Historical question - when did rubber domes eat the world?

esr

17 May 2021, 12:45

I've been doing a lot of updates to the Wikipedia article on the Model M recently. Also, Wendell Wilson of Level1Techs and I are planning to do an in-depth video on the history of the buckling spring, with keyboard comparisons. A beam-spring 3278, New Model F, and Northgate OmniKey will be featured along with a couple of Model Ms and an original Model F.

For both projects, it would be excellent if I could document when rubber domes drove the original Model M out of the mass market. I realize of course this was not a single event, rather I'm interested in trying to reconstruct a timeline of key events in the transition.

I've tried searching the web, but this is not the kind of question that is readily researched that way, so I'm looking for expertise and people with long memories.

Focusing the question: two key events would be the Lexmark spinout in 1991 and the end of Model M production at IBM in 1998. I'm more interested in the other end. Can we document when Lexmark first shipped a rubber-dome Model M? Looks like it might have been 1994 from the Wikipedia P/N table, I'd be interested to know of earlier specimens if any.

I'm looking for sources I can cite that will meet Wikipedia standards.

Reference:

https://en.wikipedia.org/wiki/Model_M_keyboard

User avatar
Muirium
µ

17 May 2021, 13:34

You're specifically interested in the IBM perspective, by the sound of it, given your documentary. But your original question "when did rubberdomes eat the world?" is broader and sets off my mental cogs all the more. Where did rubberdomes come from? What's their history? When did they get to the mainsteam, who brought them there, and how soon did they come to dominate in shipments sold? It's a dark history for us mechanical folks, yet it's absolutely a major part of keyboard history!

I dare say even rubberdome's complete domination is fading into the past now, with the way laptop-style chiclet scissorswitch keyboards have taken off, after you-know-who brought them to the desktop more than a decade ago. Sloppy, greasy, mushy domes are still with us, but they're not as ubiquitous nowadays as they were back in the 2000s.

As for your documentary, I'd advise comparing a vintage Kishsaver with one of Ellipse's remakes, but then I would, wouldn't I? :lol:

esr

17 May 2021, 14:33

@Muirium: I think we're more focused on the "buckling spring" aspect than just IBM's history - Wendell and I are both huuuuuge buckling-spring fans. But yes, the more general question about rubber domes is interesting, to me at least. We know why this slow-motion disaster happened - cost pressure - but the "how" and "when" is unclear.

We don't have a vintage Kishsaver for the lineup. Are you offering to donate one? :-)

Actually, I think we're well covered there between the XT keyboard and Ellipse's remake. What I'd really like to get my lunch hooks on is a prototype made from the Hall Effect beam-spring switches Kono has been talking up for a couple of years.

Findecanor

17 May 2021, 14:38

I think that "eat the world" is a misnomer. The history of the buckling spring is really a history of the IBM PC, and the world is a lot bigger than the IBM PC and compatible computers.

Before the IBM PC was introduced, IBM was the most influential typewriter manufacturer in the world. Their Selectric line was the gold standard. The IBM PC was supposed to replace their typewriter in offices, and the keyboards were clicky so as to have them be more familiar to typists making the transition.
The beam-spring and buckling spring switches were purposefully made to sound and feel like the Selectric (this is documented in patents and elsewhere), so much so that the clicky version of the "Model M2" was even branded the Selectric Touch Keyboard.

Clicky keyboards were associated with IBM. No other major computing platform came with clicky keyboards (even though there were third-party keyboards for some that were). Various types of rubber dome/sleeve switches and linear-feel switches have existed for at least as long as buckling springs.
Pretty much all other clicky keyboards were made to resemble IBM's, IMHO.

I believe the shift happened when a "PC" was no longer defined as being an IBM PC, a "clone" of IBM or "compatible" to an IBM PC, but more defined as a "computer running MS-DOS or MS-Windows".
That's when the clicky keyboard was no longer the norm.

I think you and Wendell should look for:
1. People who were there back in the day, who you could interview about the business and design decisions: When and why IBM/Lexmark started to make rubber domes. Start with Unicomp and work your way back there to people at Lexmark and IBM. (Unicomp was founded by ex-Lexmark people, and Lexmark originated with IBM)
2. Sales figures for IBM and other PC manufacturers, year for year. When did the IBM PC clones take over? At which point was the norm for what a "PC" was defined as compatible machines from other makers running MS-DOS/Windows, and not a machine made by IBM?
esr wrote:
17 May 2021, 14:33
What I'd really like to get my lunch hooks on is a prototype made from the Hall Effect beam-spring switches Kono has been talking up for a couple of years.
You should get in touch with the guys at Input Club who are behind it. I believe HaaTa is their lead guy.

User avatar
Muirium
µ

17 May 2021, 14:40

esr wrote:
17 May 2021, 14:33
Actually, I think we're well covered there between the XT keyboard and Ellipse's remake.
For what it's worth: I find the OG Kishsaver to be very noticeably better in feel than any XT or AT that I've tried, and I've had a few over the years. (I've also had the pleasure of multiple Kishsavers, briefly! And they are *all* that good.) It's likely the original 77 and 104 keys are similar. The feel is extra smooth and stable, and as LOUD AS ALL HELL. The wee barrel plate is held absolutely rigid by that solid metal case. Right across the keyboard, every key feels the same. True apex buckling spring!

Findecanor has an excellent point about linear keyboards being the norm, outside IBM. Quality linear keyboards really were another age. And rubberdomes wiped them out in a truly mass extinction event! Only Cherry kept them going, resulting in MX being the dominant standard the revived hobbyist mech world revolves around today.

It's not just clicky the world lost, though that word lingers on in the imagination ever since. What people really miss is "clack." :P

User avatar
fohat
Elder Messenger

17 May 2021, 15:27

I would propose that the simple direct answer to the question is:

Beginning slowly in the mid-1980s and mostly complete by the early-1990s.

When I bought my first computer, a 286-based IBM "clone" made by PCs Limited, the keyboard was a line-item peripheral that cost, as I remember, between $100 and $200.

When sellers and consumers discovered that a keyboard could be made and sold for tens of dollars rather than hundreds of dollars, the race to the bottom was on. I think that buyers were far more to blame than sellers - a keyboard became a giveaway (ie a cost) to the seller rather than a profit-generator on its own - and buyers liked that but sellers didn't.
Last edited by fohat on 17 May 2021, 15:29, edited 2 times in total.

esr

17 May 2021, 15:27

@Findecanor: I'm well aware there were lots of other things going on than the PC - I've been a programmer for more than 40 years and started on DEC minicomputers in the mid-1970s. My first computer keyboard was literally an ASR-33 back around 1972.

That said, I think "when did rubber domes eat the world?" is still a meaningful and interesting question. Because the buckling-spring-equipped IBM PC (and later, PC clones using mechanical switches like the SKCM ALPS) ate the world first, including the entire generation of minicomputers and early personal computers I cut my teeth on. I was there to watch it happen.

I think your point about when clones overtook the IBM PC in market share is sound. That wasn't the end of mass-market mechanical keyboards, not with things like the Northgate OmniKey still shipping. But it does put an earliest bound on when the rubber-dome won out. I say "earliest" because it's possible most clones shipped with mechanical switches for a few years afterwards before cost pressure moved them to rubber domes.

esr

17 May 2021, 15:46

@fohat: I'm not willing to say the victory was "complete" before IBM terminated its purchasing agreement with Lexmark in 1996.

I'm willing to write off IBM's remnant production of buckling springs 1996-1998 because I think those were going into lower-volume product lines that were statistical noise compared to the number of PCs being shipped.

User avatar
fohat
Elder Messenger

17 May 2021, 15:57

esr wrote:
17 May 2021, 15:46

compared to the number of PCs being shipped.
My point was that the peak came quickly, while the tail is still trailing out (to this day - with a small but significant ramp up starting in the early-mid-2000s when gaming got serious).

Hence the words "beginning" and "mostly" in my comments. My first rig cost $3K without a printer or a mouse, but by 1990 complete systems were selling under $1K. That was not possible with a costly keyboard.

Findecanor

17 May 2021, 16:11

esr wrote:
17 May 2021, 15:27
@Findecanor: I'm well aware there were lots of other things going on than the PC - I've been a programmer for more than 40 years and started on DEC minicomputers in the mid-1970s.
I meant no disrespect. Quite the contrary. I've known who you are for ~20 years, and listened to Wendell every week since the TekSyndicate days. ;)
I was just trying to help you guys by giving you a better answer, ... and with what I thought would be interesting points and facts. I'm looking forward to your documentary.

esr

17 May 2021, 18:52

This is pretty informative:

https://www.sfgate.com/business/article ... 047969.php

It says Compaq passed IBM in PC sales volume in 1994. But that was already three years after the Lexmark divestiture, so cost pressure was probably making buckling-spring keyboards uncompetitive *sooner* than the clone-makers got the upper hand.

User avatar
Tritian

17 May 2021, 19:31

I'd consider also comparing Unicomp's new Mini M since it is continuing the legacy

esr

17 May 2021, 19:50

Hm. I think I spoke too soon. Maybe I actually had all the information I needed...

In 1991, IBM made a 5-year agreement with Lexmark (which IBM had just spun out) to buy keyboards from them. At the time, Lexmark was only producing buckling-spring keyboards. So IBM's product strategists must have believed in 1991 that they could still make a net profit on buckling springs until at least 1996 - cost pressure can't have been terminal in the future they foresaw at that time.

I think that pretty much rules out an eat-the-world date sooner than 1991. I mean, unless we think IBM's people were incapable of analyzing their own production costs and margins!

On the other hand, the agreement was not renewed in 1996. Which suggests pretty strongly that their incentives had flipped by that time.

So 1991-1996 is looking like the window. And Compaq passed IBM in sales in '94. And 1994 is the first year in which I know that Lexmark shipped rubber-dome keyboards itself.

At this point, if I had to pick a year, 1994 is looking pretty plausible.

Thoughts? Arguments?

User avatar
fohat
Elder Messenger

17 May 2021, 20:15

Here is an article from early 1990 that is long and dense but interesting.
https://money.cnn.com/magazines/fortune ... /index.htm

Choice quotes:

"The economics of our business are frightening. We can deliver a lot more power for a lot less money each year, but the industry isn't growing as fast as it once did."

"Apple Computer CEO John Sculley tersely concludes, ''This industry may have hyped itself into a recession.'' "

User avatar
sharktastica

17 May 2021, 21:20

esr wrote:
17 May 2021, 12:45
Focusing the question: two key events would be the Lexmark spinout in 1991 and the end of Model M production at IBM in 1998. I'm more interested in the other end. Can we document when Lexmark first shipped a rubber-dome Model M? Looks like it might have been 1994 from the Wikipedia P/N table, I'd be interested to know of earlier specimens if any.
The first "Quiet Touch" Model Ms were built by/shipped in October 1993 as per this IBM announcement for IBM AS/400. It mentions P/N 71G4644, which is the standard US ANSI rubber dome Model M. ClickyKeyboards seems to have a few 71G4644 store pages for past sold 1993 71G4644s.

Also on another note, it seems IBM's in-house production ended in at least 1999. Here's a birth certificate of a 1999 Greenock-made M122. Apologies it's in poor condition, but the date is at least clear. Also, not my photos (from a volatile eBay listing from a few weeks/months ago):
Birth cert.jpg
Birth cert.jpg (191.65 KiB) Viewed 1446 times
Front.jpg
Front.jpg (201.57 KiB) Viewed 1437 times
Last edited by sharktastica on 18 May 2021, 01:08, edited 1 time in total.

micmil

17 May 2021, 21:50

There seems to be a lot of confusion regarding the adoption of rubber domes. It was not solely a function of cost. Rubber domes have the added benefit of being super quiet. There's a reason all of the early ones are named "Quiet Touch" or "Silency Silentnonoise" or something to that end. While the cost may have been lower the dB rating was much more attractive in an office setting. I can sit here and listen to my buckling springs and Box Jades all day long but even over the phone I get comments about how loud my keyboards are if I'm using one during work.

The lower noise drove adoption by businesses which drove prices down even further but the REAL doldrums of rubberdomery didn't start until late in the 90's when eMachines, Dell, HP, and others started the race to the pricing bottom. Keyboards were the easiest way to cut a few bucks without anyone caring much. Either someone knew their stuff and had their own keyboards or they didn't and ended up not caring how the keyboard felt, as long as it works. THAT is when the slow erosion turned into a landslide because at that point almost every new computer sold had a rubber dome board included.

Rubber domes existed before, and had a large market share, but it was that late-90's pricing war that really destroyed all market share that mechanical keyboards had prior.

User avatar
raoulduke-esq

17 May 2021, 21:53

micmil wrote:
17 May 2021, 21:50
I can sit here and listen to my buckling springs and Box Jades all day long but even over the phone I get comments about how loud my keyboards are if I'm using one during work.
Coworker: What was that keyboard you were using on the call last week?
Me: The one that sounds like a spray paint can or the one that sounds like a teletype machine?
Coworker: The latter. It was godlike. I have to have one
Me: 3278 - provides link to wiki
Coworker: - proceeds to read wiki and look up pricing
Coworker: - ummmmm... never mind

User avatar
fohat
Elder Messenger

17 May 2021, 23:35

micmil wrote:
17 May 2021, 21:50

it was that late-90's pricing war that really destroyed all market share that mechanical keyboards had prior.
Probably true, and I would add that serious gaming probably powered most of the renaissance a few years later.

User avatar
hellothere

17 May 2021, 23:41

micmil wrote:
17 May 2021, 21:50
Rubber domes existed before, and had a large market share, but it was that late-90's pricing war that really destroyed all market share that mechanical keyboards had prior.
I think that's about right. In a very general sense, if I see a keyboard with a Windows (95) key on it, I assume it's a rubber dome/membrane unless I have information to the contrary. There were still a bunch of Alps and other switches out there, though. Easy example is the Dell AT101 bigfoot. Black Alps.

Personally, I am interested in when the first rubber dome/membrane non-mechanical came out and what it was. I'm thinking it was probably in the mid 1980s.

esr

18 May 2021, 02:16

"it was that late-90's pricing war that really destroyed all market share that mechanical keyboards had prior."

I could easily believe this if not for the fact that IBM failed to renew its agreement with Lexmark in 1996, and Lexmark then bailed out of its keyboard business. Impending doom must already have been visible in 1996 or earlier, which is a couple years too early for the launch of eMachines in 1998.

Mind you, I don't doubt that the late 90s price wars accelerated the decline. But the actions of these major players signal that the buckling spring had already gotten too expensive for the mass market before the real bloodletting began.

As for low noise being a draw, color me skeptical. I was there while all this was going on - I remember rather vividly what a shock the low-ball prices on the first eMachines were - and I don't remember keyboard noise being an issue for anybody. I was on the board of a Silly Valley startup between 1998 and 2001 and I did a lot of walking around the offices to get a feel for what was going on at the company. People used to pull stunts like binding their keypress event to the TCHOONK sound of an ASR-33 and nobody turned a hair.

micmil

18 May 2021, 04:02

esr wrote:
18 May 2021, 02:16
"it was that late-90's pricing war that really destroyed all market share that mechanical keyboards had prior."

I could easily believe this if not for the fact that IBM failed to renew its agreement with Lexmark in 1996, and Lexmark then bailed out of its keyboard business. Impending doom must already have been visible in 1996 or earlier, which is a couple years too early for the launch of eMachines in 1998.

Mind you, I don't doubt that the late 90s price wars accelerated the decline. But the actions of these major players signal that the buckling spring had already gotten too expensive for the mass market before the real bloodletting began.

As for low noise being a draw, color me skeptical. I was there while all this was going on - I remember rather vividly what a shock the low-ball prices on the first eMachines were - and I don't remember keyboard noise being an issue for anybody. I was on the board of a Silly Valley startup between 1998 and 2001 and I did a lot of walking around the offices to get a feel for what was going on at the company. People used to pull stunts like binding their keypress event to the TCHOONK sound of an ASR-33 and nobody turned a hair.
For IBM and others to start abandoning mechanical keyboards there would have necessarily been demand for change from customers. The appearance of so many "quiet" keyboards around the same time, and even on the same base as various mechanical keyboards, illustrates that the demand was there for something that wasn't as loud as most conversations. A Silicon Valley startup in the late 90's hardly represents, even anecdotally, anything resembling the standard office at the time. Think law firm. Think accountants. Think... the kind of companies that bought IBM products because nobody ever got fired for buying IBM. Think about dozens of Model M's clacking away against the noise of each other. Or don't, because that sounds horrific. But then think that some guy named Todd in middle management has to work next to that noise. When the only computer in the place was shared by five people that knew what they were doing and it was isolated in its own room the noise wasn't a problem. In the early 90's PC's started appearing on more desks and just a handful of years later they were on every desk. That would get a bit tiresome, and I can sit here going fdsajkl;fdsajklf;dsajklf;dsajklf;dsajkl;fdsjkl;fdsa on buckling springs all day.

But that's still just a fraction of the PC market and can be seen as a bellwether, not the actual movement. The office market may have demanded something quieter but look at what you had at the time. Much like mechanical keyboards ten to fifteen years prior, rubber domes were undergoing a bit of a patent war (this was the real impetus for so many of the differing switch designs back then, which nobody seems to really talk about). Look at all the different types. Dome with slider, dome without slider, capacitive domes, dome over membrane, dome over PCB, there were about as many variations on domes as there were mechanicals. They were quieter and the best ones could type just as well as anything else. And then something happened.

PC's became a commodity.

While the business market was trending towards rubber domes, there was no reason for the home users to abandon mechanical keyboards and enthusiasts largely didn't. But your grandma had no idea what a buckling spring was and didn't care. As long as it was easier to type on than that damn Remington it didn't matter. Your uncle Steve didn't care about keyfeel. He cared that for $500 and an AOL subscription he could finally realize his dream of creating nakedchicks DOT com. Your parents just wanted you to do your damn homework and somehow this was supposed to help. The keyboard didn't matter and the manufacturers knew this.

So they provided the absolute cheapest thing they could get away with. Dell, HP, eMachines, and others destroyed the mechanical keyboard market by completely cloaking the fact that they even existed to what was now the majority of computer users. Grandma didn't know about Alps. Steve had no clue about Cherry MX. Your mom wasn't aware Acer switches were a thing.

The commodity companies absolutely cloaked the fact that mechanical keyboards even existed and as a result nobody even knew to ask for them anymore. Prior to this mechanical keyboards may have been falling out of favor but they would have remained as a niche product rather than become the critically endangered oddball products they were from roughly 1998 until the Black Widow hit the market.

And like it or not, the Black Widow is what you have to "thank" for the modern resurgence of mechanical keyboards but that's a different post for a different topic.

esr

18 May 2021, 04:51

"For IBM and others to start abandoning mechanical keyboards there would have necessarily been demand for change from customers."

Assumes the conclusion. I don't remember users caring about noise at the time; I think the change they wanted was lower prices, and that created cost pressure to abandon buckling springs that was quite sufficient. The "it's quieter!" marketing has always seemed to me to be a head fake intended to provide cover for the cost-cutting change in case anyone noticed that the result was flimsy crap, not a reflection of demand.

Indeed, there's strong evidence of demand for noise - a couple of pre-M IBM keyboards had solenoid thumpers in them so that typing would be reassuringly loud, like a Selectric. The buckling spring was apparently intended to sound like a Selectric without needing a thumper.

You speak of the noise of dozens of Ms, but at the law firms and accountancies you reference enclosed offices were still normal, making keyboard noise a non-issue. (My wife is an attorney who was a junior partner at a white-shoe law firm in Philly at the time, so I speak from direct knowledge.)

I've heard this argument about demand for quiet a few times before, and I've always thought it was back-projection of today's priorities onto conditions that didn't really exist at the time of the transition. I never heard of anybody hunting for a typewriter quieter than a Selectric to squelch the clatter from the pre-computerized secretaries.

micmil

18 May 2021, 05:19

esr wrote:
18 May 2021, 04:51
I never heard of anybody hunting for a typewriter quieter than a Selectric to squelch the clatter from the pre-computerized secretaries.
https://www.youtube.com/watch?v=gJ1Ij38bCO8

https://www.youtube.com/watch?v=L51AQ-VSPgA

https://www.youtube.com/watch?v=JzG0If8J4gU

And just because you mentioned the Selectric...

https://www.youtube.com/watch?v=nu8JVHH2UYA

If you didn't hear of anyone looking for a quieter typewriter it's probably because you couldn't hear them yelling at you over the clatter of typewriters. Office noise has been an issue for as long as offices have existed.

esr

18 May 2021, 11:27

And yet, the Selectric massively outsold all these niche "quiet" typewriters. IBM's product designers thought the noise was an important feature to add solenoid thumpers to the 3278. The buckling-spring patents speak of intent to generate Selectric-like noise. And there were other period keyboards with an artificial click generated through a speaker because the switches were too quiet.

That's not the far past, either. I just discovered that the Microsoft Surface. marketed as cutting-edge dream tech, has an option to generate an artificial keyclick sound to this day.

User avatar
Muirium
µ

18 May 2021, 11:45

Bosses liked their secretarial staff to sound busy, and bosses bought the equipment. No one ever got fired for buying IBM, as they used to say.

The typists themselves might have had very different opinions!

User avatar
Bjerrk

18 May 2021, 12:20

micmil wrote:
18 May 2021, 04:02
Think... the kind of companies that bought IBM products because nobody ever got fired for buying IBM.
Muirium wrote:
18 May 2021, 11:45
No one ever got fired for buying IBM, as they used to say.
They sure did. Long, long ago.

micmil

19 May 2021, 00:28

esr wrote:
18 May 2021, 11:27
And yet, the Selectric massively outsold all these niche "quiet" typewriters. IBM's product designers thought the noise was an important feature to add solenoid thumpers to the 3278. The buckling-spring patents speak of intent to generate Selectric-like noise. And there were other period keyboards with an artificial click generated through a speaker because the switches were too quiet.

That's not the far past, either. I just discovered that the Microsoft Surface. marketed as cutting-edge dream tech, has an option to generate an artificial keyclick sound to this day.
I could post dozens, possibly hundreds, of keyboards, typewriters, and other bits of office equipment marketed as quiet, silent, quiesilent, silequiet, or other variations of not making much noise...

...and you'll ignore all of those and point to one or two early models of keyboard that shipped with noisemaking solenoids, most of which weren't even produced within in my 40 year lifetime, all of which were discontinued to hell and back well before the 90's, therefore only reinforcing my statement that nobody wanted to pay for noisy-ass keyboards otherwise the solenoids would have remained in production as a requested feature.

Which, again, goes back to my point that nothing you're talking about contrary to what I'm saying. You just have your timeline skewed.

esr

19 May 2021, 13:11

@micmil: With respect, if you're only 40, you weren't there. I was; by the time the rubber-dome transition cranked up around 1990 I'd already been a processional programmer and heavy keyboard user for over a decade, working in offices that bought the kind of equipment we're discussing.

You're asking me to believe a story about noise reduction being a major element of consumer demand during the 1990s that contradicts what I experienced. I was there, and I'm telling you that you have overinterpreted and misinterpreted the evidence. You've bought into a a myth that is more reflective of 21st-century concerns than what people were worrying about at the time.

And it's not a useful argument to be having. I opened this thread to get a better handle on the when and how of the rubber-dome transition, not to be in a dispute about why it happened. The "why" was never unclear to me; I had a front-row seat for the period's price wars. Hell, as a board member of VA Linux Systems during the dot.com boom I was an actual participant in them. I had skin in the game - I know why we didn't OEM a decent mechanical keyboard for our workstations in 1998 and noise wasn't it.

So can we stop doing this, please? I have learned a couple of useful things from this thread. I'd like to learn more. I'm mainly interested in facts, not interpretations. Especially facts that I had not previously realized were relevant.

Let's refocus on the question: what was the timeframe, and what were were the crucial events, in establishing the market dominance of rubber-dome keyboards over early mechanicals like IBM's buckling-spring keyboards and the Northgate OmniKey?

User avatar
Muirium
µ

19 May 2021, 13:31

The crucial events? If you're after facts, over speculation, you're going to need some sales data. Shipments will tell the true story.

esr

19 May 2021, 18:30

Muirium wrote:
19 May 2021, 13:31
The crucial events? If you're after facts, over speculation, you're going to need some sales data. Shipments will tell the true story.
Those would be the best kinds of facts to have, all right. No idea where to find them, though. The Web didn't start to become a significant archive of those kinds of facts until after 1996.

Post Reply

Return to “Keyboards”