genkey VS a200

semilin/genkey

if you can extract the text from those emails and put it in a single file what you can then do, download genkey, (docs can be found here, and load that file in to do analysis. If you wanna try handmade swaps without generation, you can use this little thing I made which lets you drag and drop to swap and rightclick to disable keys which can give you some immediate feedback on the changes you make.

genkey is usually what's used, you can also use your own corpus with it. a200 is also pretty useful but might not work on windows

https://github.com/semilin/genkey this is the one I have in mind currently. I currently use a keyboard layout made by this, and I also know semilin.

More importantly it's got things to look for in section 2: stats to rate a layout by, usually generated by the DH analyser, genkey (from the person who created Semimak) and lately the a200 analyser (by the person who created the Pine layouts). In a very general sense, sfb >> dsfb > redirects with onehands, rolls and alternation being personal preference, given a sensible hand and finger balance. Stats shouldn't be a crutch to justify a certain layout, but they can give valuable insight in how a layout does in certain aspects and how it will perform generally. Even people who do know what they're doing rely on analysers a lot for sanity checks as they make misakes really easy to spot. Creating layouts is a lot of fun though, and me and the homies in the alt layout discord are always willing to help :)

I'm guessing you have no prior experience with creating layouts, so I think you should probably make yourself familiar with the general idea of what makes a layout good or bad. I think section 2 of this document should give a good overview (I think it's useful in general, there's also information about the data used for analysis and contains a tonne of good layouts, sorted by sfb). A lot of those you can find with analysers like the colemak DH analyser or genkey, and they can give you a good idea of when a layout is 'good'. To be very brief, generally it comes dowt to sfb >> dsfb > redirects with alternating vs rolling and onehands being personal preference, assuming that you keep a sensible finger and hand balance. Stats aren't the end-all be-all, but are really useful guidelines. Even if you do know what you're doing, it's almost inevitable you analyse your layouts at some point.

About your last paragraph: I 100% agree with you! And that's the beauty of alt layouts, it's all preference. But I want to make this clear, if you are looking for an optimal layout for portuguese, Colemak is not the way to go. And in my experience, all layouts that I've seen are trash for portuguese. If you want to create one, go ahead, as long as you know what you are doing. You can use Semi's analyzer for this, just insert your own corpus and see what you end up with.

Results were geneted using Genkey, the analyser used to create Semimak. The analyser grades layouts with a score (based on a bunch of metrics), the lower the number the better the layout basically which is what the numbers in the bottom mean: [191, 195] are all layouts that scored between 191 and 195

I think they are simply a % at the moment, not really weighed for distance/finger strenght. Semi's genkey has 'finger speed' which has not yet been included in the doc but which should be what you're looking for. It's essentially how often, far and fast a finger has to move, weighed by how strong a firger is supposed to be (stats garnered by Semi), so a 2 row sfb would mean very high speed (bad). I think that's what you mean by finger weighed?

Back to SFBs: I'm getting the same result with the tool you provided, but the before mentioned a200 tool (https://github.com/ClemenPine/a200) gives me 2,73% and now I'm wondering where does the discrepancy come from.

genkey is usually what's used, you can also use your own corpus with it. a200 is also pretty useful but might not work on windows

More importantly it's got things to look for in section 2: stats to rate a layout by, usually generated by the DH analyser, genkey (from the person who created Semimak) and lately the a200 analyser (by the person who created the Pine layouts). In a very general sense, sfb >> dsfb > redirects with onehands, rolls and alternation being personal preference, given a sensible hand and finger balance. Stats shouldn't be a crutch to justify a certain layout, but they can give valuable insight in how a layout does in certain aspects and how it will perform generally. Even people who do know what they're doing rely on analysers a lot for sanity checks as they make misakes really easy to spot. Creating layouts is a lot of fun though, and me and the homies in the alt layout discord are always willing to help :)