This is a quick port of Martin O'Leary's naming-language code from JavaScript to Go. He describes in detail what it does in his blog post.
Why?
Why not? I felt like it.
What does it do?
It generates a naming language based on a bunch of tables which specify phonemes & orthographic mappings along with some other rules.
What's a naming language?
This is a language which is intended to look good on paper, but doesn't have any real semantics (meaning) behind it. It might be used to name cities, regions, or land features on a fantasy map for example. Because the words are all generated from the same rules, they should be consistent with each other.
Should I use this code to learn Go?
Nope. Absolutely not. This is pretty non-idiomatic Go code. It might even be anti-idiomatic.
I started with a direct port of the JavaScript, then moved some things around and added a couple of capabilities like choosing random syllable structures from a list.
What does it looks like?
Building it and running it generates a random language based on internal tables and spits out some words and names like this:
$ go build && ./naming-language-gen
[words]: kun, eeak, iwt, mahkuz, tužnaskun, hjužkuziwt, koemahhjuž, hemwepsmot
[names]: Tounak, Tužnaskun, Hemwepsmot, Paoshaš, Hwotkun, Eeak, a Kunkiw, Itmah,
Kun Iwt, Hjužkun, Kuzužak, Eki-Kunkuz, Mahkuz, a-Kunkuz, a Iwt, Ipmonpuwn,
Mjemiwt, Iwtmiu, a Kunkuz, Umsmot
-> apply ortho: true
-> apply morph: true
-> phonemes:
C: ptkmnh
V: aeiou
S: sʃ
F: sʃzʒ
L: wj
-> restrictions:
-> consonant ortho:
j => j
ʃ => š
ʒ => ž
ʧ => č
ʤ => ǧ
-> vowel ortho:
A => au
E => ei
I => ie
O => ou
U => oo
-> morphemes:
> 'the'
a
> 'words'
kun, ak, iwt, mah, smot, shaš, kok, te, ip
> ''
e, kuz, tuž, nas, hjuž, koe, he, mwep, tis, pajp, mwez, hwot, mjem, mon,
toun, tuz, pao, iwh, pwek, už, šnes, tus, nit, kiw, kez, ki, it, miu, oh,
en, op, puwn, nwih, han, ke, um, ket, hwak, mek, pož
> 'of'
up
Huh. Weird. Can you give me another example?
Sure.
$ go build && ./naming-language-gen
[words]: igtad, nartad, sumskan, tadsurk, skan, tadirm, urdtad, schiskan,
tadig, tad
[names]: Tadig, Schiskan, Skan Tad, Igtad, Tadirm, Unulmingstu, mu Stugam,
Surkumulstu, Nartad, Urdtad, Unstu, Skanig, Tadsurk, mu Tad, Imkiasguag,
Sumskan, Tad Saschan, Tadkim, Artstu, Tad-Saschan
-> apply ortho: true
-> apply morph: true
-> phonemes:
C: tkdgmnsʃ
V: aiu
S: sʃ
F: mnŋ
L: rl
-> restrictions: [sʃf][sʃ], [rl][rl]
-> consonant ortho:
x => ch
ʃ => sch
ʒ => zh
ʧ => tsch
ʤ => dz
j => j
-> vowel ortho:
I => ii
O => oo
U => uu
A => aa
E => ee
-> morphemes:
> 'words'
tad, skan, stu, gut, guag
> 'of'
sasch
> 'the'
mu
> ''
ig, nar, sum, surk, irm, urd, schi, gan, nusch, sar, ti, kasch, ming,
kas, nug, un, ul, gam, isch, giis, um, art, schan, im, kim, kias, sa,
kang, saim, schim, sching, tik, urm, dik
That's it?
Yep.
Why doesn't the program take any input?
That would be the next logical step. The internal tables should be made external so they can be easily changes and fed into the program along with a language description. This would allow experimenting, fine-tuning, and reusing of language descriptions that work well.
What else could it do?
Martin has some suggestions at the bottom of his post which outline other things that could be added.