Hacker News Reader

Comment by gojomo | Hacker News Reader

"recommends users switch to dominant URL shortener bit.ly"

Often old links can't be changed – easily or even at all. (Tweets can't be edited, for example.)

URL redirection is such a lightweight service that Yahoo/Bread should commit to either redirect indefinitely or ship the mappings and domain off to someone else who will.

View on HN

Normally, I'd disagree, but part of starting a URL shortening company would seem to be a commitment to keep it working in the future, partially because its so easy and partially because URLs shouldn't die.

Charitably, we'll say they have a billion URLs. Each mapping is maybe 8bytes->256bytes. So they have a max of 264GB of data. That would fit on a heroku postgres dev plan, and the code to run it would take a single engineer half a day to put it on heroku.

They could also just join 301Works[0], who seem to some degree affiliated with the Internet Archive and will keep on running the redirect if they are given technical control of the domain.

[0] http://archive.org/details/301works

While I agree this is not a huge engineering endeavor and Yahoo could do the Internet a solid by making this work, our free dev databases are limited to 10k rows (and, as the name suggests, intended for development). Please don't put a billion in them.

Fair. I was just trying to find a thing they could stick this on for basically nothing. The $49 plan would have been fine too.

FYI, I didn't notice this limitation on the pricing page.

Oof, you're right--sorry about that. The addons page [1] has the right pricing, but our own pricing page is incorrect. I'll make sure we get this updated. Thanks for pointing it out!

[1]: https://addons.heroku.com/heroku-postgresql

No, a URL shortening company was a commitment to breaking the internet by adding redirection. The URLs should die to discourage the others.

still not zero cost

Sure, but with 6 engineers it would be the polite thing to do.

Frankly, I hope they don't, for the simple reason that we need some high profile link breakage before people learn to treat URL shorteners accordingly (that is: avoid them whenever possible if they're outside your own control).

It's not the right solution, but URLTeam is stepping in to archive the links[0]. They estimate 121M links, and it'll be a shame that there's no easy way for a user to automatically look up the link from URLTeam when bre.ad is gone.

[0] http://archiveteam.org/?title=URLTeam

Reply on news.ycombinator.com