Not a huge change today, in terms of work, but I think this makes each generated list of names (instead of individual names) more interesting.
Up to now, except for some special name types (e.g. military operation names, English names, etc.), the “size” selection worked like this: “short” names had 2 syllables, “medium” names had 3, and “long” ones had 4.
As of now, the above values haven’t changed, but some variation was added: for each generated name, there’s a 10% chance of having one fewer syllable, and another 10% chance of having one more. So, a 20-name list of short names, for instance, will, on average, consist of 16 2-syllable names, 2 1-syllable ones, and 2 with 3 syllables.
One consequence of this is that, for the first time, 1-syllable names (and 5-syllable ones, if you choose “long”) appear from time to time (and some of them are even decent!). But, more importantly, each list of generated names is less uniform, and therefore less predictable. I hope you’ll agree that this is an improvement.
I also added the two most recent name types, “reptilian” and “Aztec-like”, to the Portuguese language version.