I briefly explain there how I had done it in the past, which inspired me to take a second look at the algorithm and see if there were ways I could improve it. I wound up completely rewriting it to calculate the most popular domain name prefixes and suffixes used in all domain names. This article is about that process and the results.
Let me know if you have any questions.
You should consider a premium subscription where members pay 5-10 bucks to see even more domain options. Folks who buy lots of domains should have no problem paying for the utility you provide.
I say this as someone who has found bunch of domains using your site but not once used your affiliate link to buy because you don't have my preferred domain registrar choice.
I'm very interested in a more premium service, however, where I can pay money to have someone who's done this before find a name that's either free, or available at a cheap price. I've tried things like PickyDomains, and some others like it, but most of what I get is the same as if I just used tools like LDS myself.
Consider if I were writing a new Email package.
I might tell your tool "Mail" or "Message"
As great as LDS is, it's going to give me "GetMessage.com" or "HiveMail.com" not anything like Pegasus Mail, Cyrus, or Hermes.
Those are much more creative names!
These are all stories about mail delivery, to an extent, but any computerized system is going to have a hard time with it.
The best that I suspect could easily do would by synonyms - If I give it Dark, it could suggest Black and Dusk, and give me names using those..
But I'd love to see something with a less direct connection.
So you know, LDS does list synonyms on the right side of the search results page, which should help to some extent. Searching for "mail" wouldn't turn up Pegasus Mail, though I could add "pegasus+" to the search results and it would. There's a balance fine line between including a variety of results and including poor results and I need to be careful not to cross it. Any recommendations?
I don't think the problem is that you don't have enough generic words to add to either side.
I think the problem is that all your doing is adding words to either side ;)
I like your site. It's really useful to get the low-hanging-fruit.
But for instance, Cyrus is a awesome name for a Mail program, because Cyrus the great is credited with the creation of the first postal service.
That doesn't mean that CyrusCoupons is a good name ;)
Many names are good because of the metaphors they invoke, or because of the story associated with that character, place or object.
"Teapot 3d" would be a great name, because it's a Teapot is a traditional object to render.
Lena is a great name for image manipulation, because of.. http://www.ecogito.net/articles/lena.html
Again, I think your site is great for what it is, but a lot of my favorite names.. Really great names.. come from much more oblique or abstract approaches, for that particular problem. These don't generalize well, by their very nature.
Does that make sense?
There's a lot of smart folks on here... can anyone think of a way to algorithmically come up with domains like Colin is describing?
Another idea: get a dump of Wikipedia and do a TF-IDF analysis on all of the words. When someone types "mail", you look at wikipedia.org/Mail and look for the most unique words. As it happens, the Mail article does mention Cyrus the Great, so as long as Cyrus is a pretty rare word in the corpus, it might be a suggestion you can make.
That said, there is such a thing as too clever. If I typed in "mail" as a domain search query you told me to create the website "CyrusMail", I would wonder what the heck that was all about because I don't happen to know the relationship between mail and Cyrus the Great.
More abstract relationships (mail -> Cyrus, image -> Lenna) might be gleaned by crawling Wikipedia.
Adding "i" to any word seems to be all the rage these days. And "e" would be another popular single letter prefix, I think. I can't think of any single letter suffixes though, except maybe "y".