MEI switches

[Daniel Beardsmore]

Anyone know if these are U mount?

http://deskthority.net/wiki/MEI_switch

[Icarium]

I had to look up "U Mount". >_<
Now it seems that they obviously are. What am I missing?

[Daniel Beardsmore]

Yeah, several people have asked where "U mount" originated. It's possible Cherry just copied it. Back in these days, both Cherry and Alps were toying with various mounts, including cross and tee switches each. (The Cherry tee switches may take M7 keycaps; don't know.) There are other switches with crosses — the Futaba or Futaba clones on the BBC Micro look U mount, but the dimensions are slightly larger and Cherry caps won't fit.

Cherry settled on a cross with 1 mm by 1.5 mm arms. What cannot be proven from that one photo, is whether MEI are using a cross of those exact dimensions, or something annoyingly different.

It gets worse. The [wiki]Cherry M8[/wiki] switch (which predates MX) has a cross with those dimensions, but M8 and MX caps are not interchangeable due to differing stem lengths, so, perhaps to assist companies with M8 tooling, there are MX switches with shorter M8 stems! This is in addition to the MX Orange (with a non-standard cross) and the MX switches with Z mount (Alps) stems ...... Without a vintage catalogue, I cannot prove that the Orange and Z mount variants are actually MX (i.e. MX* product codes), because if so, then the U mount categorisation at the wiki needs to be moved onto the individual switches that actually are U mount, not MX as a whole. (If not, then MX Orange isn't MX, just Orange :-)

The more work you do on the wiki, the more you realise is missing from it. So much knowledge scattered about in forum topics that needs to be collated. (If people each just covered one switch they're familiar with, we'd get there.)

In terms of MEI, it may be slightly different to 1 by 1.5 mm cross arms, or it may need special keycaps like the M8 did. I've e-mailed MEI for specs, but I don't know that it's even the same MEI. Probably not. Can't find another suitable MEI.

[Daniel Beardsmore]

Confirmed over at Geekhack previously. Also: http://www.smcelectronics.com/swkey.htm

Next people for me to harrass for specs ;-)

[webwit]

It's a fit.

mei2.jpg (181.36 KiB) Viewed 6790 times

mei1.jpg (219.93 KiB) Viewed 6790 times

[Daniel Beardsmore]

Interesting. Keycap spacing proven to be 19 mm. Someone at Geekhack previously claimed 18 mm, but the Datalux spec says 19 mm (19.05 specifically), which you have proved correct. Plus, a top down shot for better tracing for the switch diagram.

I forgot to ask Datalux specifically whether the current models are Cherry ML, but I imagine they are. Hopefully that will make them more willing to divulge who MEI are if I can't actually buy MEI switches now.

BTW, are you actually sure they're mechanical, and not some weird membrane contraption inside? :)

Maybe I'll draw a MEI later. Either that or SMK cross slider.

[webwit]

I never opened one up, but considering the feel, it could easily be something really lousy.

[Daniel Beardsmore]

OK, that's one for you. A keyboard spotter can have the fun of completing my model table, although I know it will still be my job …

http://deskthority.net/wiki/Datalux_SpaceSaver

[Ascaii]

Haata has pics of the opened switch on his page.

[Daniel Beardsmore]

Hm ... You mean that page with fifteen million pictures on it? Now that I've got a new PC, it might actually load. I tried it on my previous PC, and finally managed to close the tab before the whole system crashed and burned.

[Ascaii]

Yes, that page xD

[Daniel Beardsmore]

If it's http://kiibohd.com, that's screwed up beyond belief. Half the commands don't seem to do anything, and "keyboards" just vomits a ton of stuff off the top of the window where you can't read it (there's no scrolling). It also has HP linked to Fujitsu for some reason.

I guess that's an improvement over striving to crash my PC, but seriously, sometimes (i.e. always) it's better to be accessible than a dog-panting eagerness to impress to the point of absurdity and uselessness.

[Daniel Beardsmore]

Haata has pics of the opened switch on his page.

Sure? I managed to find the URL that dumps all the photos out, and I found the Datalux pics, but no opened switch. (There are other MEI switches that may or may not be the same MEI, but they're all 70s stuff.)

What sort of prize do we need for the 2013 Deskthority wiki award, to get people to commit their knowledge somewhere accessible and organised?

[webwit]

Uncharted territory, Daniel. I don't think there are a lot of MEI owners / deskthority members / currently into wiki. You're better off if you see that as an advantage rather than a disadvantage. You're on your way to become deskthority's sandy55 who will be referenced forever for your research and documentation. Would it be fun if it was already covered this way? If you subtract from deskthority members the number of members who are currently into wiki and the number of members who have the hardware to say something about the current point of interest, I think you have to live with the fact that most writing will be serial, not parallel with many authors. Serial like daedalus, sixty, 002 and others contributed greatly to the wiki before and others will contribute greatly in the future. But perhaps not many on the same subject at once.

[Daniel Beardsmore]

I don't expect anyone to know anything about MEI switches. I just lament how much information people have committed to the wrong places (including forum topics), and aren't going to commit it to anywhere useful now. Take Sandy — he seems to have ripped down a load of pages (those on Alps switches), but at least the Internet Archive still has at least one of them. Sadly that's still no use as Japanese and Chinese is not presently machine translatable — the results just come out a stoner garbage. Hilarious, but useless. Asians prefer to communicate in their native languages, while Europeans prefer English.

So much knowledge exists. It looks like the only solution is to spend all night and day reading every forum topic like it were freshly uncovered Egyptian carvings in some pyramid somewhere, reassembling lost knowledge. Which is just absurd, but is this what the English-speaking keyboard community has been reduced to? Compare with trainspotters, who were actually keeping Wikipedia (not even a private wiki) up to date with every BR Class 43 locomotive entering and leaving Brush Traction for engine refit (lousy quiet engines, too, grr; locos should be seen and heard!). Not just "they're being refitted", but actual locomotive numbers and the dates entering and leaving.

Of course, you do have to entrust the guardians of your data to actually back it up ;-) That said, Geekhack's wiki was a real mess and a nightmare to follow. Maybe mocking Geekhack is unfair, though — are we any better here? (Not backups, just being a driftwood community.)

That is one rationale, though, for a separate keyboard spec database — it would be editable without the need for learning markup languages. Just type number into box and press Submit. Someone else can cross-post later. Certainly better than faffing around with Google Docs spreadsheets, as at least a custom website could generate CSV dumps and BBCode and MediaWiki markup for rapid data export. While I agree in principle with SemanticWiki, I think you'd lose what little editing you have right now, or you'd have to chase everyone round rewriting in terms of the new code, so basically I guarantee it won't happen. Should, but won't.

[webwit]

You're too much in a hurry. It's a hobby, it's for fun. It's not a job and it won't change the world. And you can do it exactly in the way you think is right at whatever tempo. This is all the point. Who cares if it takes 5 years or 1 year to get the wiki complete? You should do it for your enjoyment and not some higher purpose. PS. Since deskthority wiki is an open wiki it will live on forever even if deskthority itself would go tits up.

[Daniel Beardsmore]

Don't forget that a go-to source of information is only as useful as it is complete. If you typically can't find what you're looking for, you won't bother with it. It's important to get it to the state that it's actually a viable source of information. Right now it's terribly sketchy.

It's also hard for anyone to be motivated about something when they realise that no-one else is bothering. It's too easy to give up on the basis that no-one's going to care whether you bother or not. Yes you can win some lame wingnut key — from what I can see, that's not made any difference — no-one's eager to get next year's prize. I can't say I blame them. Though in my case, the only prize would be to see real progress in getting this thing finished. There's a vast, yet distinctly finite amount of knowledge, considering that most switches are now gone forever.

It's all catch-22s. It's like buying a book and finding three quarters of the pages are empty. And maybe it really is less fun to do everything alone. However, here we have a community where you still work alone.

[Soarer]

le mieux est l'ennemi du bien.

[Daniel Beardsmore]

Really? The French for "enemy" is "ennemi"?? Dear me.

I wouldn't call the wiki "good". It will be "good" once people can actually take a keyboard they found, work out what it is, and learn all about it. Even with Cherry, it's only good for current switches. All the knowledge is in Asia. No-one in the English-speaking world would produce such awesome disassembly photos of Alps and clones, although funnily enough I'm relying heavily on Ripster's imgur site for pictures of switches. Maybe the state of the keyboard community is what drove him mad. Wouldn't surprise me honestly.

I see 002's posted a Sony BVE keyboard to the wiki, though. I've played with one briefly — fantastic switches, couldn't get a keycap off easily enough though to figure out what they were. Topre apparently. That explains it. Also confirms it wasn't worth buying — can't harvest the switches. (There weren't enough keycaps either for a normal keyboard. I was considering buying it just to sell on to someone who'd actually appreciate the parts, assuming it was some really high grade discrete switch module.)

[Soarer]

Or is the English word for 'ennemi', 'enemy? And before that we just had 'foe'?

Sigh. Commonly it's taken to mean that an insistence on perfection prevents the good from coming into being.

Perhaps some strange convolution of the 80:20 rule tells us that 80% of people would find what they where looking for, even if the wiki was only 20% 'complete'. That would be a 'good', and realistic, goal.

I don't know what will motivate people to contribute, but having to meet overly high standards will not help.

[Daniel Beardsmore]

There's no shortage of info on the Web if people want to know about Cherry MX or buckling spring, or even Topre. You could just ask Cherry. Or Unicomp. Or Topre. The stuff is still being made!

Never did care for tin-canned sayings. They're usually wrong anyway.

[Soarer]

Well I could write a lot of words instead, to say essentially the same thing. But why should I bother, just to have it dismissed with a single line? If you don't get what I'm saying, it's not my problem.

[webwit]

There's only one truth which is visitor stats! Wiki will reach 1 million views in January.

[Daniel Beardsmore]

Well I could write a lot of words instead, to say essentially the same thing. But why should I bother, just to have it dismissed with a single line? If you don't get what I'm saying, it's not my problem.

Because you're wrong. The wiki won't be perfect, but it should be good. (Hell, if you want perfect, you'd have to rewrite MediaWiki, because it's awful. Is that going to happen? Of course not.) Choosing an irrelevant canned expression doesn't serve any purpose other than to perpetuate them — it's like a meme from before memes were "cool".

Being mostly empty for years doesn't qualify as "good" — it just becomes one more dead lead in a world-wide web of dead leads. One more browser tab that gets closed because you still can't find what you're looking for despite Google suggesting you might. I used to find with my own website (which sucks, I grant you) that despite all the hits, the entry and exit pages were all the same — nobody was reading around. Hits aren't relevant — what's relevant for a wiki, considering how they work — is that people move from page to page and leave a different way that they came in. If you don't explore, you won't have read very much, because each page is very focused, and interlinked with separate pages with the additional information you just have to read. (Which is why I'm adding inter-page links across the wiki, so that people actually realise that they have access to this additional information. This is critical for wikis as they don't have a natural structure like traditional sites: these links are your navigation.)

I can't say whether this is true of the visitors statistics here, but raw numbers aren't any more useful than canned expressions.

[Soarer]

... Being mostly empty for years doesn't qualify as "good" ...

That is exactly opposite to what I was saying. So even though I tried to explain my use of the canned phrase, you still don't understand what I was trying to convey. Attacking it simply because it's a canned phrase is not a valid argument against it.

You want collaborators, yet your attitude is destined to drive them away. You do great work on the wiki, but then harshly criticise anyone who fails to use the 'perfect' image filename, or clutters the log with too many edits, so is it really surprising that "no-one else is bothering" to join you in your efforts?

I welcome any contribution of content, even if it is flawed. I also welcome any contribution of editing/rearranging/correcting/tidying, even if the result is still flawed. I'd reserve harsh criticism for vandalism or deliberate misinformation. I think many more would be encouraged to participate if they thought that 30 minutes now and then would make a useful difference (which it would), than if it seemed that doing anything useful would take all day (which is how putting a multitude of hurdles in front of people makes it seem).

But what would I know, I'm simply "wrong".

[Daniel Beardsmore]

You're still mixing up "perfect" and "good". And it was webwit who suggested that being empty for years is fine.

Think about Wikipedia. When I first encountered it, it was in a good state — most things I wanted to look up, were there. I didn't see what it was like before most of the articles were written, but it wouldn't be all that useful. Wikipedia is so great because it has so much information. It's far from perfect, but it's a great go-to source because of the breadth of coverage.

The Deskthority wiki is so incomplete, that you can't treat it in this way yet. We need to get it as fast as possible to the stage where it has decent all-round coverage.

Also, don't forget that the clock is ticking on anything vintage.

[webwit]

We're hardly empty. Two years ago there was nothing. Now we have a substantial amount of information. In a year it will be even better. I don't see the problem.

[Daniel Beardsmore]

If you like Cherry and Topre, sure. (There's a huge table of Topre Realforce models, and even that's wrong — but I couldn't make enough sense out of it to correct the entries for European keyboards. The information on other pages isn't sufficient.)

If you look up any vintage switch, virtually nothing, assuming you can figure out whether it's in the wiki or not, or where (great for all those unbranded switches, or ones that are branded on bottom, which is why people keep mixing up SMK and Futaba). I've written something on NMBs just by looking at Ripster's photos (I've mostly figured out what they do, and no other photos are even close to good enough). Futaba clicky? Some pictures, but you need Ripster's to get a better idea, and I still don't know what a "square" Futaba is. SMK? Nothing. MEI? A really scrunched up photo. Futaba linears? Nothing. Non-Z-mount Alps? Nothing. Real vintage Cherry? Almost nothing. (If the marketplace didn't suck, I could actually buy some of these boards. The ones that weren't destroyed for switch sample packs, that is.)

All switches that people do know about. And when I'm digging up photos to trace for illustration, I keep turning up all sorts of switches and keyboards that aren't documented here.

I've tried to make the wiki less haphazard, but it's still confusing and messed up. Maybe by these five magical years, it will be possible to put text in diagrams.

It's improved within the last few months because I tried to kick life into it.

[Soarer]

You're still mixing up "perfect" and "good".

No, I'm not mixing them up. It's just that your idea of 'good' is closer to my idea of 'perfect' than my idea of 'good'!

Of course time is a factor - that's where the 80:20 rule can guide where it's worth spending time. Naturally such a simplistic rule is flawed, but even so, it's reasonable to say that for many articles the first 20% of time spent on them will produce something that's sufficient for roughly 80% of people who want information on the article's subject. At that point it's "good", and further effort has diminishing returns. Sure, at that stage it's evidently flawed to anyone knowledgable on the subject (i.e. it's original author) so it's hard to leave it at that stage, but overall it's better to spend time elsewhere than to raise an article's sufficiency by even 5%. The same applies to the task of getting everything properly filed and linked etc., or to the 'completeness' of any given area.

So I just fear that your idea of "good" is 90% sufficiency or more. More importantly, that holding other contributors to that standard will discourage them - losing both time and effort.

PS. Just to clear an earlier point, I didn't ever mean to imply that the wiki could be considered "good" with only information on Cherry, Topre etc. Even though that might suffice for 80% of queries from the wider web, it almost certainly would not for the queries from DT members.

[Daniel Beardsmore]

No, it's just that the world's standards have gone through the floor.

I never said articles needed to be complete coverage of the subject in question; that would be silly. I just said to submit each edit so that the article is in a self-consistent state when done, and proofread as best as you can, with the spelling checked. Just basic presentational skills.

Basically, make articles suitable for reading, and the internals suitable for other editors to make sense out of.

Seriously this is basic stuff. But then, if people can't even run a spelling check on their CV of all things, there's no hope left.