On Tue, 8 Jul 2014, Derrick Meury wrote: So, you do not want a 5150 (8088) 160K AND 320K? The 320K could do 160K, so that wasn't needed. IBM never really supported 8", but, in the declining final years of their DOS machines, they introduced 1.2M 5.25" drives (kinda a miniature 8"), 720K ("Quad") 5.25" drives (but not in USA), 3.5" 720K "shirt-pocket" disks, followed by 1.4M and 2.8M. Alas, the 3.9" drive didn't make it to market. Alas, IBM never actively supported more than 4 floppy drives. But, the 5150 only has five slots. Video FDC Serial Parallel Joystick To have more than 64K total RAM, you would have to give one of those up, go to "multifunction" (aftermarket) cards, upgrade to the "NEW!" 5150 (that supports 256K of RAM on the planar (IBM never called it "motherboard" due to use of the word "mother..." by the Panthers at Merritt College)), OR, . . . upgrade to a 5160! (EIGHT slots; some can be modified for 640K on the planar, often came with a giant TEN MEGABYTE! Fixed drive (IBM didn't called it a "hard" drive), but no cassette port, and the hard drive filled up one of the floppy bays.) The slot closest to the power supply was missing some buffering, so IBM gave a free serial port to keep the slot plugged up. -- Grumpy Ol' Fred cisin at xenosoft.com

On Tue, 8 Jul 2014, A. P. Garcia wrote: "FIXED" disk?? "FIXED" In a veterinary context??!?

On Jul 8, 2014 4:42 PM, "Fred Cisin" <cisin at xenosoft.com> wrote: :-) they should've just gone with 'winchester'.

On 07/08/2014 02:22 PM, Fred Cisin wrote: DASD? RMS? --Chuck

On 07/08/2014 04:22 PM, Fred Cisin wrote: Neutered disk.

On Wed, 9 Jul 2014, Tony Duell wrote: True! I started off with CGA, simply because the IBM monochrome monitor was not cheap, and old used B&W video monitors were. (CCTV, Sony AV and CV series) (I already had a few of them when the TRS80 came out, and all of my TVs in the 1980s were monitors with other devices doing the tuning - I got two very cheap because a "repair technician" hadn't noticed slide switches in between positions, and one where a flea market peddler was juat too annoyed with "where's the channel knob?" questions. So, I stuck with "sorta RS-170" video I couldn't resist a analog to digital converter, even without joysticks. Yes, IBM had a different name ("Analog Input Adapter"?) and didn't call it "joystick". I played with it with CDS photocells long enough to realize that it did not have enough bits of resolution to make a decent enlarging meter, nor enlarger color balance meter. In the IBM manual that I had at the time, there was a drawing that was obviously the Radio Shack Color Computer joystick, although that called for some trivial rewiring. Yes, the original was 16K to 64K, using 4116s. What I called the "NEW!" 5150 used 4164s, to provide 64K to 256K. 5160s used 4164s, to provide 64K to 256K, but some?/all? could be trivially modified to use two rows of 41256 plus two rows of 4164, to provide 640K. The first aftermarket board that I bought for my 5150 was a Boulder Creek Systems 192K? ECC memory board. -- Grumpy Ol' Fred cisin at xenosoft.com

On 07/08/2014 07:28 PM, Derrick Meury wrote: Go hang out over at vintage-computer.com. They seem to have a mania for the things. --Chuck

Sure. And sometimes having a graphics dispaly was more important that clearer text. However, you can have both the MDA and CGA cards in the same machine, even in a 5150. If you're going to take up a slot to provide a printerport, you might as well put an MDA card there, which will give you said printer prot and a mono display output. You can have the CGA as well, and use that as the default display if you want to. I thought it was the 'Game Control Adapter' or soemthing like that. IIRC, it was just a set of 555-like monostables (actually a 558, which is effectively a chip containing 4 555 times wired as monostables). The joystick pot is the timing resisotr (there is probably a series resisotor on the card) and there's an timing capacitor.Bascially, software triggers it, starts counitng up, and stops when the monostable times out. I am not sure what resolution you could get if you wrote your own software (ie. how high a count you could get with the standard components), but possibly if you tweaked the R's and C's on the board you might get enough resolution to read out an LDRR for this Not just a plug change[1]. The CoCo joystick inputs are genunie voltage inputs, the joystick pot uses all 3 terminals. Track wired across the 5V supply, slider to the input. The IBM (and Apple ][) joysticks are varable resisotrs, betwee nthe +5V line and the input. It is, of course, tricial to rewire the Radio Shack joystick when you realise this. [1] as an aside, one of my favourit cartons shows a chap handing over a car radio for reapir. The caption is 'I think it needs a plug change and the valves ground in' Somewhere I have a fairly high end joystick for the Apple ][ which uses all 3 termianls of the pots and uses said pots to control currnet sources that are fed into the Apple inputs. Give the operation of the 558 (discharge the capacitor, wait for it to reach a certain voltage), I think this would work very well. It would obviosuly also work with the IBM Game Cotnrol Adapter AFAIK all could. Certainly the oldest schematic I've seen in the TechRef shows the necessary parts fo the modification (the socket for the 74S158, etc). Later on, IBM actually seem to ahve sold a 256K-640K Planar for the 5160 (which is, of course, the smae board, wit hthe 74S158 soldered in, pads E1 and E2 jumpered, anf the approrpatie RAMs fitted). Ar at least my Techref has an entire set of schemaitcs for it (which are almost identical to the previous pages coverign the 64K-256K planar). Nwo, IIRC, pads E3 and E4 o nthe 5160 planar were also a jumper on an input of the RAM mampping PROM. So there are actally 4 pssible memory cofigurations -- the 2 we all know about (4 rows of 64K chisp, 2 rows of 256K, 2 rows of 64K) and 2 othes. I think one of those was to have 4 rows of 16K chips (4816s) for a 16K-64K XT Planar. I can't remmber wha the last one was. Something like 32K rows??? -tony

On 07/08/2014 08:23 AM, Sean Caron wrote: That leaves you in the position of having to find MCA cards and ADF files for any that you might purchase. Not to mention special floppy and hard drives, should they fail. I like boxes that can be loaded with more than one operating system and have a mixture of expansion slot types. Some of the Socket 7 boards are really good. My favorite for this type of work is an ASUS P5A with a K6-2 CPU. Accepts larger IDE disks, can run BSD, Windows and DOS just fine and ISA cards are still relatively plentiful. But that's me--and my needs are probably very different from the original poster's. --Chuck

On Tue, Jul 08, 2014 at 06:55:34PM +0000, Rich Alderson wrote: Nicely played, in both posts! :-) You can't go wrong with (MS-DOS-like) DOS. It runs fine on modern hardware. I do my main software development daily on a DR-DOS V8 system and I just don't want to hear the whining of some kid who hated using it for a single day so much that he had to write an article complaining about it. DOS works as well as it ever did, except INSANELY faster and with basically infinite disk space. John Wilson D Bit

On Jul 8, 2014 4:55 PM, "Joseph S Barrera III" <joe-ml at barrera.org> wrote: about how slow and painful Windows 95 was on it. I'm looking at rebuilding it as a strictly DOS machine. I have copies of several Microsoft DOS apps (I worked there at the time) but I'm curious what else you have in your toolkit. the simtel collection is pretty sweet. all kinds of goodies...

On Tue, Jul 08, 2014 at 01:55:59PM -0700, Joseph S Barrera III wrote: Well I think we each build a little nest to fit our own personal needs. So for me the most important things are Turbo Assembler V5.0 (which I bought back when that was still possible -- there was talk of making it freeware ages ago but AFAIK that stagnated) and the OpenWatcom collection (V1.9 is available for free from www.openwatcom.org). Other than that I use a lot of homegrown stuff that's of interest only to me -- a text editor that I started writing using EDLIN in 1983 (and is now built in to E11), terminal program, MAKE replacement, PIC programmer stuff, etc. Each program has the 5% of a commercial program's features that I would ever actually use, plus a bunch of stuff I needed that the commercial programs don't have. I'll bet lots of people do the same ... especially when they have three decades to do it. John Wilson D Bit

On Tue, Jul 08, 2014 at 03:19:08PM -0400, John Wilson wrote: [...] Prompted by this thread, I fired up FreeDOS up on an i7 with 8GB of RAM and a 2TB disk. It nearly had kittens, and the cluster sizes were silly. But there's still about 1.9998TB left on the disk. Now what do I do with it? :) Somewhat curiously, it reported the full 8GB of RAM as available, although it's not entirely clear to me how DOS apps can make use of that short of switching into long mode and bootstrapping a real operating system.

On Wed, Jul 09, 2014 at 12:24:58PM -0400, John Wilson wrote: [...] That's a bare-metal mentality, which doesn't translate well to how things are done in the Unix world. The short answer is that if you're processing data that doesn't fit in your available address space, you're just going to have to spill over to a file and let the disk cache sort it out. However, you can use mmap()/munmap() to map files into and out of your address space and use them like memory. The long answer is basically a lecture on modern virtual memory operating systems in which Linux and Windows are both held up as typical examples of the genre, but I don't currently have the time or inclination to give it.

On Wed, Jul 9, 2014 at 6:18 PM, Peter Corlett <abuse at cabal.org.uk> wrote: Nothing in the architecture or implementation of "modern virtual memory operating systems" (where "modern" hasn't really changed significantly since the mid-1970s) precludes what John is talking about. The Windows Address Windowing Extension (AWE) API he mentions is an existence proof. That Linux, xBSD, etc. on x86-32 don't provide anything directly equivalent only demonstrates that there's little demand for such a capability (probably due to the widespread availability of x86-64), but not that there's anything fundamentally difficult about it or that it is in any way incompatible with "modern" operating system architecture or implementation. Eric

On Wed, Jul 9, 2014 at 10:24 AM, John Wilson <wilson at dbit.com> wrote: Did you have the "Lock Pages in Memory" privilege? I think you have to enable that with the policy editor. And I couldn't find a Linux equivalent. Short of writing a specialized device driver to allocate the memory and allow it to be mmap()ed and munmap()ed, I think the closest 32-bit Linux equivalent to Microsoft's Address Windowing Extensions (AWE) API would be to use mmap() and munmap() on a non-disk-backed file larger than 4GB, such as a file in a tmpfs filesystem.

On 8 July 2014 21:19, John Wilson <wilson at dbit.com> wrote: Yes, I thought that, too. :?) :-o Did you RTFA? It wasn't like that at all. Blame the subeditor for the title. He's an old-timer -- one of the first things he does is get WordPerfect, which he rhapsodises about. He points to the legit free download of MS Word 5.5 for DOS (which for laughs I grabbed and got running under ``dosemu'' last night) but his preference is clear. This guy is from the era of MS-DOS, as I am myself, and he's taking a trip down memory lane... which I've done, too. I quite liked the piece. It has some insightful stuff -- like he got a TCP/IP connection going over a packet driver and Crynwr, which is a non-trivial exercise, particularly today -- and got a *current* DOS WWW browser working. This was a more dedicated exercise than I've ever bothered to do. Getting DOS networked wasn't hard, but TCP/IP on it is not child's play. I bet it does! Do you use a multitasker, JOOI? DESQview or something? -- Liam Proven ? Profile: http://lproven.livejournal.com/profile Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)

On 9 July 2014 18:11, John Wilson <wilson at dbit.com> wrote: Well, for all that Robert J Sawyer & George R R Martin swear by WordStar, I normally swore *at* it and think that 30y later it's eminently whinable-about. :?) OIC. Find it hard to live without, myself, now I've got used to it. Going app-launcher's task swapping capabilities sometimes. :-o -- Liam Proven ? Profile: http://lproven.livejournal.com/profile Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)

On Wed, Jul 09, 2014 at 11:18:49PM +0200, Liam Proven wrote: I never really used WordStar, but wasn't the Turbo Pascal editor supposed to use sort of the same command set? I remember some wonderful line in the TP manual that claimed that with practice, the commands would become an extension of your own mind, or some mildly creepy thing like that. So if I have that right, *they* liked it. I had a text editor and a homegrown RUNOFF-like thing that was good enough for what I was doing. Never looked into that. I was always sorry there wasn't any attempt in DOS itself (apart from some overseas version of V4 maybe?) to do multi-tasking, since it would have been easy enough to do at least what RT-11 does. I was always careful to retry my GETBLOCK calls endlessly, just in case in the future (or even the present, thanks to pop-up TSRs), the largest available block turns out to already not be available afterwards, if some other thread is nibbling away at memory while I'm trying to get some. John Wilson D Bit

the only dos i ever used was ms-dos. also most computers i ever found used ms-dos and the fact that all computers were all over the board with hardware configurations. some computers had onboard controllers and audio and even on board processor and ram slots and some computers used add on cards for everything. the other thing i find is that some add on tape drives were never the same size as the slots that were in the computer where some were a smaller drive with a bigger bezel and wouldnt fit in a 5.25" slot but also were to big for a 3 1/2" slot. for a dos computer i would go with something that has both size floppy drives, a tape drive, maxed out ram and processor, best video card it can run, professional sound card, and a joystick port and scsi card -----Original Message----- From: Liam Proven <lproven at gmail.com> To: General Discussion: On-Topic and Off-Topic Posts <cctalk at classiccmp.org> Sent: Thu, Jul 10, 2014 2:33 am Subject: Re: best dos machine On 10 July 2014 01:39, John Wilson <wilson at dbit.com> wrote: multi-tasking, Did you say you used DR-DOS 8? DR-DOS had multitasking from DR-DOS 7.0 and ViewMax could act as a GUI for it. Just full-screen, virtual-console switching type, but it worked well. IIRC though you do need to use DR's 386 memory manager, not the Microsoft ones that you get if you install any version of Windows later than 3.0. -- Liam Proven ? Profile: http://lproven.livejournal.com/profile Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)

On Wednesday (07/09/2014 at 07:39PM -0400), John Wilson wrote: We used to use a thing called DESQ (and later DESQview) on DOS to "multi-task" several standard DOS programs. Granted it was a third-party add-on and not part of the OS, it did pretty much work for a lot of DOS applications. http://en.wikipedia.org/wiki/DESQview Chris -- Chris Elmquist

Hear, hear! I wonder if I can find that spoof advertisment that appeared in "Microcomputer Printout"... Olivetti had an advertisement in which you "asked a computer" what was the best machine, and all these different machines came up with the Olivetti M20. So Printout ran a spoof, in which you asked a load of M20s what was the best machine, and they all came up with different answers, because they had all been programmed with different requirements. Are you sure? The keyboard I know as the "Model M" was the AT enhanced keyboard (and with a different plug and slightly different trim, the PS/2 keyboard). It is a total anachronism for the 5150. I'm not sure it would even work with it! Philip.

-----Original Message----- From: ben <bfranchuk at jetnet.ab.ca> To: cctalk <cctalk at classiccmp.org> Sent: Tue, Jul 8, 2014 11:59 pm Subject: Re: best dos machine On 7/8/2014 10:53 AM, Philip Belben wrote: Other than UNIX did any DOS really support hardware on the smaller machines? True Serial I/O was never handled under DOS. One can forgive the lack of DOS on say a PDP 8, not on a cpu like 8088. Ben.

On Tue, 8 Jul 2014, Philip Belben wrote: You are absolutely right. Good catch. The model M was available with several different connectors. Most of the original AT (5170) had a 5 pin DIN connector that looked the same as the 5150/5160 keyboard connector, but they are elctrically very different. I don't remember the name/model number of the 84? key 5150/5160 keyboard. It had a model M like feel (buckling springs), but different circuitry. -- Grumpy Ol' Fred cisin at xenosoft.com

I know of 2 (5 pin 'type A' DIN, 6 pin mini-DIN). What others are there? Actually, I seme to recall tht the signals are the same -- clock, data, +5V, ground. And on the same pins on both the XT and AT. The 5th pin, origianlly a reset outptu from the PC to reset the keybaord at power-on was, AFAIK, never used. Hwoever, the serial protocol is very different betwee nthe XT and AT keybaords. So while the clock and data are on the same pins, they keyboards are not compatible. The origianl PC keyboard is capacitive. The buckling spring flips a flap onto the PCB, increasign the capacitance betwene traces. The Type M is, beleive it or not, a membrane contact unit. The flap pushes the membrane layers together. I think the 84K PC/AT keybaord is capacitive, but it's been a long tiem since I've been inside one. -tony

Agreed. I've had them totally aparttoo (well, I _would_). I've also had a tyype M totally apart (to the point of cutting off the heat stakes (they were replaed with small nuts and bolts) and taking the flaps and membrane layers out. The one I worked on was certainyl a membran contact, not capacitive. I was going to suggest looking at the schematics in the AT TechRef, and then I rmmembered that for oome odd reason the only scheamtic fro the 84 key keyobaard is the one for the little LED PCB. The keybaord itself is not there. I have the 84 key keybaord somewhere, bot not to hand, or I'd open it up. The sens amplifier chip i nthe capcitive units is one of those square IBM metal cans, it's very distinctive. -tony

And don't forget the RS/6000 keyboard. The 7012 could exchange mice with the PS/2 (the RS/6000 mouse showed as "Logitech" under Windows), but not keyboard. They looked very, very similar, but the protocols the machines used to talk to them were very different. I found out the hard way after an office move... The AT keyboard had LEDs for num/caps/scroll lock from the first. I don't think there was anything in the 5150 keyboard protocol to handle them, so they would have had to do something different anyway. Philip.

i do know the model m also had a phone cable type connector. i happen to like anything made exclusively by ibm and not by any other company. -----Original Message----- From: Paul Berger <phb.hfx at gmail.com> To: ""General Discussion: On-Topic Posts Only <cctech" <General Discussion: On-Topic Posts Only &lt;cctech&quot;(a)classiccmp.org Sent: Fri, Jul 11, 2014 1:03 am Subject: Re: best dos machine On 2014-07-10 3:00 PM, Philip Belben wrote: The RS/6000 keyboard works well on a PC but the PC keyboard will not work on a RS/6000, after a bunch of the RS/6000 developers, who all also had a PS/2 in their office, where relocated we got a lot of calls for keyboards that did not work. The easiest way to tell them apart was the RS/6000 keyboard had a speaker on the underside below the LEDs, in fact now that I look I am typing on a RS/6000 keyboard. The XT version of the model M does not have any LEDs on it. Paul.

On Fri, 11 Jul 2014, Paul Berger wrote: That plant is still in operation, making the successor to the Model M. http://www.pckeyboard.com/ g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://scarlet.deltasoft.com - Get it _today_!

On 11 Jul 2014, at 10:58 AM, Guy Sotomayor <ggs at shiresoft.com> wrote: I second that, in spades. I picked one up a few months ago and was amazed by how much I've missed a proper clicky-keyed mechanical keyboard; I was particularly pleased to see the speed and accuracy of my typing approximate what it was decades ago. -- Christian Kennedy Ph.D. chris at mainecoon.com AF6AP | DB00000692 | PG00029419 http://www.mainecoon.com PGP KeyID 108DAB97 PGP fingerprint: 4E99 10B6 7253 B048 6685 6CBC 55E1 20A3 108D AB97 "Mr. McKittrick, after careful consideration..."

Now that I think of it, the original Apple ADB Extended Keyboard (the "wedge" with the rainbow apple in the lower left hand corner, Rev. 1) was very nice too. I'd probably slot that one above HP's. They cheapened the Rev 2s a little bit when they streamlined them... not terrible but not really remarkable, either. Best, Sean On Sat, Jul 12, 2014 at 5:18 PM, Sean Caron <scaron at umich.edu> wrote:

I'm still using an '87 M myself on my desktop, just needed a cleaning when I got it. Some of the older ones can be used on a PC\XT, such as the 1390120 and 0131 which are a bit nicer to use if you're not a huge fan of the PC\XT keyboard layout.

On 14 July 2014 16:19, Jules Richardson <jules.richardson99 at gmail.com> wrote: I remember switchable keyboards from the 1980s. Usually there was a little slide-switch underneath, labelled "PC" or more commonly "XT" and "AT". -- Liam Proven ? Profile: http://lproven.livejournal.com/profile Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)

On 14 July 2014 18:48, geneb <geneb at deltasoft.com> wrote: I only ever had 1 of those -- and it was the later cheap version that Gateway 2000 bundled, which didn't have great feel. I gave it to a friend of mine who lost the use of his right arm in a motorbike crash -- I thought he could program some macros onto the dozens of function keys, to make one-handed use a bit easier, but it was a bit beyond him, I fear... -- Liam Proven ? Profile: http://lproven.livejournal.com/profile Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)

I always wondered about this because I could swear that I saw at least one XT Model M in the day... Does anyone know if you can identify a switchable Model M from the keyboard cable being terminated into a modular type jack on the keyboard side? Or was going from the modular jack to a hard-wired PS/2 cord just a cost reduction at some point in time after the keyboard was introduced? Best, Sean On Mon, Jul 14, 2014 at 2:12 PM, steve shumaker <shumaker at att.net> wrote:

On 07/14/2014 01:41 PM, Hayden Kroepfl wrote: No matter, if it doesn't. Go look over on vintage-computer.com forum for a very inexpensive AT/PS/2 converter to XT using an 8 pin PIC MCU. It's so simple that it can be constructed on a small piece of perf board, but PCBs may still yet be available. A lot of keyboards could auto-switch between the two protocols. It usually starts with the keyboard sending a bunch of PS/2 NAKs. If the host doesn't respond, XT mode is assumed. The keyboard that I used for many years was a made-in-Germany NCR Cherry-switch model. 30 function keys, but no F11 or F12. Otherwise, very nice. Currently on a Model M. --Chuck

On Mon, 14 Jul 2014, Liam Proven wrote: Yep. BUT, do you remember any such FROM IBM?? -- Grumpy Ol' Fred cisin at xenosoft.com

I can confirm that it works as per I use one on my XT. There are some black cables for the Ms which have the 5-pin DIN although adapters do work on the more common PS/2 cables. The one I have here is a 1390120 (Determinable from the lack of LEDs the 1390131 has and the silver badge later models lack). The 1390131 which has LEDs also works fine, I've tested as I own one as well. Here's a picture of one on my XT: http://goput.it/i05.jpg

AEK and AEK II both used various Alps switches. The Design keyboard is where they quit using Alps (or any decent switches for that matter).

On 12 July 2014 17:18, Sean Caron <scaron at umich.edu> wrote: I have one of the UNICOMP keyboards. In fact I have a slightly older (seven years I think I've had it) PC/5250 version. PS/2 not USB, beige and very much not at all matching my current desktop (black tower, black mouse, black monitor, black speakers, black desk... beige keyboard). The thing is, if you will pardon my swearing: FUCKING FANTASTIC. I have a vision problem, so the normal cheap as sin, made in China for fifteen cents keyboards a really not something I can work with reliably, tactile and auditory feedback being very important to me. My 122-key Model M is pretty much the only keyboard I can actually get up to my full typing speed (and low error rate) on. Get one for yourself. They do customization on almost everything, and reasonably too. I think the only thing they might not do is specially labeled key caps. Cheers, Christian P.S. I have nothing in any way to do with UNICOMP, I am just a very satisfied customer. -- Christian M. Gauger-Cosgrove STCKON08DS0 Contact information available upon request.

The tactile feedback should be quite similar between those 2. Although I must admit I was disapointed when I took a type M apart and found it ws a membrane switch mechansim at heart :-(. I was also disapointed when I fond the HP9830 used keyswtiches. The smaller machines in the range (9810, 9820) used a wonderful electromagnetic system. Under each key were 2 spiral traces on the PCB forming a transformer. 2 keys were conencted in inverse series, so the induced signals in the secodnary 'windings' cancelled out. Pressing a key brought a metal disk onto the associated 'trasnformer', damping it, so the signal from the other one got through. Very reliable [1]. The main problem was it didn't easily hadnle mutluipt keypresses, which is I guess why it wasn't used on the 9830 (whioch had a shift key), but... [1] The electroncis is quite compleicated, but not hard to repair. The comapratoer IC in the sense circuit is hard to find now, though. The contact mechanism in the VT100 keybaord is also used in the VT5x, HP85 (etc), HP9816, some HP terminals, TI99/4A, early TRS-80 Mdoel 1's, etc, etc, etc. A lot fepends on how well it is debounced :-) Odd. I hate the feel of the emchanical HP-HIL keyboards (is that a 46020?) The laster one (46021) which is an odd capacitive membrane thing is perhaps a little better, but not much. -tony

On Sat, 12 Jul 2014, Paul Berger wrote: Yep. The folks that run the place are top notch too. The only thing these new keyboards lack are the removable key caps. You can replace the whole key, but not just the cap. They do offer color key sets too. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://scarlet.deltasoft.com - Get it _today_!

On Mon, Jul 14, 2014 at 4:12 AM, geneb <geneb at deltasoft.com> wrote: Are you sure about the removable-caps thing? I have 4 of Unicomp keyboards, and all of them have keytops that are in two parts (i.e, stem and cap). They may have changed recently (or since 2006, anyway - my newest Unicomp keyboards were manufactured in 2006).

On Mon, Jul 14, 2014 at 08:10:35AM -0700, Mark J. Blair wrote: [...] I had the opportunity to use a ~2009 vintage Unicomp when a friend imported one know for sure, but the Unicomp certainly felt as good as I remembered. It really is the kind of keyboard that the BOFH could use to club a luser to death, and then write his obutuary with it afterwards. So when I finally wore out my beloved decade-old Cherry keyboard in January, I decided to order a Unicomp, and found a UK seller via Amazon. The unit I received claimed to have been assembled in late 2013, and was a badly-assembled creaky thing with sharp edges on ill-fitting plastic parts, and failed within a couple of months. Comparing it to the other one is like night and day to the point I'm wondering if mine is counterfeit, and if so, who would actually go to the bother of counterfeiting Model Ms! (I tried to reject the goods as Not Of Merchantable Quality and the seller is giving me the run-around. I'm sure they'll change their tune when the court paperwork drops through their letterbox.) What's the build date on yours? I bet it's quite recent and akin to the junk one I've got.

On Mon, 14 Jul 2014, Raymond Wiker wrote: I just checked mine and it's dated October 2nd, 2002. I popped off the F12 key and it's a single injection molded part - the labled key portion doesn't come off the post. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://scarlet.deltasoft.com - Get it _today_!

I didn't know about that difference. Come to think of it, I have at least one RS/6000 keyboard - I shall have a look at it. The way I tell them apart is that the RS/6000 keyboard has an "ID" number in white in the trim of the LED panel - just above the label that says "Num Lock" iirc. I don't think I have ever seen an XT model M, even when I worked at IBM. Agreed that the existence of such a beast probably does refute my earlier conjecture that the Model M wouldn't work with a 5150, but it doesn't affect what I was saying in that paragraph - when they brought out the AT, IBM introduced the LEDs in the keyboard, so it is not surprising they changed the interface protocol at that point. Philip.

On Tue, 8 Jul 2014, Rich Alderson wrote: He did mention ISA. So, would an XT370 be a good one to seek?

On Wed, 9 Jul 2014, Rich Alderson wrote: admittedly, not in his first message, but: On Tue, 8 Jul 2014, Derrick Meury wrote: He said, "isa", instead of "ISA", but it looks like his shift key isn't working.

what i would like to do with the machine is run anything dos, games, audio, even a little chatting and such also run it like a business machine -----Original Message----- From: Chuck Guzis <cclist at sydex.com> To: ""General Discussion: On-Topic and Off-Topic Posts <cctalk" <General Discussion: On-Topic and Off-Topic Posts &lt;cctalk&quot;(a)classiccmp.org Sent: Tue, Jul 8, 2014 1:56 pm Subject: Re: best dos machine On 07/08/2014 11:53 AM, Rich Alderson wrote: Add a 1403 and some 729s and a MFCM and you'd be in business. // JOB // EXEC SILLY /& --Chuck

On 7/8/2014 11:53 AM, Rich Alderson wrote: If you really insist on DOS and want to run it, use the vm370 turnkey 6 pack system. ONe of the OS's that is supported is DOS. Rich can let us know how that 360/40 project goes. H seems a bit much, didn't it run in a lot less core? Jim

On 07/09/2014 09:07 AM, Paul Koning wrote: I seem to recall that the smallest variant of DOS/360 was a 4KB resident, but 8KB was more the rule. Lots and lots of transient phases loaded to do anything meaningful; e.g., $$BOPEN just to open a file--no permanently resident code for that. And people think that CP/M had modest storage requirements. I recall that COBOL and PL/I both took forever to compile. If you wanted to assemble a program with a bunch of I/O macros, that also seemed to take forever. On the other hand, using the small FORTRAN compiler (USA Basic FORTRAN) was pretty zippy. 128KB on DOS/360 would get you a sizeable background partition and two foreground ones. You could use the foreground ones for running interactive timesharing applications. Several universities did just that, hooking TTY dialins for running BASIC interactively. --Chuck

well your right but what i would like to do is have a killer dos machine and be able to have it run anything that was made for dos or any current home brew dos program. also i would like it to have all the ports in use with all kinds of cards like scsi, networking, video, sound, joystick, serial etc.. also want it with both size floppy drives, hard drive and a dat or some form of tape drive and maybe cd rom if i had to have it -----Original Message----- From: John Willis <chocolatejollis38 at gmail.com> To: General Discussion: On-Topic Posts Only <cctech at classiccmp.org> Sent: Tue, Jul 8, 2014 11:46 pm Subject: Re: best dos machine I'd be inclined to look for a 16MHz 80386 machine from the '92-'94 era, preferably with IDE and 8MB of RAM (or more). With IDE, you can get an IDE to CompactFlash adapter, which will be more reliable than any mechanical drive that will work with a machine from that era, and CF media is fairly cheap. Single-ended SCSI stuff is getting long in the tooth and hard to find these days, and ST-506 and ESDI stuff is rarer than hen's teeth, and pricey in working condition. ISA bus is desirable. Look for an early (non-plug-and-play) Sound Blaster 16 card, and for ease of data transfer, I'd recommend picking up a NIC too--I like the Intel EtherExpress 16TP or (if you don't mind using an AUI transceiver) 3Com 3C509. Version of DOS: 5.0 is the earliest I'd go with, for several reasons. I'd recommend 6.22 for maximum flexibility though, or IBM PC-DOS 2000, if you can find it. Get yourself a version of Windows for Workgroups 3.11, as it's easy to get Microsoft's 32-bit TCP/IP stack running on it, as well as Win32s for certain apps. If you do get a 386, I'd make sure it's a DX and not an SX, as well as not being one labeled "FOR 16-BIT S/W ONLY". I say 386 because it should still be slow enough to run a lot of timing-sensitive games and such. However, if you're like me, and you're more interested in running serious software (development tools, etc.), I'd look for a 90MHz Pentium or higher, with 16 or 32MB of RAM, a few PCI slots, and something like a Diamond Stealth 64 video card. Stick to the SB16 for sound though, or you'll run into lots of compatibility problems--Creative Labs was king in that era. On the topic, if your'e doing the DOS thing, why not build up a dual-boot system that also runs OS/2 Warp Connect? If you do your disk layout right, you get a huge amount of flexibility from this--and OS/2 runs each Win16 app in its own private memory space, which is great for stability, as well as being able to dish out more conventional memory to picky DOS games. I have a machine running OS/2 Warp Connect, DOS 6.22, and Windows 3.11 on a 233MHz Pentium with 32MB of RAM, a Sound Blaster 16, and an IDE-to-CF adapter with 2GB CF media. I even use it for MIDI sequencing occasionally, with good old Cakewalk 3. Has 5.25" and 3.5" floppy drives, as well as a 52X CD burner. An absolute pleasure to use, as Win3.11 pops up in about 3 seconds, and OS/2 Warp is rock solid. Feel free to contact me if you have any questions about writing good CONFIG.SYS files: at any given time, mine allows me to pick one of several configurations for different uses. jpw P.S., and not for the OP: Flames about top-posting will be redirected to /dev/null. Take your flames to Google for making it the Gmail default and making cut-and-paste so unusable on Android. On Tue, Jul 8, 2014 at 4:02 AM, Derrick Meury <rmx44 at aim.com> wrote: each machine is different from the hardware to the way it runs the software. some have the use of a hard drive and some don't i have thought of wanting a sorta high end dos machine one that has the use of a hard drive and can run any program that dos can run and not have issues. i remember with dos some programs would run to fast on some machines but with others they would run just fine or to slow. around in seattle its hard for me to find a good dos machine since most of the older machines that became unwanted got sent to a computer ecycler where it got stripped down and killed off for free or that the store sells the stuff for more then what its worth as a whole. maybe one of you might have something or know some stuff.

i would say a huge tower machine instead of a pizza box machine like some of the ibm ps/2 model 95 units were a tower -----Original Message----- From: Peter Corlett <abuse at cabal.org.uk> To: General Discussion: On-Topic and Off-Topic Posts <cctalk at classiccmp.org> Sent: Wed, Jul 9, 2014 3:13 am Subject: Re: best dos machine On Wed, Jul 09, 2014 at 03:18:54AM -0400, Derrick Meury wrote: There is no single such machine that will run everything as in the DOS era, software developers cut corners so it could be hit-and-miss. It basically boils down to finding a representative machine that has an acceptable spec, and the case isn't so massive and hideous that it ends in divorce. I have a 2001-vintage Celeron 633 in a cute micro-ATX pizza-box case for this purpose. However, thanks to decades of compatibility cruft, DOS will still run on state-of-the-art Haswell boxes.

On Wed, 9 Jul 2014, Sean Conner wrote: OR software. "Getting adequate performance" was used to justify doing a lot of stuff that prevented compatability. Some, such as writing directly to video memory became widely accepted, and [ALMOST] all manufacturers made allowance for it. But, there were some programs, such as "Flight Simulator" that wouldn't run on a lot of machines. (Note: by the 1990s, versions such as the "crash into the twin towers" (where do you think Osama got the idea??) had solved most of the incompatabilities. And, some copy-protection schemes were too sensitive. And, some was deliberate - my publisher (may they Rot In Pieces), in order to try to make a pitch to IBM, had me create a version of Xeno-Copy that would fail to run on anything other than the actual IBM machine. But, then they screwed up and sent THAT out to magazine editors, and it ended up being used in 1983 (PC? PC-World?) as "the acid test" of compatability. There were even some idiots who wrote programs on their "Compatible" machines, who failed to ever test them on a REAL PC! along with, "You might experience some minor discomfort"?

On Wed, 9 Jul 2014, Derrick Meury wrote: An excellent choice. MS-DOS 6.20, 6.21, 6.22 have the unique distinction of being the only versions of DOS for which the primary design goal was to improve reliability, rather than add new "features". By the way, it is NOT 6.2.2, although we tend to use the nomenclature as if it were. There is one "major version", and one "minor version", which is actually a two digit decimal number, between 00 and 99 It is actually, internally, SIX point TWENTY-TWO! If you do a function 3 of the DOS functions: Go into DEBUG.COM, type: A MOV AH,3 INT 21 INT 3 (you must hit enter (to get out of assembly mode)) G You will find that register AX contains 1606, which is 6 in AL for the major version, and 16 (sexadecimal) or twenty-two as the minor version. One thing that I don't like about the later versions, such as 6.22, is that many of the DOS programs, such as DEBUG.COM are actually .EXE files, falsely renamed as .COM (notice the first two bytes of the file are "MZ".) It's generally not a big deal, but it does make it less convenient to modify the contents of the program (using DEBUG) Each version of DOS has many differences from all other versions, but, we tend to only pay attention to the specific differences that we care about. For example, for many years, the only characteristic that really mattered to me was which disk formats it supported. To other people, different things are important. So, feel free to add comments to this list for "what really matters": CP/M What it was intended to replace QDOS "quick and Dirty OS" A place holder, to use for development while waiting for CP/M-86 to be finished SCP-DOS Seattle Computer Products Disk Operating System - "Maybe it'll be a while before CP/M-86 is ready" PC-DOS 1.00 Purchased from Seattle Computer Products, and cleaned/finished a little. First commercial release of Personal Computer Disk Operating System Note that that is a description, NOT a name. it is NOT trademarked, although IBMs lawyers will respond if somebody else tries to trademark it (see DRI) Single sided 8 sector per track (160K) floppies. MS-DOS 1.00 Same as PC-DOS 1.00, except, . . . Sold ONLY to OEMs, NO "legal" retail sales (although massive gray market). IBMBIO.COM and IBMDOS.COM are renamed IO.SYS and MSDOS.SYS. Has GWBASIC ("Gee Whiz"), instead of the ROM dependent BASICA. MODE.COM may differ from the PC-DOS version PC-DOS 1.10 Added DOUBLE sided 8 sector per track floppies (320K) MS-DOS 1.25 Same as PC-DOS 1.10 ("OEMs ONLY" "May only be sold with a new computer") Always look for WHOSE version of MS-DOS! MODE.COM and a few others may have some customization! PC-DOS 2.00 9 sector per track floppies (180K,360K). Hard disks! sub directories, "Unix style" internal DOS file functions (INT21h), in addition to the "CP/M style" file functions. Some talk of making it "unix-like", but IBM insists on '\' instead of '/' as path separator, and defaulting to '/' instead of '-' as switchchar. Addition of function 30h to check DOS version (running a version of CHKDSK MS-DOS 2.00 same as PC-DOS 2.00 PC-DOS 2.10 minor fixes, such as SLOWING the track seek of floppies to work with the [redacted] Qunetrak 142 floppie on the PCJr MS-DOS 2.11 Same as PC-DOS 2.10. Always check for WHOSE MS-DOS! MODE.COM, FORMAT.COM, and numerous other supplied programs (they are NOT "external commands"!) may differ from PC-DOS, and between one and another. Several OEMs add 360K and 720K floppy support, mostly mutually incompatible. PC-DOS 3.00 add 1.2M floppy support (for PC/AT) MS-DOS 3.00 same as PC-DOS 3.00 PC-DOS 3.10 some internal additions, such as network redirector; first version that can handle CD-ROM without major aftermarket add-ons. (CD-ROM is NOT seen as a drive on the computer, it is seen as a drive elsewhere on the network (try doing a CHKDSK of it!)) MS-DOS 3.10 same as PC-DOS 3.10 PC-DOS 3.20 added 720K 3.5" drives, and standardized the format of them MS-DOS 3.20 Same as PC-DOS 3.20. OEMs with 3.5" drives switched to IBMs format. EXCEPT HP. PC-DOS 3.30 Added "1.44M" drives, with 1.40625M capacity. MS-DOS 3.31 Same as PC-DOS 3.31, EXCEPT addition (most? OEMs) of hard drives larger than 32 MB! One of the often heavily customized variants. Always check WHOSE MS-DOS. (Zenith is one of the favorite ones) PC-DOS 4.00 Addition of >32M drive support. Norton fUtilities was not notified in advance to make appropriate changes, so malfunctioned, causing lots of complaints that there was "something wrong with DOS" because Norton fUtilities didn't work. Also, some real bugs, but the most commonly reported ones were because there were changes, that were not supported by Norton fUtilities. PC-DOS 4.01 Bug fixes. New release of Norton fUtilities to handle changes. Some bugs still remain, including function 30h returning wrong DOS version number! MS-DOS 4.01 ? MS-DOS 5.00 First version with LEGAL (not gray market) retail sales. MS-DOS 6.00 Major shopping binge! Lots of third party stuff bundled, including compression, smartdrv, etc. PC-DOS 6.10 Same as MS-DOS 6.10, but compression, etc. bundled stuff SMARTDRV was first bundled with Windoze 3.10, but, in MS-DOS 6.00, it was installed by default. Power lUsers installed compression. SMARTDRV caused data loss, which was inevitable with resequenced and delayed writes if the machine went down (such as the usual practice of hitting the power switch as soon as DOS prompt reappeared after saving the file (which hadn't actually been written yet!) That got blamed on compression! For example, Infoworld "PROVED" that compression caused data loss with a PC with compression (AND SMARTDRV). When Bill Gates told them that their test procedure was bad, and that the problem was not caused by COMPRESSION, they called that attempted intimidation. MS-DOS 6.20 Microsoft repaired the compression! They did that by having SMARTDRV NOT default to rearranged nor delayed writes, and not displaying the DOS prompt until all delayed writes were finished. Many other fixes for reliability. STAC Electronics sued Microsoft over the compression. There had been technology exchange in preparation for a licensing deal that fell through. STAC won ~100 million dollars in their suit. Microsoft won back ~30 million dollars in their counter-suit (technology exchange went both ways. Bill Gates said, "I'm having a bad day." MS-DOS 6.21 Same as 6.20 with NO compression bundled. (in response to the STAC legal vistory) MS-DOS 6.22 Same as 6.20 and 6.21 with a new non-infringing compression bundled. PC-DOS 6.30 similar to MS-DOS 6.20? MS-DOS 7.00 Hidden under Windoze95. PC-DOS 7.10 Similar to MS-DOS 7.00?

On Jul 9, 2014, at 12:07 PM, Fred Cisin wrote: Speaking of good DOSes, I'm not sure what the general consensus is here, but my favorite DOS was IBM PC-DOS 7.0 before upgrading to Windows 95 back in the day. To me, there was a lot to like about it: A lot of good Central Point tools were included with it, REXX, a nice graphical DOS shell, a disk defrag tool, Stacker (which I used quite a bit with some ST-225s back in the day), a decent editor, ability to load drivers from the command line via dynaload...