You mean the mapping between twi.bz URLs and the real URLs? Yes, I absolutely plan to release the data. I'm just working on what the best way is, and I wanted to get some feedback.

It's worth knowing that I hacked this together in a two hour session very, very late last night. This is its public debut.

I want to be able to release the mapping and also the trending information.

What format would people prefer?

Please release a separate file without trend information, as not everyone will need it. For the data format: How about a simple CSV file with 3 fields: "domain-part,path,url". This would translate to such entries:

amazon,a, http://www.amazon.com/Geek-Atlas-Places-Science-Technology/d...

If you compress the file (or start a new file every week/month) this shouldn't be much of a burden for your server.

Great to hear that you also plan to release the trending information. For the URL->twi.bz mapping I would opt for a simple text file with a mapping per line.

Always wondered - what value is this database? short of editing the /etc/hosts or browser plugin to redirect from $nowDeadURL/dh36 to $newURL/dh36.

Do people have urls so valuable locked in a tiny url service that'll be lost forever if the service dies? serious question.

Imagine you search for a solution to a very annoying bug in your favorite program. Now imagine you find someone posting on a mailing list with a link to the solution/a patch/something that solves your problem. Imaginary problem solved.

Now imagine the same, except that the link is a tinyurl link and tinyurl went out of business 2 days ago. You will never be able to solve your problem.

If there were some kind of tinyurlarchive.org, you could solve your problem again.