It is odd that he would say that. If one really thought SFU didn't matter, and that only key position did, then it would be better to pick Norman over Workman. Norman has high same-finger ratio but otherwise quite well-positioned keys, and has the benefit of being easier to learn (i.e. closer to Qwerty) than both Workman and Colemak.

To play devil's advocate, I would at least agree the same finger on some keys is not as bad as other keys. For example (qwerty) QZ would be worse than MN. But unfortunately, Workman's worst incidence of same-finger is PO, which is on a weaker finger (ring), so even that argument doesn't really hold up very well.

I have heard Workman aficionados also complain about Colemak's L, and certainly Workman's L is in a better position than Colemak's. But it also has a worse D and F in addition to the same-finger problems already mentioned.

It's ironic really - the motivation was to fix Colemak's overuse of the center column - about which he also says this:

Yet as it turned out, the DH mod does "just move a few things around" and the results are better than the much bigger change that is the Workman layout.

To be fair, Workman came along before Mod DH. There is always some benefit to being first, and he did do us a favour I suppose by accurately identifying the centre-column issue.  However I agree there is no longer a compelling case for Workman.

Workman isn't entirely off it's rocker. Just... mostly?

I will give it credit for forcing us to re-evaluate the difficulty map for modern keyboards (instead of the manual typewriters Dvorak was using). Prioritizing keys that cause no hand motion is a valid goal. Maintaining L.Hand shortcuts, and some Qwerty familiarity, also good. The problem is that it comes at the expense of SFU. Which as we all know, is the single greatest typing bottleneck.

There is also the other problem I've always had with Workman. Why is the Qwerty V position higher rated than the C? C is where the index finger naturally falls with a straight wrist. Unless you are using a ortholinear keyboard. (likewise, x on middle, z on ring and pinkie on shift)

But what can you do? Workman is almost mainstream, even being an option on ChromeOS. Colemak DH? Not yet. (And probably won't be until it presents a single recommended variant across all keyboards. Personally, I think the xcdvz model has the best chance for widespread adoption. Not sure why it isn't offered widened.)

Wait, XCVdZ is also on the table? No wait, nevermind. What I meant was that I feel the cycled lower left (XCdVZ) has a better chance of adoption than allowing users to pick from the multitude of ISO and ANSI specific angle options right up front, as is done on the website. Instead, take your best option for the greatest number of users, and sell the hell out of it.

As for Qwerty CV positions, I find the C moderately easier to hit. Also, just watching my fingers move -  if I keep the rest of the hand anchored to the home positions, F to C is a straight curl. F to V requires at least a little lateral movement. I can see the muscle shifting sideways, which probably means it is more stressful in the long run. This kind of motion only gets more pronounced on a laptop keyboard, which is usually more centered and closer to the typist, increasing inward arm angle and making the sideways reach more pronounced. (Of course, when floating the hand the opposite is true, as a slight inward shift of the arm for V is easier than a finger curl for C. *sigh*)

As for the cut/copy/paste block: That is an excellent point. I admittedly hadn't thought of that, mainly because XCVdZ wasn't listed as one of the options on the Mod-DH website. On the other hand, XCdVZ maintains the Qwerty V position. To quote Shai from the last time the Workman layout was brought up:

So, while a bit of a stretch, Qwerty V would at least maintain paste...? Also, Metrics. Plus 1 to Qwerty Positions!

Actually, this is where your Bag of Tricks shines, as it replaces CTRL ZXCV with Caps ZXCV. So the shift becomes a moot point.

... you know, now that I think about it, I think getting the BoT to be mainstream is far more important than Colemak-DH. No offense to the curl mod - it is an excellent tweak to an amazing layout - but in the end, it is just a layout. The BoT changes how we interact with the computer though, regardless of layout. It is almost like inventing the mouse all over again. DreymaR, you need some marketing people. (Yes, the irony of me telling you this, when I failed to properly market the Wide Mod does not escape me.)

*PS someday I'll finally post something once, and only once. No edits. ...though, not today apparently.

When you say BoT, do you mean the Extend mod? The BBoT is everything in that topic, pretty much (well, maybe not the TextBlade but anyway).

Yes, I/we do need marketing people. :-)

Errr, you're very very confusing Cevgar (as usual, lulz!). Both DH variants do the exact same thing to the LShift distance, which is to LOWER it by using the Angle mod. This is easiest on an ISO board, where the Angle mod is pretty much a given. On an ANSI board, I agree that Angle(Z) seems the easiest choice although I haven't tried it out.

Using an AngleWide mod, my hands are well separated and my wrists straight whether I type on laptop or full board. Luckily, I have good eyes so maybe I haven't experienced the distance problem you describe; I see how that could be annoying.

Well, they can. There are plenty of more ergonomic keyboard solutions including fully split ones (TIE Fighter boards are cool). These cost money though, and I prefer to work with the stuff my workplace and laptop already have.

If you want to dream very sweetly indeed, dream of a pair of DataHands.

Yes, this!  DreymaR and I disagree a little here, but I think my version of the mod is significantly better, whether on matrix or traditional boards, for exactly this reason. The point about maintaining XCV is legitimate, but in all honesty, applying the angle mod at all causes these keys to change. Sure, the same finger is used, but there is a still disruption to familiar learned motion of, say, Ctrl-C after applying the angle mod.

Therefore I don't buy that XCVDZ is as good as XCDVZ. If you are willing to apply the angle mod at all (which although is worth it, I can understand why some people wouldn't want to, especially if they are ANSI users), then you might as well go for XCDVZ over XCVDZ, especially as the former keeps V exactly in place, and you end up with a more comfortable D, which is easily the most common of those five keys.

If however you have already applied and gotten used to the angle mod on default Colemak first, then I can see the attraction of XCVDZ, as it seems like a smaller change from what you're used to. I'm sure that's why DreymaR prefers his variant. For those not already using an angle mod variant, I don't see any benefit to XCVDZ at all.

I don't think there is any prospect of getting unified agreement on DH mods here, especially since relying on the angle mod already introduces a difficulty for ANSI users. There are plenty of options of ANSI alone. But actually I'm not very concerned about that. I would rather concentrate on promoting Colemak to make that as mainstream as possible, and then have the add-on mods available for those who want them and/or have the centre-column problem.

So far as arguments over distance to L.Shift is concerned, I am increasingly of the opinion that the standard position of the shift keys is so terrible that it is better to use a different key (a thumb) for Shift. If I was an ANSI user, I think the best bet is to define Z on L.Shift, then use one of the Alt keys as Shift. Perhaps that's a change too far for most though. That option currently isn't even shown on the Mod-DH page, I'll have to add it.

Did my last post come across as a crazed rant? Sorry, it wasn't intended to. Although I remain highly sceptical about the XCVD pattern for the bottom row personally, I do think it's good to have options and I respect others' right to choose.  Also, in fairness I do think your DH mod "lite" version for the right hand does has a decent case for it - the frequency difference of H vs M is less severe than D vs V, and there is some benefit to not moving M from the Qwerty/Colemak position.

Anyway, with so many mods and possible configurations around, having "one layout to rule them all" is not feasible. So better to accept and enjoy the many options that are available. What we really need is to make it as easy as possible for people to change things to their own preference - and that means better keyboard design (programmable!), easier software configuration, or even for those usb remappers to become readily available.

I also got around to creating a illustration of the ANSI option with Z on left shift, which I think is a potential winner for ANSI, but only if people can be weaned of the (terrible) default shift keys!!

Well, although you can't switch LAlt or other modifier keys with default windows apps, it's pretty easy to do with SharpKeys. And as far as I know that only does a registry change, so I guess you could apply it with just a reg file even.

The LAlt position is not perfect, and I agree I would rather see thumbshift keyboards become mainstream. But I still think that LAlt is better than both the default Shift key, and even Caps Lock. So if you are limiting yourself to traditional keyboards, and you don't mind switching things around a little (which, if you decided to learn an alternative keyboard layout, you probably are such a person), it's an idea worthy of consideration at least.

I actually use my LAlt all the time (except mine is Extend). Admittedly my main keyboard is the Matias Ergo Pro, which has very nicely positioned and comfortable Alt keys. But even when I am using my laptop, with it's standard key positions, the LAlt is still just about satisfactory.

Alternatively, ANSI users could use Right Alt for shift, which, if using in combination with the wide mod, should be very comfortable indeed!  I don't see the point of having two Alt keys (presumably US ANSI users don't need AltGr).

If Google, or even Apple, started producing keyboards with thumbs (either extra keys between Alt and Space, or a split-spacebar) that would be brilliant! It would look sufficiently normal not to scare away "normal" people, but I could imagine such designs becoming mainstream very quickly once people realized the advantages. Yeah, just a dream though.

I'd be tempted to make it in two pieces, myself. There are good projects out there.

Using an AtMega compatible controller, TMK takes good care of the programmability.

This is close to a perfect keyboard.

I think I'd still have an extra column on the far left/right so that you can have TAB etc in the normal place.
And I'd split it down the middle so that your central row of blue keys is divided into two.

Good suggestions. I am seriously considering building this one - will be my first