• Print

Four short links: 2 Apr 2009

Predictions, PDF, source code control, and recommendation engines:

  1. Wrong Tomorrow — track pundits predictions and see how accurate they really are. From the ever-awesome Maciej Ceglowski.
  2. PDFMinerUnlike other PDF-related tools, it allows to obtain the exact location of texts in a page, as well as other layout information such as font size or font name, which could be useful for analyzing the document. It also infers text running within a page by using clustering technique. Entirely written in Python.
  3. Migrating from svn to a Distributed VCS — to decide which distributed VCS to use, Brett Cannon gathered Python use cases and then showed how they’d be done with different dvcses. The result is a very useful comparison document for svn, bzr, git, and hg.
  4. Online Monoculture and the End of the Niche — interesting post summarising and explaining research into recommendation engines, drawing the conclusion that although Internet World recommendation engines show everybody lots of new stuff, we’re all seeing the same new stuff and the end result is less the “riches of niches” Long Tail fantasy and more a monoculture.
tags: , , , , ,
  • AP

    As to number 4, beware of simulations without real data. That’s not science, it’s fantasy. It’s weather forecasts without the weather stations. The exact opposite has been argued — that people pursue more an more niche interests and points of view on line. Where is the data?