... like in the PC AT, PC XT[1] or the Compaq DeskPro 386[2] that the article discusses didn't have those ports at all.
Those were instead on ISA expansion cards, just like the floppy controller that would often share a card with the UART controller for the serial interface.
IDE was just coming in (in the UK) in 1990. The acronym got updated to "AT Attachment" because "Integrated Drive Electronics" was generic, and it wasn't as if the older drives had no electronics on them. Much later when SATA showed up, the name evolved again as ATA became known as Parallel ATA to distinguish the two.
Before that, when you installed a hard disk you had to go into the BIOS to specify the geometry of the drive. 46 types were already defined, to match individual drives on the market. "Type 47" allowed -- required -- manually specifying the drive geometry in terms of cylinders, heads and sectors. So for a short while some traditional MFM or RLL drives would be informally classed as Type 47 because their geometry and capacity differed from earlier drives.
Yes, the earliest mainboards I know of with on-board I/O including ATA is around Socket 5, the first mainstream Pentium boards. Some slightly older Socket 4 boards (circa 1994) have on-board I/O, but they weren't as common.
My 486 and earlier systems have all I/O provided by ISA cards, other than the 5-pin DIN keyboard port which was standard since the original PC.
I remember my dad's Dell 486P/33 from 1991 had integrated IDE, but that was a fairly high-end machine at the time (the forerunner of their "Precision" workstation range).
Wow! Very impressive board, I had no idea. It's kinda cool how we can directly see some of the chips that would be on the SuperIO card but directly on the mainboard. Thanks for sharing.
White box systems didn't really acquire onboard I/O til the late 486/early 586 era, but it was pretty common on name-brand systems to integrate IDE/floppy/serial/parallel and usually video.
Sure looks like it, the commit that introduced this is from July 7th, the affected version (Apache 2.4.64) was released on July 10th. Today (as of writing this in CEST) is the 24th.
It looks like not even Arch Linux had that version in their repo yet (currently 2.4.63-3) [1]
That's interesting, the few times I tried playing with whisper, I had the impression that YouTube style videos or random cellphone videos was something it did particularly bad with (compared to movies). My guess at the time was that most of the training material might be sub titles and raw screen plays.
The videos I tried to transcribe were also Mandarin Chinese, using whisper-large-v3. Besides the usual complaints that it would phonetically "mishear" things and generate nonsense, it was still surprisingly good, compared to other software I played around with.
That said, it would often invent names for the speakers and prefix their lines, or randomly switch between simplified and traditional Chinese. For the videos I tested, intermittent silence would often result in repeating the last line several times, or occasionally, it would insert direction cues (in English for some reason). I've never seen credits or anything like that.
In one video I transcribed, somebody had a cold and was sniffling. Whisper decided the person was crying (transcribed as "* crying *", a cough was turned into "* door closing *"). It then transcribed the next line as something quite unfriendly. It didn't do that anymore after I cut the sniffling out (but then the output switched back to traditional Chinese again).
The original IBM PC used an Intel 8048 microcontroller inside the keyboard and an 8255 I/O controller on the main board to communicate with the keyboard.
The PC AT (which had an 80286), later replaced the 8255 with an 8042 microcontroller too. It was running firmware, so re-purposing it for a Hodge-poge of other tasks became trivial. A single GPIO pin was used for masking the A20 line and another for handling CPU reset. Having a total of 24 programmable I/O lines, I guess this could have been done with the 8255 too, but the microcontroller probably allowed simpler interfacing with the CPU and bought them more flexibility for future expansion.
For the article, this is mostly irrelevant. As somebody else noted, the Xbox wasn't supposed to be able to toggle/mask the A20 line, but later x86 CPUs had already integrated the A20 masking feature into the CPU itself and exposed an A20 control line. The Xbox simply tied the A20 enable line to a fixed potential, the hack described in the article requires a simple hardware modification to change that (https://xboxdevwiki.net/File:Haxar-a20m.jpg).
From what it looks like, the difficulty shouldn't be in getting a board made or soldering them, but sourcing the parts. Other than the odd 74 series ICs, the design includes some (I think memory?) chips that are nowadays more difficult to find than the 6502 & 6520.
If you remove the "blog." subdomain and go to https://pipetogrep.org/ you are greeted by a typical "about me" page. It includes a link back to the blog, as well as a mailto link, and links to GitHub and LinkedIn profiles.
> This feels like a nothingburger. There isn't even a hint of [...] a demo or anything.
The visible human project has been around for quite some time and has, in fact, accomplished their initial project goals. And they do have a demo application, complete with a "guided tour", interview snippets and everything, all on a single CD-ROM! I remember playing around with it at the local library in December 2003.
reply