Trunk.ly picks up the torch of del.icio.us | Hacker News Reader

Trunk.ly picks up the torch of del.icio.us

trunk.ly

98 pointsalexdong42 comments

I’ve been using delicious since 2004. It’s no surprise to see Yahoo abandon delicious finally but it still hurts to see a promising product failing to get the attention/resource it deserves. Back in 2005/06, when tagging and folksnomies were at their peak, I could see delicious as the “central” memory for links.

Fast forward to 2008. I was in foocamp. There was a night session called “Computational Photography” and it was Joshua who hosted it. I went there and found PG, Tim O’Reilly and a couple of other folks interested in AR were there. Joshua wanted to show cool videos on youtube but just couldn't find it. Here is the conversation between Joshua and Tim that has stuck in my mind ever since.

  > J: Forget it. I just can’t find it.
  > J: I wish there is a service I can share my links with others. 
  > T: That’ll be delicious!

3 months ago, @timbull and I started developing trunk.ly. To me, it’s picking up the torch of delicious and delivering a promise of “never forget a link again”. Trunk.ly pulls in all the links you’ve shared on different social networks, twitter, facebook, and of course delicious, and make them full text searchable.

Trunk.ly is still in its invite-only pre-beta stage. It's not as aesthetically attractive as we'd like it to be. It still has minor bugs. But we'll try out best to send out as many invites as possible so that you can pull in all of your delicious bookmarks before Yahoo completely shut it down.

Here are what others have said about trunk.ly:

Fenn Bailey, founder of Adioso, YC09 alumni

  > I honestly thought I'd only need to use it 
  > sporadically ... however within the day of 
  > installing it I was already reaching for 
  > Trunkly to find something I'd said

Rand Fish, founder/CEO of SEOmoz

  > Trunk.ly is a deceptively simple … 
  > I've been surprised at how often I use 
  > it just to find the things I tweeted!

@bootload, fellow entrepreneur

  > with #delicious rotting and rumours of shutting
  > down this makes #trunkly a great #geek #christmas present

Very neat idea. I like the idea of aggregating all the breadcrumbs one leaves across the internet. I just dropped my email for an invite, it looks like a fun service.

Likewise, looks nice.

How close is it to being out of beta? It seems existing in time for people to import their links to your site and not dispersing to different areas of the web is pretty vital for the act of picking up where delicious left off.

You should launch right now.

You are front page on HN and delicious is all over the news. If you have an architecture that will scale any way well at all, get out of public beta, and try surf this wave.

So it's basically historious + packratius?

This is funny. Tim and I woke up this morning thinking about the contract job we've been working on. Then we saw the delicious thing and I wrote this post. Not expecting anything major. But 3 hours later, we're on the HN front page. And our invite queue is jammed with requests.

What the heck. Tim and I just decided to pull out our credit cards, remove the invitation thing and open up the registration.

A big Thank You to the HN community. You've made our day.

We just launched!

I see you get a lot of traffic - cannot authorise my delicious account for the last couple of hours. Is there any way you can distribute the traffic? I guess you wouldn't want to use other people's hosts due to passwords being required, but can you spin up some cloud vps type of hosts to cleanup the queue the next days and import more links?

You don't seem to have any flattr/paypal/whatever link on the website, but I'd be happy to click it, if it meant more accounts migrated ASAP :)

Also, just adding an idea: if you provide a delicious-compatible api endpoint, people could reuse the existing browser plugins - is there any chance you could do that?

We're working on EC2 instance to catch up! We've been flooded. Yes, a delicious-compatible api endpoint is a brilliant idea. Already in our etherpad. Thanks.

On "import links from delicious"

In trunk.ly, you can pull in all your delicious links and notes with only one step. Your tags will be imported as well.

For 99% user, your links will be imported very smoothly. However, trunk.ly is still pre-beta, a fews of things I worth mentioning:

1. If, very unfortunately, you've merged your delicious account with your yahoo accounts, you won't be able to import.

2. Trunk.ly is not https just yet. Yes, we're working on this right now.

3. If you have over 50,000 links, you may have problems importing them. We're using lxml's DOM to process the yahoo message while we really should'of used the SAX model to deal with super large messages.

You can use ElementTree's iterparse() function (also available for lxml) to do incremental parsing while still having XPath-like functionality available for the elements you like (manually calling the .clear() method to release memory).

http://effbot.org/zone/element-iterparse.htm

A few trunk.ly pages:

* mine: http://trunk.ly/alexdong/

* tim's: http://trunk.ly/timbull/

* rand's: http://trunk.ly/randfish/

interesting, my first impression is "hey, it's a linkblog!" rather than "hey, it's a delicious clone!". This isn't necessarily a bad thing, just my observation.

a few things I see that I'd kinda miss from delicious:

1) instead of seeing how many people are linking to something as a number, it instead has an "also shared by:" feature instead. I honestly like just seeing a number of how many people are sharing it more than seeing if a few random people are also sharing something. It also helps me judge whether someone linked a "big news item" or something they were personally digging for info on.

2) multiple tag filters - i.e. search multiple tags for my personal bookmarks. I don't use this much, but it's comfortable that it's there.

> I honestly like just seeing a number of how many people are sharing it more than seeing if a few random people are also sharing something. It also helps me judge whether someone linked a "big news item" or something they were personally digging for info on.

Nodding.

2) multiple tag filters - i.e. search multiple tags for my personal bookmarks. I don't use this much, but it's comfortable that it's there.

I use this frequently enough that I probably won't use trunk.ly without it.

I wonder if Libya will take back the domain name if we save pr0n links.

I wonder if the US will seize their servers if we save Wikileak links ?

alex, very cool.

so I used to use delicious to track others bookmarking and I have a nice group of people that I followed. is trunkly able to import my connections from delicious? this is how I used it mostly and I think that it is a very valuable source of info that I'm losing

hmm, even if we can import the "following" list, I'm not sure how we can invite those people in though. There is no way we can notify those people, that's the challenge.

How cool would integration with Greplin be?

This seems like a really good idea. Ingeniously simple.

We love Greplin! I think we're using similar technologies and facing similar challenges as well. Greplin guys, if you see this, please drop me a message at me AT alexdong DOT com. I'd love to talk to you and see whether we can do something together.

The invite form Submit button on your front page won't work without JavaScript enabled.

I hope you fix that, but more importantly I hope you make sure trunk.ly works correctly without JS enabled. Delicious does, which is very nice.

Note that I'm not some anti-JS gremlin. I've been using the web since 1997 and have only disabled JS in my primary browser (Opera) last month. Wish I'd done it sooner. (I browse JS-heavy pages in Chromium or FF; previously they were reserved for pages that depended on Flash).

I'm also a w3m/vim guy so I appreciate what you feel.

Guess what. Tim and I just decided to open up the registration. Screw the invitation thing. Let's see what's going to happen.

It's still asking for a beta code.

Neat! Do you plan to add the ability to add new links (and add tags to them)? I saw your trunk.ly page (http://trunk.ly/alexdong/), and the first thing that struck me was - besides building a searchable index of all the links that I shared on different services, if you give people the ability to tag/bookmark new links, you might just be the service that Yahoo failed to grow out of Delicious.

That's exactly what historio.us does, no?

A suggestion: When I submit my email address for the invite, the email field does not get cleared, and instead a small message appears underneath saying "Thank you for your interest!" - I completely missed it, and kept trying to submit again and again!

Looking forward to using the product, good job.

I wish you luck! Check out archivd.com and dowser.sf.net; there are lots of good ideas to steal. (I wrote them both, but they are not under active development.)

achivd.com looks really interesting. But it threw back a 500 page: http://demo.archivd.com/list/detail/87UfnvACqpGz5z

yeah -- it's been rotting. :( Here is a "public" list:

https://www.archivd.com/public/detail/dEuLp5fXxS772U

It also has fulltext search, etc. You can sign up for free.

FYI, it keeps crashing Safari on my iPad when I try to visit the home page.

dude, upgrade your iOS. This is a bug in Safari. http://liftinteractive.com/news/2010/sep/22/why-your-site-mi...

Now that iOS 4.2 is out Typekit has changed how we support for iOS devices. We've stopped supporting iOS3 and earlier, which means users that haven't upgraded yet won't get browser crashes.

There's more details at http://blog.typekit.com/2010/12/06/updates-to-typekits-mobil... but to summarize, you should probably republish the kit for your site.

How does it compare with others? Diigo, Stubmleupon, Xmarks...

Trunk.ly is not designed to be yet another bookmark service. It's your personal search engine. It builds a search index just for all the links you've shared.

Why bother tagging each link if you can full text search them?

Try mine at http://trunk.ly/alexdong/ Search for "chinese financial", or "ruby erlang" and you'll see.

I enjoy tagging. Google has automated tagging on his Chinese QA system (they used Lated Dirichilet Allocation algorithm for this), and it is in general usefull when you want to have personal opinion on a link.

Besides that, it seems nice.

I do notice myself tagging more after I start to use trunk.ly daily. And I agree that "tagging" is a great to do attach personal opinions on a link.

Trunk.ly can import delicious tags. We'll soon add the ability for you to add a tag to your link as well.

Having said this, we found most people do not tag, they just search what they want and continue with their life. Based on the existing user behavior data we have, 1% people clicks tag in two weeks. 7-10% uses search at least once.

that's a bit similar idea I got some time ago and even started building, later abandonned it. are you planning on releasing an API?

API is definitely on our roadmap. May I know what are the API methods you're having in mind?

currently the basic ones that would allow to integrate with trunkly - get favorites, add favorites

Just signed up, can't wait for my beta code!

Chrome bookmark imports coming?

Reply on news.ycombinator.com