Tuesday, October 23, 2007
F#
F#, Microsoft's functional programming language, will be fully integrated into Visual Studio. Ars Technica has a nice description of what F# is and the advantages of functional programming.
Labels:
functional programming,
microsoft,
programming
Firefox 2.0.0.8 unstable on Windows
If you have not installed the Firefox 2.0.0.8 update on Windows, wait for a bit. It keeps crashing. Mozilla recognizes the problem and plans an update soon.
Labels:
firefox
Monday, October 22, 2007
Digitizing content and the Open Content Alliance
The NY Times published an article today about how several large libraries have rebuffed offers from Google to scan their collections, aiming for broader distribution of their content by preferring to go with the Open Content Alliance. The OCA has recently reached an agreement with the Boston Public Library to scan content, and will be scanning 136,000 books from the Library of Congress. This content will be indexed and made available through the OCA website and Yahoo!
It is laudable that several institutions are trying to make their content available to the widest possible audience. They are standing on principle, disagreeing to cede complete control to a private entity.
It is laudable that several institutions are trying to make their content available to the widest possible audience. They are standing on principle, disagreeing to cede complete control to a private entity.
Labels:
digitization,
libraries
Wednesday, October 10, 2007
Google, IBM and cloud computing research
Google and IBM recently announced setup of a data center for access to students to work on cloud computing. They expect to setup 1,600 processors by the end of the year. This should be a fantastic resource for students to build and run some large applications.
Google has published a bunch of research that has gone into building their applications. In particular, MapReduce - a programming model based on functional programming - allows programs to be automatically parrallelized. According to Google, "a typical MapReduce computation processes many terabytes of data on thousands of machines."
Google has published a bunch of research that has gone into building their applications. In particular, MapReduce - a programming model based on functional programming - allows programs to be automatically parrallelized. According to Google, "a typical MapReduce computation processes many terabytes of data on thousands of machines."
Labels:
google,
high-performance,
scalability
Subscribe to:
Posts (Atom)