Projects

harpoon-web-publisher

First of all, Harpoon is not yet released and will not be until late 2007 or even 2008. There is as of yet no code here. Harpoon is currently used to publish the Harper's Magazine website. The code running in production needs to be refactored and the site needs to be redesigned, and as it is ... [More] refactored it will be open-sourced here. The goal of Harpoon is to lower the cost of creating new narrative forms on the web--for instance, to make it easier: to convert summaries into timelines, to add news popups to a Google map, or to turn a blog into a book. Harpoon is structured as a set of Saxon XSLT 2.0 templates, functions, and extensions (i.e. plugins), that run on top of Jetty or Tomcat. In the case of Harper's Magazine, the content managed by Harpoon includes blogs, monthly magazine content, 250,000 page images, 350,000 pdfs, 10,000 or so timeline "facts" and "events," running on one okay machine. Everything is cached, slowly. It took a week to cache the millions of pages that make up the Harper's website at first, but about two minutes per day of CPU time thereafter to cache new content. Rather than seeking faster page-loading time, several-megabyte query result data structures are held in memory to generate hundreds of interrelated pages that are then written to disk. Content from everywhere. Blogs and magazine articles from the 1850 all have to live together. Using RDF provides flexibility; just drop data into the database and style it using XSLT. This proves to be incredibly useful with some up-front planning. Only one part of a complete breakfast. Harpoon generates views of data and writes those views to disk. That's all. Subscriber/user management, advertising management, site administration, design, content entry--these happen elsewhere. Framework? Library? Who cares. Basically, this code is something that works for us. I'm GPLing it in the full expectation that it will be ignored entirely or rewritten in five lines of Perl by a 14-year-old. Fine by me either way. PiecesTomcat or Jetty Saxon 8.X.X The FXSL functional programming framework (possibly) Batik Solr Patience [Less]

PHP - XML_XSLT2Processor

XML_XSLT2Processor is a PHP class that allows you to perform XSLT 2.0 transformations within PHP by using SAXON or AltovaXML. All with the ease of the PHP XSL extension. Documentation is available from the project's Wiki.