Tuesday, March 22, 2011

How html5-boilerplate helped drastically improving our website performance

http://blog.androidsnippets.com/2011/how-html5-boilerplate-helped-drastically-improving-our-website-performance

And this is where html5-boilerplate really starts to shine. One of it’s best although often overlooked features is the automated build system (with ant and build.xml), which automatically compresses and concatenates stylesheets, javascript files and images, producing a heavily optimized version of the website and all resources.

We just needed minimal adaptations to make it play nicely with AppEngine, in order to provide a base template all the others can extend, and a full integration of the build process (you can find our adaptation as on github).

Posted via email from miner49r

Parallelism is not concurrency

https://existentialtype.wordpress.com/2011/03/17/parallelism-is-not-concurrency/

The effectiveness of the language-based model of parallelism lies entirely in its ability to expose the dependency structure of the computation by not introducing any dependencies that are not forced on us by the nature of the computation itself.  And the key to this is functional programming, which manifests itself here in the transformational approach to computation: sorting is conceived of as a mathematical function that transforms a given sequence into another sequence.  It does not destroy the given sequence any more than adding two numbers destroys those numbers!  Since Quicksort is a mathematical function, we need not worry that execution of qs xsl 

interferes with (depends on) 
qs xsg; we can readily run them in parallel without fear of untoward consequences.  The payoff is that there are many fewer dependencies among the subcomputations, and hence many more opportunities for parallelism that can be exploited, in accord with Brent’s Principle, when scheduling the work onto a parallel fabric.

The upshot of all this is that functional programming is of paramount importance for parallelism.

Posted via email from miner49r

Monday, March 14, 2011

No, really, pi is wrong: The Tau Manifesto by Michael Hartl | Tau Day, 2010

http://tauday.com/

This manifesto is dedicated to one of the most important numbers in mathematics, perhaps the most important: the circle constant relating the circumference of a circle to its linear dimension. For millennia, the circle has been considered the most perfect of shapes, and the circle constant captures the geometry of the circle in a single number. Of course, the traditional choice of circle constant is π—but, as mathematician Bob Palais notes in his delightful article “π Is Wrong!”,1 π is wrong.

Posted via email from miner49r

Sunday, March 13, 2011