Month: November 2007

Mashup Camp fuled by Guiness

Standard

Mashups are fun, Guiness is good, Dublin is cool – combine them all and get Mashup Camp 5 in Dublin. The highlight of the Camp was too meet a lot of smart people and talk tech. Some of the highlights:

  • Chad Dickerson from Yahoo Developer Network talked about web site performance. According to Yahoo’s experience 80-90% of the performance of a web site is in the frontend and only 10-20% in the backend. I must say that that these are suprising numbers, but as the worlds number 1 web destination Yahoo should know. So improving the front end performance has a huge impact and Yahoo are listing 13 rules to help you do so, as well as the Firefox plugin YSlow to help you analyze a sites frontend performance. YSlow is already a critical part of my web toolkit, too bad that it is very depressing seeing YSlow’s reports on my websites, I have a lot of work to do! Chad also mentioned Yahoos GUI design patterns that can come in handy for anybodyworking with web design. His slides are on http://www.slideshare.net/chaddickerson.
  • Performance was also a subject for a later discussion where John Herren and others talked about improving backend performance. The two things that really stuck in my mind was to use a PHP Accelerator to precompile PHP code (instead of doing that on the fly for every request) and to use cookies as a light weight free client side cache.
  • There was an interesting discussion about how to find Mashups, especially in a near future when there are mashups built on mashups and the users are not developers but normal mortals. The Search Engines today help us answer the question “what is”, but searching for mashups or apps is more a question of “how do I”. The options discussed covered everything from ranking mashups based on popularity and meta data to using introspection to automatically figure out what the mashups are doing. It would be interesting to have a system that could track the interactions between APIs and Mashups in a way that a developer could take a mashup apart to reuse just the parts he is interested in. Programmableweb.com is the start of such a system, but it is still only built on the basis of that the mashup developers gives the right meta data about their mashups. Since John Musser from Programmableweb was part of the discussion I am hoping to see some of this implemented there soon 🙂
  • Speedgeeking was as always a part of Mashup Camp, it is basically like speed dating but for demos. 5 minute intervalls to demo your stuff for a new group, and it went on for an hour this time. I built openkapow robots on request and it went pretty well. Building one REST robot to an unknown site every 5 minutes was a risk, but most of them went very well (search openkapow.com for tag “mashupcamp” to find my creations). My Kapow collegue Benjamin demoed his cool Blackberry-LinkedIn-Xing Mashup (also using openkapow). The winner of speedgeeking was the Mashup Camp veteran John Herren and 2nd place as well as the Winner of IBM’s Business Mashup Challenge was Dennis Deery (congrats to you both). Check out one of the most pointless results of the speedgeeking at LOLCatService – the most pointless web service ever.

David Berlind, one of the organisers, have written a good summary of the camp on ZDNet if you want to read more. If you understand Swedish you can also listen to an interview with me from MashupCamp at WhatsNext.se.

All in all it was a great Camp. Not a great number of participants, but the number is less important than the quality of the people. I learned a lot, met great people and came home with my head swirling with ideas. Next Camp is in Mountain View in March and I hope to see you there, if that is too far for you Mashup Camp is also coming back to Ireland next year.

LOLCatService – the most pointless web service ever

Standard

I have just launched the most pointless web service ever constructed – LOLCatService.com. It is a REST Service that allows you to search for LOLCats in an easy way so you can embed this absolutely crucial feature in your web apps and mashups. The very idea of it is so pointless that I just had to implement it.

It all started at Mashup Camp 5 in Dublin where I built openkapow robots on request during speed geeking (basically speed dating for demos). Tom from Yahoo asked me to build a LOLCat REST Service and that was such an utterly pointless and stupid request that it stuck. So this whole service is of course pointless, but it is so pointless that it is cool.

The cool part of the service is that it literally took me a couple of hours to do it all, including writing the openkapow robot and building the site. The most time consuming part was to wait for my domain server changes to propagate.

WhatsNext intervju från MashupCamp (sorry folks, Swedish only)

Standard

Tidigare idag så snackade jag med Tomas Wennström från WhatsNext om vad som pågår på MashupCamp i Dublin just nu. Gå till WhatsNext och lyssna själva eller lyssna direkt här. Det är väl värt at lyssna WhatsNext fär att hålla sig uppdaterad om vad som händer i Sverige vad gäller webben.

Sorry folks, this post is in Swedish only since I am being interviewed by a Swedish podcaster in Swedish. More aboutMashupCamp in English later…

links for 2007-11-11

Standard

Reading the Social Graph – you are defined by your friends

Standard

Ankur Shah and Gi Fernando from the UK web development company Techlightenment had a brilliant presentation at Web 2.0 Expo in Berlin – “Disrupting the Platform: Harnessing social analytics and other musings on the Facebook API“. Techlightenment is the company behind the widely spread Bob Dylan Facebook application, so they know what they are talking about. In these days it is enough to have worked with something for a few months to be an expert, so after 4 months with the Facebook API makes the Techlightenment guys into gurus.

Facebook vs. OpenSocial
The Facebook API was compared to the OpenSocial API. There were quite a few differences, but what caught my attention was that it takes 60 lines of code in OpenSocial what it takes 2 lines in Facebook using the Facebook Markup language (FBML). In general it seems like the OpenSocial API’s are quite imature so far and that Facebook will keep ahead with it’s aggressive constant innovation for quite a while. Still, go for both Facebook and OpenSocial if you are going to do an app. If the social platform war will stay limited to just two standards we are quite well of …

You are defined by your friends
In addition to the information you explicitly have defined, like your age and if you are married or not, there is a lot of information about you that can be deducted from your list of fiends. Are most of your friends from London? If so there is a good chance that so are you. If most of your friends are working in the banking industry then there is a good chance that you are as well. This information is of course worth a lot in the hands of an advertiser that can target the ads to the right people. Now with Facebooks new Social Advertisting initiative that makes it possible for them to target ads to Facebook users even if that user is on another site than Facebook this kind of information is worth even more.

Techlightenment have developed a cool Facebook application called Socialistics that on-the-fly makes best guesses about a person based on their friends. It is a really interesting proof of concept of what can be done. Install it yourself and see what can be deducted about you based on your friends. The more friends the better guesses of course, but it is amazing what can be figured out about you. The power of these kind of statistical analyzis will explode in the near future, which is going to make both Mark Zuckerberg and the Google guys more money.

Who am I according to my friends?
Using the Socialistic app I get the following data about myself. I most likely live either in Stockholm (I did for years) or in Mexico City (not a bad guess since I do spend a lot of time thre), there is a 10% chance I have worked for Microsoft in Copenhagen (wrong) and I most likely took my university degree at Universidad Iberoamericana in Mexico City (I have studied spanish there for all of 1 hour once, but that’s it). The good stuff is that there is a 65% chance that I am a man (just that I work with IT should bring that number up to at least 90%) and that I am from Sweden (guilty as charged)

gender.png

country.png

According to my wife I am a social outcast since I only have 28 friends on Facebook. If the name calling I have to take out a Facebook divorce (ie “cancel relationship”), after all it is 28% chance that I am single already.

cancel.png

married.png