you couldn't run because your gpu only supported up to OpenGL 1.4. Linux support OpenGL 2.1 for these IGP !

The support for DX9 on GMA 950 is there but limited, I mean DX9 had evolutions, DX9c is actually what we consider DX9 today, with Shader 3.0 support, while original DX9 was Shader model 2.0 and that's what GMA 950 (and old Radeons before x1000) supported, and the GMA 950 didn't even fully support the hardware requirements for DX7 let alone DX9 (like hardware transform and lighting, and the vertex shader was done in software not hardware).

Progammers if you want to include old Intel IGPs and Radeons bellow the x1000 series, target the original DX9 with SM 2.0 and make sure it will work with the GMA 950 which will not pass a check for capabilities if it requires hardware TnL

.

Win 3.1 : HP Omnibook 425 + Toshiba T2130CT
Win 9x : Dell Latitude Cpx H500GT + Dell GX1
Win XP64 : Asus P5B Xeon

Reply 148 of 164, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

I was reading somewhere recently that there is OGL driver for that line for XP, but OGL and DX is exclusive, the DX driver doesn't work for OGL and the OGL driver doesn't work for DX.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 149 of 164, by Jo22

User metadata
Rank l33t++
Rank
l33t++

Another idea. Creation of a "Link Blaster" cable. To connect two vintage PCs via DB-15 gameport, by using the MIDI RXD/TXD lines.
Support for both SB-MIDI and MPU-401 would be neat. Allows multi-player games and file transfer similar to a null-modem cable.

The Atari ST users did something similar for their MIDI Maze gaming experience.
https://en.wikipedia.org/wiki/MIDI_Maze

Edit: Another idea (again). Writing a network applet for GEM or Free GEM and call it "Selector". 😉
Ideally, it would be compatible with the AppleTalk PC software, which had a DOS TSR version of "Chooser".

Edit: Or, how about writing of a minimal OS that acts as bootloader for DOS on UEFI PCs?
Maybe a modified version of Minix 1.x/2.x can be used? Which installs a TSR version of a BIOS-UEFI translator, then loads DOS?
Because, back in the day Minix could be loaded from a DOS system/from an image file.
Would be cool if it was possible other way round, thus.
Even if only basic text-mode worked initially (like a CRT terminal, maybe, if MDA support is too complex).

Edit: Another idea. Create a serial port card for DOS that's compatible with a Mac's RS-422 port (same pinout) ?
That way, PhoneNet adaptors (and their countless clones) could be physically connected to a DOS PC.
The AppleTalk software as such already existed on PC way back in the 80s.
Not sure if special drivers had to be written here, though or if any if any COMx port of DOS works.

Edit: Another idea. Modify vMac for DOS to support Hercules graphics card (rather than using VGA).
The Macintoshs were monochrome, anyway and vMac already supports a hacked resolution (640x480 vs 512x384 pixel).
If Hercules' 720x348 pixel aren't good enough to boot System, then the unofficial 640x400 "mode" could be used instead.
That C64 emulator used a similar approach. Alternatively, it would be possible to use AT&T graphics mode of 640x400.
That resolution would be good enough to simulate full 512x384 pixel native resolution.
Ideally, the modified vMac port would run on an V20 or 80286 processor, too, but that's optional if too much of a hassle.

Edit: How about a tiny NEC V20/V30 CPU board for ISA bus? Maybe by using I/O ports or DMA ?
Having that hard to emulate "8080 emulation mode" available to tinker with would be cool in a 286/386/486 PC.
It would help running old 8080 code that otherwise had to be emulated (CP/M emulators).
It also would make it possible to have an PC/XT emulator running on such a vintage PC.
Imagine having a DOS or CP/M "VM" software being running on MS Windows 1.0 or Digital Research GEM! 😁
In principle, something similar (NEC V or Intel 808x on a card) is possible already on the Raspberry Pi.
https://virtuallyfun.com/2021/06/04/elijah-mi … 30-on-a-pi-hat/

Edit: Or let's take another approach. Make an Z80/6502/68000-based CPU card for ISA bus.
Modified versions of popular DOS-based PC and console emulators could take advantage of this, maybe.
If not for speed, then for accuracy. For all three card types, there are cheap third-party clone chips available.
These processors all had been second-sourced or cloned thoroughly in the past 40 years..

Edit: Another idea. Development of a fast 80286 expansion slot that's quicker than ISA edge connector.
Or is electrically more stable, has less noisy and electrical capacity/load. Would be cool if it had buffered slots.
For the physical connection, how about the slots being used by GDR's Robotron EC1834?
I mean, it wasn't a great PC (no semi-intelligent ACT monitor anymore) but the slots look rock-solid, as should.
Not like the cheap ISA slot we grew up with. Maybe such a better connector can handle 25 MHz as a standard frequency? Does it have enough pins, also?
Would be cool if the 80286 platform had a bus that wasn't being a bottleneck for once.
Something that ran decoupled from the CPU speed all time but was still faster than 8 MHz.
Because, 16-Bit MCA bus wasn't that great, either. 25 Mhz because of the Harris 25 Mhz chip.

Edit: Or how about this? Create a PC-based rocket launcher. Maybe with a narrated countdown.
Would be cool if the control software ran on DOS, GEM or PC GEOS. Or DESQView /X. Or any other cool GUI!
The gameport (as an output) or a single serial port pin could be used to make it take off.

Last edited by Jo22 on 2024-03-21, 22:47. Edited 1 time in total.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 150 of 164, by InTheStudy

User metadata
Rank Member
Rank
Member
Jo22 wrote on 2024-03-21, 21:55:
Edit: Or, how about writing of a minimal OS that acts as bootloader for DOS on UEFI PCs? Maybe a modified version of Minix 1.x/2 […]
Show full quote

Edit: Or, how about writing of a minimal OS that acts as bootloader for DOS on UEFI PCs?
Maybe a modified version of Minix 1.x/2.x can be used? Which installs a TSR version of a BIOS-UEFI translator, then loads DOS?
Because, back in the day Minix could be loaded from a DOS system/from an image file.
Would be cool if it was possible other way round, thus.
Even if only basic text-mode worked initially (like a CRT terminal, maybe, if MDA support is too complex).

Given x86S this seems a little futile. This would be a project with a very small window of interest - people who want to run DOS, on modern hardware, specifically released between 2020 and ~2025.

Probably more practical would be to write JEDOS - not the Göpel thing, but a "Just Enough Disk Operating System" - a modern x64 kernel that can run on modern hardware, embeds DOSbox similar the way ReactOS embeds WINE, and provides API's through some kind of TSR to allow management of the "hypervisor" from inside the DOS environment. That would also support the 2020+ machines, but additionally the newer x86S hardware, and even potentially allow (the appearance of, if not the underlying mechanism of) "native" DOS on berry cake, fruit sand, five puns about Crash Override...

Reply 151 of 164, by Jo22

User metadata
Rank l33t++
Rank
l33t++
InTheStudy wrote on 2024-03-21, 22:42:
Jo22 wrote on 2024-03-21, 21:55:
Edit: Or, how about writing of a minimal OS that acts as bootloader for DOS on UEFI PCs? Maybe a modified version of Minix 1.x/2 […]
Show full quote

Edit: Or, how about writing of a minimal OS that acts as bootloader for DOS on UEFI PCs?
Maybe a modified version of Minix 1.x/2.x can be used? Which installs a TSR version of a BIOS-UEFI translator, then loads DOS?
Because, back in the day Minix could be loaded from a DOS system/from an image file.
Would be cool if it was possible other way round, thus.
Even if only basic text-mode worked initially (like a CRT terminal, maybe, if MDA support is too complex).

Given x86S this seems a little futile. This would be a project with a very small window of interest - people who want to run DOS, on modern hardware, specifically released between 2020 and ~2025.

Hi, good point, but I already knew this (more or less). 😁
In an ideal world, Real-Mode code can be run in 16-Bit Protected Mode. Which in turn still works in x64 long mode.
Concurrent DOS did support multi-tasking DOS applications on an 80286 in 16-Bit Protected Mode once.
https://en.wikipedia.org/wiki/Multiuser_DOS#C … _and_FlexOS_286

To be fair though, it only worked for well written (well behaved) applications.
Such as command line utilities that did rely on official DOS ABI and didn't play with the x86 segments.

But still, if the boot loader was able to simulate real-mode just good enough, something humble like MS-DOS 2.11 could be booted.
And over the years, MS-DOS 6.2x could be made run, as well, maybe. Too simple for games, but would still be cool for networking and other experiments.

Alternatively, there's still the possibility to pseudo-virtualize x86 instructions,
like a modern VM software does (or did, before HW-assisted virtualization became standard).
It would take over whenever a real-mode environment was needed.
Some sort of EMM386 in reverse, maybe, just without using a CPU feature in silicon (-> V86).

On a much smaller scale that EMU386 TSR for 80286 PCs does something similar,
it intercepts whenever an i386 instruction is being found.
Again, it's much simpler. It doesn't support any kind of 32-Bit Protected-Mode features.
Not sure if it can handle 32-Bit code and registers, even. EAX, EBX, ECX and so on.

Or other way round, the 80x86 emulator was made run by default and simply would pass unsupported instructions to a CPU virtualizer.
If an unsupported instruction/illegal exception was being triggered inside the 80x86 real-mode core, I mean.

Edit: Making an 8086 level real-mode environment isn't that hard per se. It's possible to run an PC/XT emulator on an AT or XT, even.
I once tried out QB8086 on vintage hardware, it really worked. I mean, an 8086 PC emulator written in Quick Basic (!) running on same hardware that it aimes to emulate.
Ok, it's not my product/work. The author was another guy, not me. I was merely silly/foolish enough to try! 😅
Still cool, though. Videos here: link

Edit: I mean, other users already had the idea of using a self-booting version of DOSBox.
But personally, I'd like to still be able to interface with the physical world, to be able to talk to real expansion cards and ports.
To be still being able to write my own programs in Power C or Turbo Pascal or Quick Basic that can talk to PCIe cards.
So making DOS boot as natively as possible would be really really cool.
That way, DOS knowledge and DOS as a platform would remain more than just a vintage piece in a museum's exhibition.

Edit: "X86S is a simplification of x86-64 proposed by Intel in May 2023 for their "Intel 64" products."
I find this line ironic/comedic, sort of. Because x64 (x86-64) is (was) AMD's invention, not intel's.
It had its chance with Itanium architecture and messed up. IA-64 never really catched on.
But that Wiki article makes it sound as if intel had any saying or relevance here. 😁

Edit: There's a recent article at hackaday.com about the topic: https://hackaday.com/2024/03/21/why-x86-needs-to-die/
In my opinion, though, x86 should be allowed to "die" peacefully in its full form and make place for ARM or RISC-V.

Edit: Some more related links: https://www.phoronix.com/news/Intel-X86-S-64-bit-Only
https://www.intel.com/content/www/us/en/devel … chitecture.html
What's interesting/worrying is the destruction of the ring scheme and the omitting of segmentation and i/o ports, I think.
Modern OSes like ArcaOS may still need some of these things. The way intel castr*tes x86 here, makes it usable for running Win 32 applications on Windows 10/11, at best.
But Windows doesn't need that, anymore. The current ARM version (port) of Windows does already have an x86 and x86-64 emulator included. Wine has an emulator for DOS/Win16 code, too.

Edit: Just noticed. The timeline shown on that intel site is interesting. No mentioning of AMD64, it's all about intel and its achievements.
It also mentions Windows 11 in the introduction on the top. The "simplification" process leaves me wondering, though, considering all the various SIMD instructions of x86.
So far, the only simplification I can notice lies within that propa.. technical material shown. There's little technical value, it reads more like an advertisement to me.
The technical manuals for the iAPX286/80386 from almost 40 years ago were being written on a much higher intellectual level, I think. *sigh* 🙄
To be fair, though, that PDF linked on the site does at least read like an average datasheet.

Edit: The whole story reminds me of the i376 processor: https://en.wikipedia.org/wiki/Intel_80376
Here, Intel once before tried to get rid of all the Real-Mode legacy. The old i376 datasheet/PDF was quite informative, I think.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 153 of 164, by Tripredacus

User metadata
Rank Newbie
Rank
Newbie

Getting more than 2.4 TB of addressable disk space on an MBR disk under Windows 7 or newer. I managed to get 2.4 TB back in 2017 on a 3 TB disk but never had access to any larger disks to see how far this could be pushed.

Reply 154 of 164, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++
Jo22 wrote on 2024-03-21, 21:55:
Another idea. Creation of a "Link Blaster" cable. To connect two vintage PCs via DB-15 gameport, by using the MIDI RXD/TXD lines […]
Show full quote

Another idea. Creation of a "Link Blaster" cable. To connect two vintage PCs via DB-15 gameport, by using the MIDI RXD/TXD lines.
Support for both SB-MIDI and MPU-401 would be neat. Allows multi-player games and file transfer similar to a null-modem cable.

The Atari ST users did something similar for their MIDI Maze gaming experience.
https://en.wikipedia.org/wiki/MIDI_Maze

IDK if there's a lot of point in that, bung up MIDI music and ability to use stick/pad for a network interface that's equally slow as the last choice of SLIP on XT class, and might be a quarter the speed on 16550 32bit machines and half the speed of 20Mhz plus with a 16450. At least copy Amiga SCSI or floppy networking to get some speed out of it.

Edit: Another idea. Create a serial port card for DOS that's compatible with a Mac's RS-422 port (same pinout) ? That way, Phone […]
Show full quote

Edit: Another idea. Create a serial port card for DOS that's compatible with a Mac's RS-422 port (same pinout) ?
That way, PhoneNet adaptors (and their countless clones) could be physically connected to a DOS PC.
The AppleTalk software as such already existed on PC way back in the 80s.
Not sure if special drivers had to be written here, though or if any if any COMx port of DOS works.

There may have been some attempts at that and RS485 networking in the 80s before wall to wall ethernet.

Edit: How about a tiny NEC V20/V30 CPU board for ISA bus? Maybe by using I/O ports or DMA ? Having that hard to emulate "8080 em […]
Show full quote

Edit: How about a tiny NEC V20/V30 CPU board for ISA bus? Maybe by using I/O ports or DMA ?
Having that hard to emulate "8080 emulation mode" available to tinker with would be cool in a 286/386/486 PC.
It would help running old 8080 code that otherwise had to be emulated (CP/M emulators).
It also would make it possible to have an PC/XT emulator running on such a vintage PC.
Imagine having a DOS or CP/M "VM" software being running on MS Windows 1.0 or Digital Research GEM! 😁
In principle, something similar (NEC V or Intel 808x on a card) is possible already on the Raspberry Pi.
https://virtuallyfun.com/2021/06/04/elijah-mi … 30-on-a-pi-hat/

Edit: Or let's take another approach. Make an Z80/6502/68000-based CPU card for ISA bus.
Modified versions of popular DOS-based PC and console emulators could take advantage of this, maybe.
If not for speed, then for accuracy. For all three card types, there are cheap third-party clone chips available.
These processors all had been second-sourced or cloned thoroughly in the past 40 years..

Z80s seem to emulate reasonably quickly on x86 given a programmer not brought up on MAME. So given the bus cycle misses and contention from ISA to CPU and back again, I don't know if you'd really get much speed advantage.

68000 cards for a coprocessor were a thing in 2nd half of 80s, you can see ads for them in Compute!

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 155 of 164, by Jo22

User metadata
Rank l33t++
Rank
l33t++
BitWrangler wrote on 2024-04-13, 02:36:
Jo22 wrote on 2024-03-21, 21:55:
Another idea. Creation of a "Link Blaster" cable. To connect two vintage PCs via DB-15 gameport, by using the MIDI RXD/TXD lines […]
Show full quote

Another idea. Creation of a "Link Blaster" cable. To connect two vintage PCs via DB-15 gameport, by using the MIDI RXD/TXD lines.
Support for both SB-MIDI and MPU-401 would be neat. Allows multi-player games and file transfer similar to a null-modem cable.

The Atari ST users did something similar for their MIDI Maze gaming experience.
https://en.wikipedia.org/wiki/MIDI_Maze

IDK if there's a lot of point in that, bung up MIDI music and ability to use stick/pad for a network interface that's equally slow as the last choice of SLIP on XT class, and might be a quarter the speed on 16550 32bit machines and half the speed of 20Mhz plus with a 16450. At least copy Amiga SCSI or floppy networking to get some speed out of it.

But it would be so nerdy and cool! 😎 And to many, the MIDI interface was never really be useful. Only a few PC users had a keyboard or a MIDI expander.

Let's just think about it. It's a fast, opto-insulated serial port that got never any use!
Wouldn't it be awesome to find an application for it? 😃

BitWrangler wrote on 2024-04-13, 02:36:
Jo22 wrote on 2024-03-21, 21:55:
Edit: Another idea. Create a serial port card for DOS that's compatible with a Mac's RS-422 port (same pinout) ? That way, Phone […]
Show full quote

Edit: Another idea. Create a serial port card for DOS that's compatible with a Mac's RS-422 port (same pinout) ?
That way, PhoneNet adaptors (and their countless clones) could be physically connected to a DOS PC.
The AppleTalk software as such already existed on PC way back in the 80s.
Not sure if special drivers had to be written here, though or if any if any COMx port of DOS works.

There may have been some attempts at that and RS485 networking in the 80s before wall to wall ethernet.

Likely, yes. AppleTalk as such had been supported on Mac, IIGS, IBM PC and Amiga (DoubleTalk), at least.
The Amiga version even had a high speed mode, if no Mac was on the network.
Edit: https://bigbookofamigahardware.com/bboah/product.aspx?id=918

I understand that many see this as being pointless. But personally, I just think that's fascinating.

I guess that's because in the 90s, in my place, we still had non-internet communications technology in mind when most of us saw the future.

Things like ISDN, Videotex (just think of Télétel/Minitel, used by our French neighbors), fax machines, videophones, GSM based mobilphones..

RJ45 and IP based networking was alien technology at the time. That strange "internet", too. Our media just barely touched the topic before 1995 or so.

Likewise, so did our national telco, Telekom at the time.
It still believed in his own service and proudly presented its new KIT standard in 1995/1996.
(Just think of it as an equivalent as CompuServe or AOL, with its own proprietary online services.)

The internet gateway was functioning, too, already, but most citizen here were still exploring the new fax technology and were being impressed by it.

In fact, by beginning of the 90s, many here didn't have an analogue telephone yet and have to wait for months to get a landline.
Because, the landline infrastructure wasn't being fully deployed throughout the country. Too few wires for too many people/houses.

Networking in businesses was another story, of course. 10Base2 and Novell Netware were being well known for years.
In 1993/1994, Windows for Workgroups 3.11 brought that topic "networking" up again.

Of course, there were also exceptions. Some users had been online browsing the internet since 1994 or so, by using then-young ISPs like 1&1 etc.
Many normal citizens didn't see internet and it's associated technology (RJ45 networking) before late 90s/early 2000s.

Having a fast modem, a null-modem connection between two and more PCs or a real (!) network card was still something special back then.

In these days, running an AppleTalk network through a PhoneNet-style wiring (we didn't have RJ11 wall plugs, but TAE) would have been exciting, still.

Edit: Here's a schematic for converting RJ11 (PhoneNet) to TAE.

https://www.brix.de/computer/localtalk_sx.html

English Translation

The only downside is, that in my place the wiring was 2-wire in many cases.

Our wires weren't always being twisted, sometimes in was just the equivalent to loudspeaker cable.

The states had 4-wire wiring being deployed in the houses, of which they merely used 2 wires.
The 2 "extra" wires thus could be used for PhoneNet. Very clever concept, imho!

Jo22 wrote on 2024-03-21, 21:55:
Edit: How about a tiny NEC V20/V30 CPU board for ISA bus? Maybe by using I/O ports or DMA ? Having that hard to emulate "8080 em […]
Show full quote

Edit: How about a tiny NEC V20/V30 CPU board for ISA bus? Maybe by using I/O ports or DMA ?
Having that hard to emulate "8080 emulation mode" available to tinker with would be cool in a 286/386/486 PC.
It would help running old 8080 code that otherwise had to be emulated (CP/M emulators).
It also would make it possible to have an PC/XT emulator running on such a vintage PC.
Imagine having a DOS or CP/M "VM" software being running on MS Windows 1.0 or Digital Research GEM! 😁
In principle, something similar (NEC V or Intel 808x on a card) is possible already on the Raspberry Pi.
https://virtuallyfun.com/2021/06/04/elijah-mi … 30-on-a-pi-hat/

Edit: Or let's take another approach. Make an Z80/6502/68000-based CPU card for ISA bus.
Modified versions of popular DOS-based PC and console emulators could take advantage of this, maybe.
If not for speed, then for accuracy. For all three card types, there are cheap third-party clone chips available.
These processors all had been second-sourced or cloned thoroughly in the past 40 years..

Z80s seem to emulate reasonably quickly on x86 given a programmer not brought up on MAME. So given the bus cycle misses and contention from ISA to CPU and back again, I don't know if you'd really get much speed advantage.

68000 cards for a coprocessor were a thing in 2nd half of 80s, you can see ads for them in Compute!

It's about the coolness factor, dummy! 😁😎
Sure you're right with everything, I won't disagree.
It doesn't make much sense from a practical point of view - that Raspberry Pi hat doesn't, either (except for V20/30 compatibility, which isn't fully understood by most emulators yet).

Except for an IBM PC, maybe, which is too slow to emulate a full-fledged Z80 at 100%.
On an 10 MHz and higher clocked IBM PC/AT, though, Z80MU and other CP/M-80 emulators begin to outperform a real Z80 PC.

Such a Z80 card would mainly be of interest for limited software environments, I assume.
Like GEM, Windows 1 or 2 and PC GEOS.
These GUIs do have memory limitations, so a physical Z80 would reduce amount of code in an emulator running on these GUIs (a good Z80 emulation core can be quite complex).

It would be pointless, but also cool to see an MSX1, Sega MasterSystem or CP/M emulator running on Windows 1, on an real IBM PC/XT. 😁
Let's think about it. Totally unreal. 🤯

Edit: Photo added.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 156 of 164, by appiah4

User metadata
Rank l33t++
Rank
l33t++

You covered the best bits 🙁