You *are* aware that you've thoroughly shifted the goalposts from your original suggestion of just using character replacement to a hybrid system that now has to be able to pull out "fixed" elements (and transliterate these - which will require essentially coding these "fixed" elements as either drawing from the namelists, or require the "fixed" elements to be given transliterations) and separate these from translatable elements?
Yeah, and it would be a lot of work, but it wouldn't be orders of magnitude more than finding the names for all the baronies in the first place. It would definitely be something on the order of magnitude of the workload they had in Reaper's Due, when they added the potential for seven holdings to every province, but that was something they already did once. We've already dropped the workload down from "the time it takes to make something better than the best linguistic software in the world" to "something which would take some of the dev team a few months".
It's now not just a matter of finding the names. It's a matter of finding the origins of the names, their etymology, and deciding which (if any) should be translated by etymology, and which (if any) should be transliterated. This is not an easy task, and now required the team to be not only linguists capable to coding a transliteration system, but also to do deep dives on the history to obtain etymologies that are in dispute or now lost.
And it's gone from "oh, I've fixed it in 20 minutes" to "it'll take a large portion of the team months to do it".
And it *still* has to address being able to decide which phonetic value a given letter or group of letters have in a given context, and what to do if a language doesn't have that phonetic value.
If it becomes something that takes the dev team a few months and becomes most of the feature of a DLC, that sounds like it's a DLC that won't sell well, unless it has *amazing* other features.
Why does the game need to know how how to pronounce it? And heck, why does it need to know which of borough, burgh or bury to use? if you made it choose randomly one of the three most people would never notice that there was anything wrong. And if one culture doesn't have a town suffix, then it would probably just keep the name the other culture used which does.
The game needs to know how to pronounce it so that it knows which sound the syllable represents in the native language, and which to represent it with in the new one. Or, more precisely, it has to know which syllables in one language correspond to those in other languages, and which are is the "correct" ones to use.
The word "fish" could be equally validly represented by "ghoti", but the latter would not be considered a "correct" transliteration. Similarly for "ghoughpteighbteau", which is a perfectly valid phonetic representation of "potato".
It has to be able (somehow) whether or not this is done manually by assigning something like IPA values behind the scenes be able to distinguish similar (or identical) looking names and give them a correct value if it is even going to attempt to get the names for the provinces right when assigning them to a hybrid culture that may be using the orthography of one parent culture, or a hybrid of the two.
Off the top of my head, I don't know if there's a problem with having a culture have more than one identifier for the same "equivalent" name fragment. It might be that it breaks something badly. It might have no effect. But it's something that has to be concidered.
If a culture doesn't have a suffix, then it just outright keeps the old name? It shouldn't perhaps transliterate it like would happen with other names that don't have a translateable part?
I'm not sure if I didn't make this clear, but the idea is that every province name would contain the basic blocks to tell the system how to transliterate it correctly - that's what would include things like 'translate this as "FF"' or 'Translate this as "Town"' etc.
So the devs have to go through every name in the game - including ones that have the same name but different pronunications, assign them all IPA values for the elements that are transliterated, and equivalents for every element that has to be translated.
To quote another forum member from this thread about giving them IPA values.
This is the only thing that would make it fine, and it's just too much work.
The same looking word in the same language can have two (or more) different pronunciations, and there's no short cut to obtaining the values for the phonemes in these names.
It would have to essentially be done manually.
The point isn't that this system we've come up with is easy, nor that the system as a whole is simple - the point is that by adding simple rules, finding the break cases and then adapting them we've gone from the insurmountably impossible task of "create an algorithm which understands human language" to "create a scripting system which takes a finite number of inputs and gives outputs which look similar enough to good answers that most of the time it doesn't feel jarring".
Essentially going from "Grotesquely infeasible task" to "Major DLC feature".
We haven't really begun to touch on break cases.
The simple rules aren't simple.
And again, if this were to be a "major DLC feature" it's somewhat dry and likely to drive sales down, especially with the opportunity cost of doing something gameplay related instead, and the necessity to hire specialists to do this sort of work.
Obviously yes, that's why I said as much in that the original post in the chain and every post subsequently. My point, as I have continually re-iterated, is not that this is a good idea, fully formed and ready to go, nor even that this is a good idea with a few kinks to work out.
My point was, and is, that if I could come up with one idea off the top of my head in twenty minutes, and we could flesh it out as we have done over the course of the last couple of pages - working out the flaws, possible work arounds, things it does or doesn't do well etc., that a dev team actively trying to solve the problem could also come up with ideas, discuss their merits and flaws, find things which work and things which don't and come up with a viable solution, which is the process of finding workable solutions.
And so far you've moved the goalposts so it's now an entirely different concept - and far more complicated and difficult - and you've largely ignored the point of how difficult it is, even asking why the game needs to know how something is pronounced in order to transliterate it correctly.
What I mean is that if you don't know how to fix something, come up with random ideas and work out why they don't work - it will give you a good understanding of how to make something which will; if you can come up with ideas, then you can make something which is good enough - not by implementing the first idea, but by taking the one or two good bits from it and combining them with some of the good bits from the second and third and fifth and ninth idea. It's almost always easier to fix bad things than make good ones from scratch so start with something bad and fix it. That was, and is, my point. We can have ideas, so solutions exist.
You suggest just getting some of the devs to do it - they're mostly game designers, not linguists. It isn't something that would really fall into the field of most of the dev team, and for those for whom researching the possible pronunciations *is* adjacent to their field (the historians) they'll need those for other historical elements of the research and design for other DLC.
And, whilst cramming together partially broken system after partially broken system to try to get a good solution is possibly easier, it produces complicated, unwieldy and generally poor solutions to problems, rather than good working systems. Even worse, they become solutions that are hard to replace when you *do* have a good one.
An "easy" solution is not necessarily a "good" solution. "Everyone is hungry, so let's eat the seeds we were saving for next year" is an easy solution, but long term it's not a *good* solution. Rationing food to get through to the next year and be able to plant is a harder solution, but far better.
There is also, as I mentioned earlier an opportunity cost, in that doing all this work to obtain IPA encoding for every possible place (and personal) name to enable transliteration is difficult. It may also require hiring a whole new team to do this difficult task.
