Dropsafe

by Alec Muffett

a demonstration that bit.ly is not here to help you

I don’t like URL shorteners, or rather:

  1. I don’t like the siloed URL shorteners that outsource the problem to a third party
  2. I don’t like the SEO bullshit that led to (most) long URLs in the first place

Following Leyland’s Law, I feel that a document should have either a single (“Uniform”) Resource Locator, or an infinite number of them for different users; this is one of several reasons why pymine will give you a URL shortener of your own.

Bitly exists to interpose itself between you and the web, and to track you. It is there to spy on you – on behalf of whomever – and track references in a way that DoubleClick never dreamed would be possible – that people would *offer* themselves to the service in exchange for undoing the_ridiculously_long_slug_based_URLs_that_SEO_suggests.html and in doing so provide extraordinary analytics information to an otherwise uninvolved intermediary.

The damage is immense and undoable – shorteners are now a fact of life, but they don’t have to be siloed, which is another reason why pymine will give you a URL shortener of your own.

If you want evidence that bitly’s not there to help you, witness:

bitly
Source

Long URL length: 15
Short URL length: 20

Bitly – or in this case, Tweetdeck calling Bitly – lengthened the URL by 5 characters in pursuit of registering it for tracking and analytics capture; certainly it wasn’t registering it for shortening, else Bitly would have given-up and returned the original, unchanged.

The whole thing’s a hideous mess. I look forward to making it worse, so that it can get batter.

<irony>Let a million shorteners blossom…</irony>

—
[1] Leyland’s law: in software you should make provision to supply zero, one, or an infinite number of any given resource. Cited to me by Paul Leyland, original source unknown.

⊞

Comments

6 responses to “a demonstration that bit.ly is not here to help you”

  1. Björn

    The problem is that URL shorteners ARE solving a problem of the social web. If you run your own website, you can easily track who clicked on the links you published (with Javascript, I suppose). But if you contribute to somebody else’s social network, you can’t track who clicks the links you publish.

    So it is not only about shortness, and Twitter is not the only thing to be faulted for it.

    Maybe if it becomes standard for “web2.0” sites to offer their users statistics about the links they published, the need for URL shorteners would go away. But that would also have to include things like comment sections in wordpress blogs.

    Reply
  2. alecm

    Björn – if you’re so happy about being tracked, kindly post here the means by which you arrived at my blog posting, and the last 20 or so URLs that you visited.

    We’d be obliged, so that we can send you advertising.

    Thanks.

    Reply
  3. giedrius majauskas

    Well, I would see different problems.
    1. Tweetdeck not calculating which link is shorter ( tweetdeck problem, not bit.ly). In some cases twitter clients do not shorten links they do not have to shorten. Overall, blame tweetdeck.
    2. Bit.ly showing tracking information for public ( You can see tracking info of any link in bit.ly no matter if you have shortened it or not).
    There are people that need to track clicks and there are people who dont need. Primary purpose of bit.ly is droping links in emails and twitter, where space is limited. (In emails because long activation links might be broken into several lines by client).
    Overall, the best answer is to use different URL shortener than bit.ly. You can do this in seesmic, I think you can do it in tweetdeck as well.
    Problem solved 😉

    Reply
  4. James Overton

    OK, we accept that this is a problem. Writing a URL shortener should not be hard for someone, so why has a clean one not been written and deployed? Could it be that doing it will only cost the developer money and earache as lusers complain when it is down or slow?

    People need money to run these services; we need both a clean (non tracking) way to shorten urls, but also the developer needs a built in way to make money running it. Donations are not going to cut it by the way.

    Also, I looked at pymine which i had not heard of before this post; I cant use it to shorten URLS, so why even link to it in this context?

    You have outlined a serious problem. Whilst you cant be expected to solve everything… I dunno, I just want the tool to fix it!

    Reply
  5. alecm

    @james

    i suggest we all re-read your comment and imagine that it applied to “blogs”:

    >People need money to run [blogsites]; we need both a clean … way to blog, but also the developer needs a built in way to make money running it. Donations are not going to cut it by the way.

    Funny how WordPress survives with a semi-commercial, build-your-own-for-free model and pulls through revenue.

    What stops people running their own shorteners? Nothing.

    Reply
  6. alecm

    oh, as for pymine, 1) i am the lead engineer and 2) it is / shall be usable to shorten URLs.

    Hence.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

More posts