You are here

e! Science - automatisierte Newsseite fuer die Wissenschaft mit Drupal und freier Suchmaschine Sphinx

Michael Imbeault, ein HIV-Forscher an der Université Laval in Quebec, hat eine automatisierte Newsseite und Suche namens e! Science News ins Netz gestellt. Wie er auf drupal.org berichtet, weil er unzufrieden mit den anderen computergenerierten News-Portalen wie Google News oder Yahoo News war

Die Website wird ausschliesslich auf Grundlage von freier GPL lizensierter Software betrieben und weiter entwickelt. Die Basis bilden Drupal und Sphinx, eine Volltext-Suchmaschine.

Sphinx Search LogoSphinx is a full-text search engine, distributed under GPL version 2. Commercial license is also available for embedded use. Generally, it's a standalone search engine, meant to provide fast, size-efficient and relevant fulltext search functions to other applications. Sphinx was specially designed to integrate well with SQL databases and scripting languages. Currently built-in data sources support fetching data either via direct connection to MySQL or PostgreSQL, or using XML pipe mechanism (a pipe to indexer in special XML-based format which Sphinx recognizes). (Abruf 28.6.2008, Andrew Aksyonoff, http://www.sphinxsearch.com/about.html)

Drupal setzt in der Regel auf einem sogenannten LAMP-System aus Gnu/Linux, Apache, MySQL und Php auf. In dem Beitrag auf drupal.org beschreibt Michael Imbeault die Entstehung der Newssite, gibt Hinweise fuer den Selbstbau einer derartigen Website und spricht ueber seine "Lessons learned".

esciencenews Here are a few lessons I learned in no particular order

  • Finding a good domain name is hard and takes time (and / or money) – start early and never stop searching, even if you got one or 2 good ones! You might find something even better.
  • Backup everything often! Especially on Windows.
  • Think simple; more is often less – do only what your users actually need / want as a starting point
  • Don’t be afraid to redo something from scratch if it’s not working right the first time around - I rebuilt critical components of the clustering system days before launch
  • Drupal is a market disrupting tool – it allows a single guy part time to build something great while learning it; imagine what a whole team of professionals can do!
  • I wish I knew about simpletest a year ago; I chased bugs for a long time (and sometimes the same bug that was reoccurring). Things like clustering and regex-based parsing could have been a whole lot easier with appropriate tests.
  • Release early; don’t be afraid to put off minor features to later
  • Keep a todo list through the process, and try to remove items from it as fast as they appear! (this is not as easy as it sounds!)
  • You know you built something great when you visit your own site and find it interesting :)
(http://drupal.org/node/261340)
Themen: 

Neuen Kommentar schreiben