willem3141 / plover_dutch Goto Github PK

Can be tucked into words where this seems useful (broodje, meisje, etc.)

This was Jeremy's idea, and I like it 😃

These have all the same vowel key, so loei, looi, and lui are all written as loi (and also, loi* is for luie). This kind of conflict turns out to be pretty common (mooi / mui, tooi / tui, oei / ooi / ui, sproeit / spruit, ...) and is annoying to solve (* is not always available for other reasons).

I see two possible solutions:

• write ui as au, write au (and ou) as ou instead;
• write ui as ou, write ou (and au) as au instead.

Both have the advantage that the au/ou sound is not spelling-based anymore, but there could be some conflicts there (also with the current usage of au for prefixes, for example aut is currently uit^).

In the README it would be nice to have some list with some common conflicts, and how to work around them.

Here's some:

• BA*NG: bange, bank
• OI: ui, oei, ooi
• more?

It would be nice to be able to write example sentences like i voel me naag -n beepg ziek instead of writing out all the keys like EU SRAOUL ME TPHAEG -PB PWAEUPG STKPWAOEBG.

I should make a table containing all pseudo-steno patterns. Also, we should have a Python dictionary that outputs pseudo-steno automatically (this could be used as a fallback for words that are not in the dictionary). Maybe even taking some Dutch spelling rules into account, so it would write zo instead of zoo for STKPOE.

The license in this repository is CC-BY 4.0 according to the LICENSE file. The setup.cfg however lists GPLv2+ in the metadata. Maybe I am misunderstanding which license covers what, but shouldn't these two be the same?

This was a nice idea, but -TSDZ is really horrible to stroke. I just got another idea, which is to use just -Z for this.

For words that do not have any consonant before the -t, this doesn't really create any conflicts, and can even be implemented by an orthography rule. (Special cases: weeën, zeeën.)

Some other words do get a conflict (lachen / lachten), but in that case it's usually a past tense, and the system doesn't support doing past tense forms in one stroke anyway.

SR for schr- obviously clashes with SR for v-. This results in clumsy conflicts such as vijf / schrijf.

Let's figure out some chord to do schr- with instead. Maybe make sch- SKH, and then use SKHR for schr-? That immediately gets rid of any sh- / sch- conflicts (shot / schot).

Note: I cannot do SG for sch- as I'm using that for z- already.

In the plover_dutch.py file, the field implicit_hyphen_keys should not contain '0-' and '5-'. This causes an error! I do not know how to change this myself

Compound nouns are pretty common, but at the moment there is no consistent system for writing them.

In the English system, where there are not so many compounds, this is handled by adding a star to the second noun, but only if a conflict would occur otherwise. For Dutch this doesn't seem like a very good idea, because you get many such conflicts.

Also in Dutch you can get -en, -e, or -s as an infix in between the nouns, which should probably be stroked as well (I think not stroking those would be confusing, because you don't stroke phonetically anymore).

My current idea is to always add a star to the second noun, and stroke the infix with the first noun. That is, -en and -e both get stroked by -Z, while -s gets stroked by -S.

An alternative is to stroke all of those with -Z; this avoids issues when the first noun ends with -d. This is maybe problematic if there are compounds that can have both -en/-e or -s, with a different meaning.

Another problem is that many nouns can be stroked in various ways. So then, for compounds you get a combinatorial explosion of possibilities. I'll need to find out if there's a way in Plover to handle this in a nicer way (that is, without having to add all of those alternatives to the dictionary).