Fun with words

A while ago, I had this fantastic project to be able to, somehow, translate place names into other languages. (Fantastic as in weird and wonderful, at least, and not necessarily amazingly good. Cleodhna, for instance, was not too impressed with it.)

I revamped that project recently as something more limited: a method of generating, from a list of words, a number of made-up, but similar, words. Here’s how I did it:

• I took a list of English words (the standard Linux dictionary is ideal for this purpose, as it’s a list of words separated by linefeeds, without any annoying cunning compression) as a base dictionary.
• I used a standard Perl module (Lingua::EN::Hyphenate) to split them into syllables.
• I did some basic munging to work out how likely each syllable was to follow or precede any other.
• I then wrote a script to make up new words that seemed likely, but weren’t in the original list.

Some sample words follow:

optivation

Activation in an optimal way.

mufflemancer

growlingua

How you speak to dogs

argussions

The direct result of a heated argument coming to blows

builderous

mellosities

calcularged

Describes how contractors work out their invoices

bafflegenda

Descriptive of plans that make no sense

oregotism

Selfishness that smells nice

maggotten

The quality of being a naturally-born fly

fallatuo

Oral sex based upon a false premise

collarhea

Like gonnorhea, but higher up

carnageway

A dangerous road

mellonnaire

Someone who has enough money to fit into a large fruit

bootlegian

Something that has the bootleg nature

The beauty of this approach, of course, is that you can apply it to any body of words. I put a list of hosting companies, as listed by The Open Directory, and generated a bunch of plausible-looking names. Amongst the names:

• Blue Zebra Hosts
• Veripower
• Constructure
• Simplicision
• Emerging Servers
• Virtu Request

All of which could be names of hosting companies. Some, on the other hand, were more doubtful:

• Rack my easi
• Dino Host
• No host your dot
• Global tdiots
• Lame H Nexus
• Expressfully