CCEL XML A Match Made In Ethereal Heaven

CCEL & XML A Match Made In (Ethereal) Heaven

The CCEL • The Christian Classics Ethereal Library w Online library of electronic texts

The CCEL • The Christian Classics Ethereal Library w Online library of electronic texts w Different applications require different formats

The CCEL • The Christian Classics Ethereal Library w Online library of electronic texts w Different applications require different formats § Which format is authoritative?

The CCEL • The Christian Classics Ethereal Library w Online library of electronic texts w Different applications require different formats § Which format is authoritative? • Th. ML (Theological Markup Language) w XML application w Master version of most (someday all) texts on ccel. org

Th. ML • Defined as an XML application several years ago • Perl scripts and other “hacks” used to transform Th. ML into HTML (and other formats) for presentation on site w Static transformation w Error prone w Difficult to maintain

Th. ML • Defined as an XML application several years ago • Perl scripts and other “hacks” used to transform Th. ML into HTML (and other formats) for presentation on site w Static transformation w Error prone w Difficult to maintain • Isn’t there a better way?

XSLT • XML Stylesheet Language Transformations w Stylesheets written in XML that “explain” to a piece of software called a “transformer” how to change an XML document into a different format (XML or other) w Example: Client-side transformations with Internet Explorer 6

XSLT on the CCEL • Summer of 2001 w Two students (myself and Jimmy Osborn) wrote XSLT stylesheets that duplicated the functionality of the Perl script transformations § From user’s perspective, the XSLT-transformed docs look no different from the Perl-transformed docs § From developer’s perspective, there is now only one file for every text in the library (the Th. ML file), all other formats are generated on-the-fly at the user’s request

XSLT on the CCEL • Summer of 2001 w Two students (myself and Jimmy Osborn) wrote XSLT stylesheets that duplicated the functionality of the Perl script transformations § From user’s perspective, the XSLT-transformed docs look no different from the Perl-transformed docs § From developer’s perspective, there is now only one file for every text in the library (the Th. ML file), all other formats are generated onthe-fly at the user’s request • How does this work? w Client-side transformations are not powerful enough for our purposes w We needed a more powerful, server-side transformer

• Apache Cocoon w Java webapp (webapps are servlets with an attitude) w Runs inside a Java servlet engine (e. g. Apache Tomcat), which is then connected to a web server w Not just a transformer, but an “XML Publishing Framework” § Brings together of host of different XML technologies for the purpose of publishing XML docs online § Does XSLT transformations, XSL-FO to PDF conversion, renders SVG graphics, executes e. Xtensible Server Pages (XSP), integrates with databases, washes your dog and more

• Sitemap w The sitemap is where Cocoon really shows that it is more than just a kludge of XML-processing programs w Using regexp matching (or simple wildcard, if you’re into that sort of thing…), it maps URLs to an XML document, a stylesheet, and a transformer (and a few other things I won’t get into here) § Based on the URL, you can match any XML source doc with any XSLT stylesheet and then send the result through any transformer you like!

CCEL Sitemap

CCEL Sitemap • New Problem: People browsing online don’t want the whole text in one big HTML file w Cocoon’s sitemap helps us here too…

CCEL Sitemap • New Problem: People browsing online don’t want the whole text in one big HTML file w Cocoon’s sitemap helps us here too…

Current Status • Still in “beta” stage w This new system still has a few bugs, but it will be going online in a preliminary form very soon

Current Status • Still in “beta” stage w This new system still has a few bugs, but it will be going online in a preliminary form very soon w As more and more documents are converted to well-formed Th. ML, they can be used by Cocoon and we will hopefully phase out the old system over time § There are SO many documents on the CCEL, this may take some time

Current Status • New Server w Our new Dell Power. Edge 2500 server runs Cocoon (being a Java app, its not the most efficient user of system resources)

Current Status • New Server w Our new Dell Power. Edge 2500 server runs Cocoon (being a Java app, its not the most efficient user of system resources) w We are currently in the process of moving the entire site over to the Power. Edge § When this is complete, the new XML-based system will be available for people to try out

Current Status • New Server w Our new Dell Power. Edge 2500 server runs Cocoon (being a Java app, its not the most efficient user of system resources) w We are currently in the process of moving the entire site over to the Power. Edge § When this is complete, the new XML-based system will be available for people to try out w Some examples from the new site…