#trunk.ly

Trunk.ly is AMAZING! You can have all of your links from all of your social media sites linked back to this site. 

They even provide a code for a search bar to be added into your blogs so that anyone following you can search all of your links from that search engine.

It's brilliant and I highly recommend it to anyone who blogs or just sends a lot of links from various sites.

Trunk.ly is a great substitute for Delicious, with all the sad news about Delicious it is time to move onto a fun way to curate all your links in one place. Trunk.ly is very easy to use and with all the links we share on Twitter and other social networks Trunk.ly is very useful. It simply collects all the links you share and saves them to your account, then you can search and do what you what. We've had a wonderful chance to speak to Tim Bull about the past and future of Trunk.ly, plus we also get to know him a bit better. This interview with Tim Bull, Co-Founder of Trunk.ly is brilliant. Keep reading on!

Co-Founder: Tim Bull

Link: http://trunk.ly

We like to know a bit about the founders here are Startupli.st and I’m sure our readers do as well. Can you tell us a bit about yourself?

Alex (Co-Founder) and I met around 18 months ago at a startup camp in Sydney. My background was with PwC where I was a global IT architect, while Alex had previously co-founded and exited a startup in China.  The startup camp was a great opportunity to work through some ideas and get to know others looking for partners.  

As we both came from outside Sydney (I'm from Melbourne, Alex was traveling) we shared a backpackers room together.  After 2 weeks of 16 hour days, sleeping in the same room, sharing the same toilet and working on the same project we knew we could work well together.

We formed BinaryPlex (our company) in November 2009 and we've been working full-time together ever since, Alex based out of Shanghai and me in Melbourne.

What other things did you build/work on before Trunk.ly?

We've iterated through several ideas on the road to Trunk.ly, the most mature of these (in that it is launched and acquiring some paying customers) is http://tribalytic.com, a Twitter based market research tool.

We also started with an Enterprise Expertise location engine, briefly launched a tool to find people on Twitter based on what they talk about (using the alpha of the Enterprise Engine), and just prior to Trunk.ly, launched an event specific Twitter client called Distlr (which is now closed).

Although they seem unrelated, each idea has had a core thread to it – how do you make sense of the digital trails people leave behind every day and turn that into something useful.

Trunk.ly seems like a very unique idea, what gave you this idea to build it in the first place?

I don't think there was one single thing, but a range of factors came into play at the same time back in October.  One was our own observations about the way in which people use Twitter – we could see it was becoming a very heavy link engine.  Then there was the general commentary and observations about Delicious – you could tell from the stats that the service was gradually dying.  

We wrapped several of these things together and hypothesised that maybe the reason tools like Delicious were becoming less popular is because the main stream had moved on. If your motivation for recording a link was to share it, why also bookmark it (which would be two clicks)?  So we asked ourself the question “what would this look like if it was automatic?”.

I think the other core problem we saw was our own frustration with just trying to find links we'd shared a few days after we'd shared them – you constantly end up trolling back through peoples timelines to find that nugget you saw. Unlike many of our other ideas in which we could intellectually see there was a market, but perhaps didn't actually need ourselves, Trunk.ly was something we knew we wanted to use.

We wrapped up these ideas together and tested them with a few people and got some really positive feedback and it's kind of grown from there.

How long did it take you to build Trunk.ly?

We had the idea for Trunk.ly in October, so we've got to where we are since then. That's not been full time, we did a little bit of contract work along the way. While there is a lot that's unique about Trunk.ly, there are also some components we reused from previous projects – for example, we've got our own inhouse Twitter library which we can pretty much plug and play with that handles Oauth and what I call “production ready” Twitter downloading.

It looks like Trunk.ly is doing really well, were there any difficult problems to get to where you are and how did you over come these problems?

There are always issues with any new service and Trunk.ly is no exception. Perhaps the biggest issues have been that we never actually intended to launch when we did.  We had this mental plan that we'd launch sometime in mid January 2011, early February. Then all of a sudden, the news that Delicious was thinking of shutting down leaked from Yahoo and suddenly it felt like everyone was looking for a service like Trunk.ly.  

The opportunity to launch then and there was just too good, so we threw the doors open, even though we weren't close to ready at all.  In hindsight this was a good decision, but at the time we got hammered. Architecturally we had lots of issues where code wasn't scaling as well as we'd like and all sorts of things were going wrong. We overcame these with a combination of raw processing power and hard work.

As an example of raw processing power – at one stage we had 10 Large AWS EC2 instances spun up just processing links coming in. To put this in perspective, now we've done more effort on refactoring and optimising our code, a single 4 core machine can handle significantly more load now than we could then even four weeks ago with all those machines.

And hard work? Well that's nothing new for a startup. Since the 23rd of December until this last weekend, it's been 16 hour days with pretty much Christmas day the only real exception.

What is the architecture behind Trunk.ly and how hard is it maintain a service like Trunk.ly?

We use Python / Django for the web site itself and most of the core processing is done in Python. As far as databases, we use a tiny bit of MySQL (which Django requires), but most of the actual data is in MongoDB.  We also use REDIS as the heart of our distributed queuing and messaging system.

The site itself as users see it is very stable and quite scalable now, so there's not much on that front, the issue is the backend pipeline. We've invested a lot of time into making this more and more robust and there's still more to go.  When you're connecting as we are to lots of different services, there are all sorts of issues you have to deal with, a long with just the simple crawling issue – there's lots of unicode challenges, badly formed web pages and so forth.  

I think the best thing now is it's at a point where when we wake up in the morning, we expect it to be up and running smoothly still – that wasn't always the case in the first week or two.

It must take more than one person to run a service like this, is there anyone working with you or is this a one man startup?

There's two of us, full-time. We've been really fortunate as well that there are others out there who also passionately believe in Trunk.ly as much as we do and they've been very generous with their time and skills, everything from helping out with some of the graphic design to writing connectors into other systems, to producing a library for our public API.

If you had to describe Trunk.ly in 3 words, what would they be?

Never forget links

Are there any future plans for Trunk.ly? Anything awesome you can share with our readers?

Lots of future plans! I think in the short term it's to stabilise where we are at – there are lots of niggly things which need polishing, so we'll get those completed. We've rushed a lot of the coding, so there's a bit of refactoring as well. It feels like we've reached a bit of a plateau at the moment where we can take a breath and catch up with ourselves.

We've got a new design in the works which will won't significantly change things, but will tidy up the UX a bit, make things clearer and improve the fonts which are currently problematic across browsers and OS.

We need to look at how we start making some money from the site.

More in the medium term, a lot more on how to make the social experience more powerful – what are the “hot links”, how do you find people that share the type of links you're after and so forth.

How will Trunk.ly be monetised in the future? 

It's not fully decided, but we there'll almost certainly be some form of advertising and some premium features.

Last but not least, what is your favourite startup excluding Trunk.ly?

There are lots, but I can't go past a local Melbourne startup http://adioso.com because I just love the way they've tackled such a difficult domain (natural language travel search) and created this “wow” experience where once you've used it, you wish every flight search worked as intuitively as theirs.

FROM STARTUPLIST