So, how does Peer work?

You may be wondering, behind the sleek black curves and dark beauty of Peer, what it's actually doing behind the scenes. In our FAQ entry about how Peer works we describe a rough process that starts with parceling data up, and sending it out to the cloud - but what does that actually mean?

As we surf the web a record of our internet history is kept by our browser. At the very least the browser stores the url, title and time of each page you visit. Some browsers such as Firefox 3 or Chrome actually store quite a bit more information; "links" between pages showing the path you take through the web, or the number of times you revisit a page.

Peer uses this history information from the browser and shares it within groups. First of all, let's be clear: Peer doesn't share secure (https) sites. We also do our best to avoid webmail and social network messaging by default, and support blacklists to try and catch all the rest that you don't want to get out.

Roughly every 10 to 15 minutes, for every active group you have Peer checks to see if you've visited enough new sites and produces a snapshot of your browsing history. What's in a snapshot? Really basic stuff - a list of url's and titles of pages you've visited, plus counts of the number of times you visited each page and the amount of time you spent there. That's all. Since this information may still be sensitive to you and the group, Peer then compresses and encrypts the snapshot and uploads it to the group's snapshot holding area on S3.

To get your groups browsing history Peer periodically polls for new snapshots from other members. This lazy sync happens in the background and keeps you up to date with the browsing history of your groups.

So that's Peer history sharing in a nutshell. We think it works fairly well but it can definitely be improved in terms of stability and efficiency. Help us improve it by testing Peer out! As always any feedback is welcome.

Explaining Peer

So why did we decide to build Peer?

We were struck by how much work we do when we filter through search results from the web, looking for the pages that will satisfy our information need.

If you think about your close network - we're talking the team you work in, or your close friends, you're very likely to share considerable interests with them - because your team all works in property in London, or clean tech in LA. Or, broadly, you enjoy similar activities socially- cycling, or going to Italian restaurants.

So, armed with some previous projects we'd worked on which explored this idea, we set out to build a way to let people share the sites that they find on the web, with the minimum possible effort, and to make these searchable by people in their network.

We feel like we've got to the stage where Peer is making a real contribution to the way that we search the web on a daily basis. We're releasing it into the wild to get feedback from others on how it works for them. We're not pitching ourselves as a Google killer, because what we're doing is quite different. We're trying to get the best possible results for a narrower set of searches - the ones that your network will know about. So if I do a search about the best places to fish in Sydney, then a web search engine will most likely give me a better result. (I live in London and don't know anyone who fishes). However, searching for information about OpenCoffee (an entrepreneurs weekly meet in London) brings up some more interesting information, including the blog post that kicked off the OpenCoffee movement as the top result.

We think that with some more hard work, the range of queries that we return great socially aware results for will increase, as will the accuracy - we're not there yet by any means. But we really need your help - please let us know where you find Peer to be useful, how many people you reckon is the optimal size of your groups, and other features that you think would be useful. We've already incorporated Twitter and Delicious links from your network, and plan to add Yammer very soon, but please continue to let us know what you think!

Our favourite (desktop) apps

We've spent quite a long time building Peer, and over the course of that time, we've come across some absolutely brilliant software which has shaped how we've developed Peer. I thought that it would be good to say a little about the best, so that you can see where we've got some of our inspiration from:

So, in no particular order:

Tweetdeck

If you use Twitter, you really should try Tweetdeck. It's a cross platform, Adobe AIR twitter client, which allows you to arrange the people you follow into columns. This means that you can prioritise people and make sure that you catch all of their tweets, whilst being able to scan the rest of your twitter stream.



What we've learnt: groups are now very much necessary to handle the amount of information on the web. And that black looks good.

Digsby

Digsby is a social network and instant messaging aggregator for Windows. We've used it to keep up to date with LinkedIn, Twitter (before Tweetdeck), as well as email - it's definitely the best GMail notifier I've ever used.



What we've learnt: Digsby is written in Python - the same programming language as Peer. It gave us a great idea of how stable a desktop application written in Python can be.

Dropbox

Dropbox is an excellent syncronisation system - you simply install it onto the computers you use, and it appears as a regular file system folder. However, all of the files are syncronised across the different computers, and also available on the internet.



What we've learnt: Another great app written in Python, in my opinion the best thing about Dropbox is the seamless integration with the operating system - there is no fancy AIR uploader - you just paste the files you want to sync into the Dropbox folder.

Skype

We've been lucky enough to work with some of the team at Skype, and they've been really helpful. Skype itself has allowed us to keep in touch when we've been at different places around the globe, and until Peer was up and running was a great way to share links around a small team.



What we've learnt: Amazingly good branding of the product and the app. Skype 4.0 for Windows also evolved hugely between Beta and the final release, showing the need to be flexible to the feedback the community was giving.

--

We've come across some other great apps, both on the desktop and on the web, and hopefully I'll follow up with some more in another post.

Gearing up for launch

The Peer team is working flat out to bring you a fully fledged version of Peer as soon as possible. In the meantime, you can keep up to date with our progress at twitter.com/getpeer.