Domain Pigeon Fixed

Domain Pigeon has been down for the last few days.

Recent visitors to the site saw this uninviting error message:

apperror

The error, “unitialized constant ApplicationController”, was remedied by SSHing into Dreamhost and running the following command:

rake rails:update:application_controller

… which renames application.rb to application_controller.rb

I think Dreamhost updated something on their end which caused the error.

Regardless, its back up — more domains to come shortly.

Wordoid

A new domain name tool just launched called Wordoid.

The thing I like the most about it is how it lets you enter in a word and then it generates available domain names around that word. Domain Pigeon, on the other hand, only lets you search names that already exist on the site.

There were a few interesting mentions of Domain Pigeon in the Hacker News thread on Wordoid’s launch:

Yeah, I was a fan of http://domainpigeon.com/ for a while, but that got a bit too restricted for me (as it became more commercial), not to mention a bit unwieldy to browse through. Bookmarked!

Wow, it’s like domain pigeon only way better. Awesome.

It’s nice. It reminds me of domain pigeon, but the words it creates seem a lot more natural.

The first point is valid (visitors can’t explore a lot of the names until they sign up). The problem, in my experience, is that its very hard to make money from a site like this purely as an affiliate. Jorge, from hotnamelist.com, can attest to that fact too.

I’ve had a new feature on the todo list since the site launched that would let you do exactly that, but I’ve let a few technical hurdles stop me from implementing it.

Maybe this is the kick I need to get around to doing it.

A Long Road Forward

I’ve spent about a year working on Domain Pigeon and I’ve learned a lot in that time and I’m proud of what the site has become, but its time to start thinking about what’s next.

I’ve been tossing around a few ideas for what to work on and I think I’ve finally settled on a project. I’m very excited about it and I think it will be life changing for a lot of people.

I plan to use the following technologies to make it:

  • Python/Django
  • Flash/Actionscript
  • jQuery
  • Google Maps API

I could probably do it with Ruby on Rails and Prototype, but a lesson learned from a decade of Visual Basic work has taught me that it’s wise to learn something new for every project you undertake. In this case, I don’t have any experience with any of those things, so it’s going to take some time.

At some point in the next few months I’m also going to use the profits from Domain Pigeon to form an LLC to cover this project and any others I might start in the next few years.

I will continue to work on Domain Pigeon as time permits, but with only a few hours a day to work, this new project will be my priority. Domain Pigeon will continue to be updated daily with new domain and Twitter names and there are a few more major features I want to implement, but the days of small layout changes are at an end.

Lots to learn, lots to do.

Post Mortem

For the last two weeks I busily worked on a new addition to Domain Pigeon that would let visitors sort the domain names by pronounceability, as determined by folks on Mechanical Turk. I was excited. It was an innovative way to help people find great domain names and it gave me an excuse to post another link to it on HackerNews.

And so, on Monday morning I posted it on HackerNews: Finding Great Domain Names using Mechanical Turk.

I eagerly awaited feedback, refreshing the comments every few minutes.

Well, things did not go as planned. You know that feeling you get when you take a test and you’re so sure that you got an A that you tell everyone how well you did, but when you get it back you find out that you actually got a D? That was Monday, kind of.

In this case, it was more like the teacher couldn’t grade the test because he couldn’t read your handwriting.

Here’s a snapshot of the comments:

Matt, your site is crawling

I have to say that if this weren’t HN, your site was posted, and it timed out, I would never come back again. I mean no disrespect; I really like your idea, but going to a site these days and timing out makes the company look like it runs out of a cardboard box with a 2400 baud modem.How many people trust their credit card information to a site with slow response times.

DreamHost? Uh oh… Methinks you aren’t getting 1/1,000th of the hits you could be. DreamHost = dog slow with two concurrent connections to a fairly light rails app, in my experience. Please tell me you’re running a VPS with them or something. You might want to consider switching hosts before you lose anymore of these hits.

The “guest” links don’t work.

For me, I tried to follow the link for domain names sorted by pronouncability and they did not work. I assume this is what OP meant.link href:

http://www.domainpigeon.com/domains?length=6&sort_by=pro…

error:

Oops! That Page Doesn’t Exist

You may have mistyped the address or the page may have moved.

Please contact support@domainpigeon.com if you have any questions or concerns.

Just clicked the register link [https://www.domainpigeon.com/users/register], got “Internal server error” with a load of info and a backtrace, which I imagine you don’t want to display.

Ugh. Domain Pigeon wasn’t loading for a lot of people and for those that were lucky enough to get through, it was about as slow as me picking up country line dancing.

A few hours later I received the following email, which really drives home the impact it was having:

I hate it when that happens…

I had come back from PayPal and clicked the complete order link, when I
got the Passenger error screen.  (Probably a security hole, too.)

Did I actually pay or not?

Thanks

The worst part was that it was probably avoidable. At first I blamed the problems on the heavy traffic and using a shared host, but I later discovered that the likely culprit was a 238 MB log file which I had been remiss in clearing.

In total I got 2 sign ups on Monday and 1 on Tuesday, just enough ($120) to cover the costs of the Mechanical Turk experiment ($106).

I can’t help but think about how many it would have been if the site had worked well.