First Dual-Channel IDE? | Hacker News Reader

First Dual-Channel IDE?

os2museum.com

77 pointskencausey29 comments

For all the confused devs under age 30 in here, this is IDE meaning Integrated Drive Electronics, what later became known as ATA and then retronymmed as PATA.

I was really expecting to see a VScode plugin haha thanks

Even for devs in the late 30s, the confusion is same. First thought it was about Integrated Development Environment, was wondering what channel means here. To realise the IDE mentioned here, it took some time. :)

Kids those days. When they buy a new phone they only look at the number of megapixels /s. People who bought computers back then (except those who just wanted an expensive toy) knew what dual channel IDE is, what does it mean if a drive is master ( sorry) or slave (sorry) and what the interrupts are.

You just reminded me of the jumpers to configure master slave in IDE drives. :)

While they were annoying to deal with, took up a lot of space, and had fixed configurations...

I do miss me some ribbon cables

For the devs above 30 IDE could apply to Integrated Development Environment as well.

Since the context is MS-DOS, Turbo Basic, Turbo Pascal, Quick Basic, Quick Pascal, QBasic, Clipper, FoxPro among a few lesser ones.

Or given the Website name, Visual Age for Smalltalk, Visual Age for C++, Borland C++ for OS/2.

I'm 30+ and it actually took me some time to realise the other meaning, since I don't use IDEs (the software) much at all, but was very familiar with the IDE hardware interface from embedded systems work.

Although I am not sure about the history of the term Integrated Development Environment and its acronymisation, I don't think the early IDEs were called such --- the acronym IDE was basically exclusively used for the disk interface.

Delphi had an IDE and there were others. That's early 90s.

Did they call it that? There were IDEs of sort long before Delphi (staying with the Pascal family, even pre-Delphi Turbo Pascal sort of fits - letting you compile and run straight from the editor well back into the 80s), but most were not called an IDE.

The IDE term already pops up on Xerox PARC documentation.

https://en.m.wikipedia.org/wiki/Xerox_Development_Environmen...

You just need to dive into the Xerox Development Environment papers and documentation for the Alto (1977).

Good point. But what is a dual-channel Integrated Development Environment? ... Maybe an early form of pair programming?

EDIT: ;-)

I installed a bunch of these and loved them compared to the head-scratching misery of SCSI - which had far superior performance but also many more opportunities for things to go wrong. Those BIOS screens brought back memories of watching machines count slowly up to 8mb of RAM and thinking 'what a beast...one day I too shall own a machine of such power.'

I thought IDE was a nightmare, cheap and unreliable compared to SCSI. For instance, one has to fiddle around with jumper settings on each connected IDE device (master, slave!) whereas SCSI was a lot more intelligent and stable (and obviously a lot more expensive). I still have my old Adaptec 19160U somewhere, it’s a really beautiful piece of engineering.

You need to jumper each SCSI device to set an address too but I guess sometimes the defaults don't clash?

Larger systems would usually use drive bays with 80-pin SCSI, which was the regular 68-pin SCSI, plus automated address configuration (ie., you got the address assigned to the bay), plus power, all in one connector.

TIL! That's very cool

IDE drives had two jumpers to set master vs slave, SCSI drives had four jumpers or dip switches to configure the bus id... And having to terminate the bus was kind of annoying. I'm sure things got more convenient (but didn't late era PATA also have auto configure for master/slave?).

> IDE drives had two jumpers to set master vs slave

Or in some cases one to choose between to the two.

> but didn't late era PATA also have auto configure for master/slave?)

That was “cable select”, enabled with either an extra jumper or (common if the were separate master and slave jumpers) by jumpering for neither master or slave. Here one of the ground pins was not connected in the far end connector on the cable so a drive could decide what it was by inspecting the properties of that line. You could make a CS compatible cable by cutting that line on a non-CS cable.

I remember this being around at least as early as 1999, and drives coming set to CS by default at least once 80-conductor cables were common, though I also remember never really trusting it and explicitly setting Master/Slave as I wanted - setting one or two jumpers was no hardship (the vast majority of drives had a little jumper diagram on the printed label so you didn't have to look up the magic anywhere), and you had a touch more cabling freedom (you didn't have to care which connector master was plugged into, it was master irrespective) which was a small mercy with wide ribbon cables.

It wasn't so straightforward in the early days of IDE: https://www.os2museum.com/wp/the-dual-drive-ide-hell/

SCSI was more inteligent. The onlx issue people had was when 2 devices on the bus had the same SCSI id :)

As an Amiga user, Commodore's choice to go IDE instead of SCSI for the Amiga 4000 was almost seen like a betrayal and a sellout by some users, and it contributed to the sense of impending doom - the Amiga had long compensated for not keeping up in CPU speed increases by offloading everything, and SCSI was part of that... But of course it was also part of what made it expensive.

That unpopulated 68-pin connector reminds me of SCSI. Perhaps they were intending to make a combo SCSI + IDE controller, or reusing much of the PCB layout from a SCSI HBA?

Probably for running IDE over some external cable. The BIOS's error messages talks about "external enclosure"s.

Obscure piece of hardware! Why would anyone who needs up to 16 drives in a single system not use SCSI? IDE was never designed for more than two drives per bus connector (eg a 3.5” hard drive and a 5.25” zip / cd rom drive).

I could imagine it being useful for an imaging setup-- an OEM could use it to generate hundreds of identical drives.

SCSI was expensive. Also i don't know if your favourite Intel CPU was able to take advantage of the SCSI bandwidth for 16 devices. With IDE, AFAIK you can talk to only 1 device at a time.

Why does an IDE adapter need an FCC application?

In general, if you manufacture and sell a product that can radiate radio frequency energy, it needs to conform to regulations from governmental agencies like the Federal Communications Commission in the United States.

Reply on news.ycombinator.com