Comment by justinlloyd | Hacker News Reader

Comment by justinlloyd | Hacker News Reader

Regarding the image:

I believe those are IBM 386 computers on the right based on the front of the machine, the megahertz display and the turbo button. The ones on the left are most likely 8086 machines. IBM brand I believe based on the front and the case design.

Several PC based BBS machines supported multiline operation without the need for mulitasking, though as modems got faster, supporting 16 modems on a single machine became impossible, and it was often cheaper to buy a new commodity desktop PC rather than a much more expensive machine with a 16-port serial card capable of handling the IO.

The machines might have been running a "DOS" flavour, but for the most part, the BBS software was the entire machine. Boot from DOS, but then what we consider DOS is effectively removed from RAM and the BBS software takes over.

The 386's might have been using DESQView or equivalent, and running a single line BBS as two separate instances on the same machine.

The machine would not boot either from network or harddrive. Network boot was mostly unreliable and a bad node on a token ring network would easily disable your entire ring. Internal harddrives would have been costly, power hungry and unreliable. I can almost certainly guarantee they booted entirely from floppy disks. There was no doubt a token ring network that kept them all communicating. A token ring could be wired to tolerate failure, but it drove the cost up.

Those are Miracom Courier modems on top on some machines, a mixture of 9600 and 14.4, and Hayes Smartmodems, on others, 2400 on the black & silver, 9600 on the all blacks. Almost certainly there are no 56k modems in that picture. And they would have almost certainly be plugged in to a dual port serial card that supports dual 16550 UARTs because the 16550 could handle up to 119kbps (118,500) IIRC (typing from memory, feel free to google it), and they had a 16-byte FIFO buffer so the CPU didn't need to poll or IRQ for data quite so often. Therefore they could support 56Kbps on two lines, though at that point, it would have probably be pegging the CPU with two users transferring files using a non-compressing XModem or ZModem.

Also, all the drives have been upgraded from five-and-a-quarter to three-and-a-half.

IBM sold computers in cheesy clone-looking cases, with turbo buttons? Do you have a source to support that? And you know for a fact they're only sold with 5.25" drives, because you say they were all upgraded? Sounds like overconfident and incorrect speculation, unless you can provide a link to these IBM models.

Also, supporting more than one modem at a time is literally the definition of multitasking. You can't do it without multitasking. Are you talking about not needing a multitasking OS, because the software does its own multitasking?

Weren't the IBM Ambra, Aptiva, and Vista brands efforts by IBM to out clone the clones? As I remember, these IBM brands from the 1990s gave the impression of an IBM computer, but they were all made by third-party companies.

I believe you are correct, and they had the perfect little IBM logo slapped on them to make everyone believe that. Those were found throughout banks in the UK for the longest time. I am not sure those machines in the image are IBMs, but the case is very reminiscient of IBM's machines. We know they aren't Dell (wrong logo placement) and we know they aren't Compaq (wrong size/shape logo). But this is a minor detail, clones or IBMs isn't really relevant.

Source? Yes. IBM. IBM made crappy clones. I am not saying those are absolutely IBMs, they could easily be clones. But I do recall IBM making a case very similar to that. I also speculated, it doesn't mean I was stating it as fact under oath.

"I know for a fact" - why so accusatory? No, I'm saying those particular machines were sold with 5.25" drives and then upgraded to 3.5" because none of the plastic matches.

"supporting more than one modem at a time is literally the definition of multitasking" - no, it isn't. You might as well claim that a single threaded application that can respond to the UI and also do processing in the background is multitasking when it clearly isn't.

"You can't do it without multitasking" - yes, you can. There is a vast difference between a multitasking OS or multithreaded program and simply polling a couple of devices or getting IRQ'd. Now this I know, "for a fact" and we can argue about that until the cows come home, but frankly, I have more productive things to be doing.

"Sounds like overconfident and incorrect speculation" Take five minutes and just chill. There was an awful lot of speculation in my post, and I was very clear to say "I believe" an awful lot, based on a grainy image, and a hazy memory of wrangling those kinds of BBS setups but apparently reading comprehension isn't your strong suit. This isn't reddit, you don't need to go full bore attack because someone said something on the internet you think might be wrong.

56k modems didn’t bring any extra speed on the answering end. You needed to dial into an isp who had a digital hookup into the phone network

I am not sure where you're going with that, but the statement about needing a digital hookup isn't correct. A 56k analogue modem running over POTS offered 56kbps in one direction, and between 1200/2400/9600bps in the opposite direction.

He is correct in that the ISP side needed to be digital. Typically it was done using a PRI line, which was basically a ISDN T1: 23 "lines" (B channels) plus 1 signaling channel (D channel.) The ISPs I am familiar with used "Ascend Max" equipment.

My memory is somewhat hazy on that. I do recall helping to hook up equipment (cannot recall the brand) to the T1 DSU/CSU for our multiline "adult" ISP, but that wasn't my area of responsibility at the time.

And I am going to take back my "56k in one direction" statement and say: I think we capped out at around 33k on the analogue to analogue connections even though they "supported" 56k so I am probably quoting our crappy marketing rather than our physical capability.

My understanding is there was no way to support full 56k between two analog modems, so one side needed to be digital. In reality, I never saw a 56k connection even in that configuration. It was always ~52k or lower, but maybe we just had crappy phone lines around here.

I am recalling getting "insane" speeds during site-to-site warez transfers on the BBS's, upwards of 200kbps a second, capping the CPU on most machines, but we also used shotgunned modems on the main trading hubs, and a packetizing ZModem-like protocol to separate the transfers into 64Kbyte chunks so I am most likely misremembering a number of misadventures and smushing a few of the details together.

That seems possible with compression? Most serial ports maxed out at 115,200 bps back in those days.

56k was achieved by removing one analog-digital conversion. Instead of your my computer (digital)-my modem (analog)-phone network (digital)-your modem (analog)-your computer (digital), it was my modem (analog)-phone network (digital)-digital trunk-your computer.

A 56k modem at a bbs would just work at 33.6k

Based on the cases, those are most certainly PC clones, not actual IBM brand systems. 286s and even Turbo XTs had mhz displays and turbo buttons, so there's no way to tell.

Yes, many machines had Turbo buttons (even 8086), and mhz displays. And they might well be clones. I am not going to argue on that one way or the other. I based my speculation on the belly front design of the case, and I notice several IBM specific displays in the image. If the machines were to be running DESQView, they would be 386. If running a multiline BBS application, they could be 8086, 286, 386, or anything else. But based on the age of the image, I'd hesitate to claim anything above 386. I doubt the dual modem machines are 8086 because the CPU simply wouldn't be able to keep up with the IO, so that becomes 286 or 386, and again, more likely 386.

Reply on news.ycombinator.com