thought i'd see how this method worked. my opinion so far is that it's a semi fun language toy, but not a production ready fantasy name solution.
at first it took a lot of wading to find passable results. the best application i had for yielding actual useful names was to heavily limit it.. adding more possibilies results in wilder output, but if you work with only half a dozen letters and an intended meter you can find more useful names.. so if you do this four or five times with different parameters you'll get several sets of names which will be sufficiently diverse when mixed together.
examples.. i thought these were 'fairly human sounding but different' - they took a fair amount of wading through wide parameters... olvil, ciura, haro, lasia, cichrilh, ulsulea, rehesh, sala, sarilh, lacli, eshlish, lacev, luca, cucla, toteash, selhsee, husil, halcor, vesa, rove, ashoth, sivalh, cilara, carilh, iashhe, alhvaro, lavelh, seuclius, vaoli, cirolh.
but using a very narrow set resulted in a much higher percentage of useable names..
cenade naimde
naende caceem
mecena emidme
amniad daca
memide meeam
cedena ecimde
ciad nemiama
nanaed miad
nemine andame
adniam nane
cename naicna
naende daecned
so i manufactured names in small batches by gradually varying it..
naelnde dhaechnedh
neddicedh mecadhni
necha chanedre
meldcimedh chanchcani
nemedh emhanchelm
ninnamh caclenchnelnh
drenamaldh alchcrenhund
meddralnna enenneum
nadha nedemi
naeln cumae
nilemue ulalhen
luim icil
nelanad emhcoerd
elaelun louv
unimne eduel
avelnlae auvenna
lonaen muev
evhlu naevuon
imbra neuvh
uenalul luunon
naola nevec
anmede feuanadh
cenide naodhlo
vaene difa
abnide coulla
molna alam
oevna luliol
medhe onna
perhaps pick a word or two like firvulag, and just use those letters at first. add in a few modifiers and maybe another consonant, and you'll get better output than just setting each probability as it appears in the english language in general.
please let me know if you have any ideas for improvement, or even how this is or could be useful to you.
as is i need to add extra randomisation to the seed value.. the names displayed on the screen use sequential seeds, and if you use elementary settings, eg. 3 vowels and all syllables beginning with vowel, you'll see the same few vowels in bands.. there's only one rnd operation to the seed before this, i'll add two. might pull my finger out and add a normal file naming dialog for saving the txt file...