dyncon 2011

Peter Svensson. Peters conferences are a bit smaller and a bit sharper than the average conference. Only one track available due to the conference size but when the line up is as brilliant as at this conference it doesn't matter at all. Compare it with Robert Virding continued about Erlang where Joe Armstrong stopped the day before. Interesting about the design principles behind the language.
I swapped Björn Eiderbäcks talk about smaltalk for a nice walk since I have already heard him talk about this before. After Björn there was a grails talk by Henk Jurriens. A nice framework indeed but I drifted away due to it not being news to me. (Happy with rails for the moment!)
Then Michiel de Jong - creator of unhosted - a reaction to the centralization of data that has been going on in the web 2.0 internet world. We gladly give away our data to companies like facebook, google and twitter without having any real control over it at all. This is a bit contrary to early internet standards email being the most notable where data is distributed and I decide for myself which server I trust with my data. Michiel argued that data and applications should be kept separately with data stored on an unhosted server. Unhosted is building support for this making it possible to write applications that store data in a place decided by the user. A future nicety with this is of course that several applications could work on the same data. While this is a really cool idea it is not something that the normal consumer asks for and it will thus take some time before it becomes the new web standard. What is the killer application for this? I am thinking about the never appearing electronic wallet.
Awesome presentation with live music played by coding in Clojure by Jon Vlachogiannis. With a clojure DSL he demonstrated how to code and play interactively. Read more at mad.emotionull.com.
The conference ended with a presentation about caching and tuning by Wim Godden. Clearly there are lots of things to do in this area when needed. Wim argued that a system should be designed with caching in mind from the beginning. I don't like designing for performance at all. I suspect that the result will be the same when the code is test driven. The key is to have small manageable parts and performance tuning will be made easy whether it is caching or clustering or speed or what have you.
Dynamic languages are really awesome. Too bad there is a lack of skilled developers able to benefit from them. The conference was really great so please - Peter Svensson - do it again next year.

Old comments

2011-03-16Hardy Ferentschik
Sounds awesome.
comments powered by Disqus