Show HN: Available .com's in /usr/share/dict/words | Hacker News Reader

words

pastebin.com

43 pointsgamegoblin49 comments

Histogram of word length. Keep in mind that the median word length in English is 5:

              5     1 #
              6    14 ##
              7   104 ###########
              8   326 #################################
              9   414 ##########################################
             10   380 ######################################
             11   320 ################################
             12   200 ####################
             13   133 ##############
             14    73 ########
             15    35 ####
             16    12 ##
             17     8 #
             18     1 #
             19     1 #
             20     1 #
             22     1 #

edit: Someone points out below I had an off by one error, which I've now fixed.

Sorting by length is an interesting programming problem; it can't be done by a regular expression. One way is to use a mask:

    % sed -e "s/./x/g" < original.txt > mask.txt
    % paste mask.txt original.txt | sort | cut -f 2
    % rm mask.txt

I did this to make the above:

    sort /tmp/foo |awk '{ print length, $0 }'|sort -n|awk '{print $1}'|uniq -c|awk '{r="";i=s=$1/10;while(i-->0)r=r"#";printf "%15s %5d %s %s",$2,$1,r,"\n";}'

Golf time! in js:

require('fs').readFileSync('list.txt').split('\n').sort(function(a,b){return b.length-a.length});

Python:

sorted(open("list.text").readlines(),key=len)

Haskell:

(sortBy (comparing length).lines.readFile) "list.txt"

Eliminate the temporary file:

    sed -e "s/./x/g" < list.txt | paste - list.txt | sort | cut -f 2

We can eliminate the paste by using sed's hold buffer, but it costs another regexp to get rid of the spurious newline:

    sed -e "h;s/./x/g;G;s/\n/\t/g" list.txt | sort | cut -f2

I saw more than one six-character domain - befogs.com, dourer.com, etc. Are you sure that histogram's correct?

I think it's off by one. There is 1 five-char domain.

You're right, it was off by one. I just fixed it. Thanks!

The vast majority of the good ones are taken, unfortunately. I am guessing he's not doing whois but instead looking at DNS records.

I have a daily cron job that looks at available domains sorted by word popularity. The vast majority are in some status that makes them not show up in DNS but are registered.

That said, I've been thinking of making some of the domain name tools I'm working on public...

I'm using a modification of https://gist.github.com/peterc/63893

If you're interested in doing more of this stuff, I recommend getting in touch with verisign - they can give you free FTP access to copies of the .com and .net root zonefiles.

Not really at all. This was just a 20 minute morning hack for fun. Thanks for the info, though.

Why are those not publicly available?

flimflammed.com is available, if you are looking to make an old-timey version of Snopes

Best one I saw in here was "pledges," which is screaming for a fraternity-run social network.

    Domain Name: PLEDGES.COM
    Updated Date: 19-feb-2014
    Creation Date: 11-jan-2000
    Expiration Date: 11-jan-2014

So it's probably in redemption

These two would be good for specialized dating sites.

homeliest.com - A dating site where ugly people meet other ugly people.

outfoxes.com - Great name for a gay dating site.

So I tried registering outfoxes.com. It was available and in my cart... I get to check out and it is no longer available, but is now "up for auction" by name.com. WTF?

I was planning on running it again tonight just for fun. There were 2024 to start with. We'll see what happens.

I'm curious, how many of them were snatched?

Some interesting numbers:

    $ wc -l domains.txt
    2024
    $ grep -Ev "(ing|s|ed|ly|est|ier)$" domains.txt | wc -l
    127
    $ grep -E "^.{1,7}$" domains.txt | wc -l
    119
    $ grep -E "^.{1,7}$" domains.txt | grep -Ev "(ing|s|ed|ly|est|ier)$" | wc -l
    15

The 15 words from the last command:

    almohad
    ayyubid
    boneyer
    chaitin
    copulae
    dourer
    eagerer
    layamon
    livia
    muawiya
    neglig
    shlocky
    unriper
    unsafer
    vilyui

chaitin.com: a service that solves the halting problem

Here is the same list sorted by length:

http://applied-math.org/words_sorted_by_length.txt

The last on the list, counterrevolutionaries.com, seems like it could be a useful site for some. First, start with bios grabbed from Wikipedia, then use it to enhance your opposition of your political enemies.

Cubadebate, for instance, could list twitterers as "necrophiliac counter-revolutionaries".

Nice! Just a reminder of what's not here (and thus is registered):

All but one of the English words in a 100k word list that is less than 6 characters long. Average English word length is about 5 characters. Kind of mind boggling.

chancel.com - get a 50:50 chance of cancelling your process or killing pid 0.

cuddlier.com - our cushions are

marauded.com - rebuilt your home in this upcoming ios freetoplay.

besieging.com - the long awaited sequel to Marauded. Spent the rest of your money on useless hats.

gawkiness.com - exposing the shady business practises of the unmentionables

offstages.com - your stars in authentic interviews

Great hack!

Deeply trying to resist buying shlocky.com

Was looking for an excuse to play with Haskell threads, so I read in the dict, put it into groups of 500 words each, and then distributed each group to one of 200 threads to whois everything.

A music notation startup's field-day Multiple music terms: Decrescendoes, diminuendoes.

What's your retention rate at decrescendoes? <add joke here>

If you're creating music-related startups or software, you're probably better off getting a .fm domain, since it is easier to find good domains there and users are accustomed to it.

Funny enough, I actually considered some of the terms you suggested (my software fades out music when you play a video) before I started thinking about .fm. The obvious problem with the terms you mentioned are that they are long, difficult to spell, and obscure to non-musicians.

No .fm necessary for sextettes - the online dating site for music professionals.

New words for my vocabulary:

congaed - past tense&participle of conga scrod - young cod/haddock/white fish split and boned. (also schrod, also available).

symmetricly isn't a correct spelling as far as I can tell. This was just one that jumped out at me, looks like others might have similar issues.

Several of these were already registered for a long time, So I'm not sure how the check was done, but I did pick up UNSeats and UNsays, but I will likely use them as United Nations Says, not unsays which I don't think is something you can do.

What's interesting is how many of the words are "negative" or have negative connotations.

I'm struck by how many are pluralizations or modified in some other way (-ier, -est, ...)

I wonder who will purchase urethrae.com. I thought fetish porn creators would have bought it by now.

Why can't I find a clear explanation anywhere of when I'll be able to buy .app domains?

This is using the ~100K American English wordlist.

These are some pretty shit domain names hehe. Here is everything with fewer than 8 chars. Lorded.com seems promising though.

with open('domains.txt') as domains: for line in domains: if len(line) < 8: print line

befogs

cagily

dourer

drolly

gusted

livia

lorded

neglig

scrods

shirrs

soughs

tyroes

unmans

unsays

vilyui

You seem to be counting \0 as a char, or meant to find words less than 7 chars. Either way, there are just over 100 names available with 7 chars.

Probably counting \n as a character since Python is pretty good about never dealing with \0's.

can't beleive "liviva" wasn't taken as a mispelling of "Libia"

There are several good ones that I could see possible businesses wanting, such as "surfboarded"

Nobody reads Joyce anymore.

tingliest.com - The 9GAG alternative the world has been so eagerly waiting for!

tinglier.com - oh shoot!

livia.com would be the best of the bunch; whois shows it taken.

Reply on news.ycombinator.com