Where does the name Minimak come from?

It’s a play on the word “minimax” and the “ak” ending of Dvorak (named after the researcher who designed it) as well as Colemak (another nod to Dvorak). Minimax is a contraction for “minimum change, maximum benefit”. It’s unrelated to the game theory algorithm of the same name.

How fast are you at Minimak?

See the homepage. I try to post my QWERTY and Minimak speeds daily.

How long does it take to get up to speed?

I’m not fully up to speed yet, so I can’t say for sure. You can track my progress on the homepage, which shows the count of days I’ve been using Minimak. After 9 days with the 12-key layout, I was in the 40-60 wpm range with both Minimak and QWERTY. That’s way better than I experienced with other layouts. It’s just nowhere near as frustrating.

However I’m not really a representative case since I’ve been typing with precursor designs for some time.

It took me about six months to get up to 80 wpm with the QGMLWB layout from carPalx. I’d estimate Minimak to cut that time to a third.

Which Minimak do you use?

I’m using the 12-key layout.

Which Minimak do you recommend?

I think the 8-key layout is the best compromise for most people. I like the 12-key layout more for aesthetic reasons (such as moving P out of an uncomfortable position) than for its actual metrics. I think it’s further from QWERTY than most will find comfortable, and the 8-key layout solves the obvious problems with QWERTY.

What are the obvious problems with QWERTY?

E and T need to be in the home position and should be on the middle fingers. J and K have no business in the home position. The rest is getting the highest frequency letters within the reach of the middle and ring fingers and avoiding same finger repetition.

How many people are using Minimak?

There have been about 70 downloads by November 2012, but I’d guess only a small fraction of that are active users. I’m looking forward to more people trying it out. It really works.

How long has Minimak been in existence?

The layout was finalized and launched on October 21, 2012.

I want to learn the 12-key layout as fast as possible. Should I just learn it directly or use the 4-key steps?

I can’t say with total confidence, but I believe the 4-key route is quicker. I need more real world experiences to say for sure.

If you’re absolutely gung-ho, I’d say that the 6-key layout is a good compromise that gets you there in two steps. It’s explicitly designed for learning the 12-key layout quickly. Give it a try.

I already use Dvorak/[layout X]. Should I learn Minimak?

If you’re unsatisfied the effect Dvorak has had on your QWERTY skills, then I’d consider Minimak. It’s not going to be as easy to learn as it is for QWERTY users, though.

Otherwise I’d say no. As my metrics show, Minimak’s value is not that it’s any better than Dvorak, it’s that it’s significantly better than QWERTY while helping you keep your QWERTY skills.

Learning a new layout is hard. If I’m going to go to all that trouble, shouldn’t I learn the best possible layout?

Perhaps. There’s two things wrong with that point of view however.

First, not all layouts are made equal. Some are harder to learn than others, and when you look at what you’re getting for it, some are harder for no good reason. It’s a myth to say you have to learn something hard to significantly improve your typing experience. Minimak was created out of that realization.

Second, this point of view entirely forgets that every other computer you will ever use will be QWERTY. You will be reduced to a crawl whenever you have to use them unless you meticuously maintain those skills. My job requires that I be able to use any computer I’m put in front of, and alternative layouts took that skill away from me. It’s just not practical to try to maintain QWERTY skills while using an unrelated layout on a daily basis.

Is Minimak ergonomic?

I make no claims about ergonomics. If the fact that your fingers move less with Minimak than with QWERTY helps, then that’s a side-benefit and not the intent of the design.

Can Minimak help me type faster?

Same answer as above. That may be a side-benefit but it’s not the intent of the design.

How do you come up with your comparison numbers to Dvorak?

I use the keyboard layout analyzer to compare each layout with Dvorak. I’m comfortable with the statistics it generates from the Alice in Wonderland input that it provides, so I use that so others can reproduce the analysis more easily. The download has importable maps for the layout analyzer, which makes loading them easier.

Once the layout analyzer does its thing, I look at the difference in the absolute numbers for QWERTY versus Minimak and QWERTY versus Dvorak. I do this by computing the ratios of the mathematical differences (minus) from QWERTY for the two competing layouts. I look at the reductions in finger distance and same-finger repetition, then use a weighted average of 2/3 finger distance and 1/3 finger repetition, which feels right to me.

Are your metrics subjective?

The numbers themselves are not subjective, but the choice of what metrics I think are important is. I think they provide a strong assessment of goodness of feel, which is the goal of the design. The greatest part of it is reduction in finger movement. While reduction in finger repetition is also a tangible concern, it is a less consistent part of the typing experience, hence the lower weight.

Other people may argue for more sophisticated metrics, but there’s only so much you can try to improve in a minimalistic approach like Minimak’s, so I question the value of trying to satisfy more sophisticated metrics. I also question how valuable other metrics actually are. At the end of the day, you have to actually be able to feel the differences and I can only feel finger distance and repetition.

Why compare to Dvorak and not [layout X]?

I know more people who use Dvorak than any other alternative layout and it has more OS and hardware keyboard support than any other.

There are layouts which have better results than Dvorak on my metrics, notably Colemak. I don’t try to compare to it as well so as not to confuse things with even more metrics than there already are. Dvorak is just a name people can relate to.

How can you call changing 4 keys a layout?

The numbers speak for themselves. E and T are so important and are so misplaced in QWERTY that you can spend enormous amounts of effort trying to finesse the other 25 keys and still not get the improvement that moving these two keys gives you. It may take a bit for that to sink in, but when it does, you may realize that by some measures, the 4-key layout is the best Minimak layout.

This is because most people assign a high cost to relearning keys…high enough that most won’t even consider it at all. If you associate a high cost with learning keys, then the only keys you’ll be willing to learn have to meet a certain threshold of improvement. For some of these high-cost users, the 4-key layout is the only layout that has enough improvement per key to be considered.

If Minimak is such a great idea, why hasn’t anyone come up with it before?

My theory on this is that as a layout designer, it’s only natural to assume that the only way to replace such a massively successful layout as QWERTY is to design the best possible layout you can. And let’s face it, you have to be a real nerd to even think about redesigning your layout in the first place. Someone that obsessive is going to resist the idea of offering less than the best they can come up with. So the natural inclination is to start off by throwing away QWERTY and designing from first principles in the pursuit of a pure optimization strategy. After all, we all agree that QWERTY is a terrible layout, so why use it as a foundation?

But unless you spend time looking at the improvement you get on a key by key basis, you may miss an important fact. That fact is that there is a dramatic range of impact to the spectrum of key changes. Even if you do notice this fact, it’s not immediately obvious how it makes a difference. But what it means is that when you can make a large difference with a small number of key changes, an incremental improvement strategy has a lot of value by lowering the cost of the learning curve. This observation is what makes an incremental improvement strategy favorable to pure optimization, seemingly against common sense. At least, that’s my take on it.

What are your thoughts on [layout X]?

First, I’ll say that a lot more thought and research has gone into most of the other layouts to which I’m comparing Minimak. Minimak is great for all of the reasons I’ve outlined, but there are a lot of great layouts out there as well.

The carPalx layouts are the result of an amazing amount of modeling. While I can’t say I agree with the particular balancing act he uses for quantifying a good layout, I can say that I learned the fully optimized layout and it felt great. I should also mention that I haven’t measured it against my own metrics, but I’ll guess it does very well. The designer has done something right. If we were inventing keyboards for the first time, I’d be willing to bet one of his layouts is the best.

Colemak is a better layout for most people though, since it does take QWERTY into consideration, even if it does fall short of having been designed for QWERTY. Colemak scores really well on the metrics I use. However, I think it is mostly the result of a pure optimization strategy and doesn’t employ the learning principles found in Minimak.

Colemak also has a set of stepwise intermediate layouts like the Minimak progression, called Tarmak. Because Colemak is a much different layout from QWERTY, there are more steps (5) and each is more complex than the Minimak steps, but they in part inspired the Minimak progression and I believe they speed the learning of Colemak for similar reasons to Minimak’s.

Dvorak is well-designed, but when compared to Colemak, it seems to pale a bit. Colemak offers better metrics for a lower learning curve.

I found the Workman and Asset layouts interesting and incorporated some of their ideas. The big one is that valuing key positions by row is wrong. You have to value them by their reachability, which varies by finger and hand due to the asymmetry of the standard layout. carPalx seems to get this wrong.

That said, I believe that Minimak’s approach is both unique and better than the rest for QWERTY typists (meaning just about all of us) because of the way it addresses the realities of QWERTY.

Do you have any other keyboarding tricks?

I use Autohotkey quite a bit for making frequently-used parts of the keyboard more accessible, such as the arrow keys. I haven’t documented it, but it’s reminiscent of what is described as home row computing.

There’s also DreymaR’s big bag of keyboard tricks, which has some interesting ideas (which I haven’t used personally, but are still interesting), most notably the idea of a split-keyboarding on a regular keyboard by remapping the right side of the keyboard one column further to the right, and putting punctuation keys in the gap that is created. It widens the natural arm position as well as relieving the pinky of a number of keys it’s responsible for and lowers the pinky distance to the frequently-used Enter key. These seem like worthwhile ideas if you don’t already have an keyboard which already does these things like my TypeMatrix does.