= Future development MPDL =

== Software selection: frontend (web GUI) and backend system (datastore, indexing, query system) ==
We examine different software systems by means of their main functions and features for 
our user requirements in the area of web based access to XML-documents. We limit the 
examination to four systems: Magnolia, mediaWiki, eXist and eSciDoc.  
Functions which are not mandadory for our requirements are marked italic. 

=== Basic needs ===

|| ||Magnolia||mediaWiki||eXist||eSciDoc||
||Scalable||++||++||-||++||
||Stable||++||++||-||+||
||Performant||++||++||++||+||
||Customizable||++||++||++||++||
||In use at MPIWG||++||++||++||+||
||Free software||+||++||++||++||

=== Datastore ===

|| ||Magnolia||mediaWiki||eXist||eSciDoc||
||User management||++||++||-||+||
||Version control system||++||++||-||+||
||Index / Query system||++||++||++||++||
||Many document formats (xml, pdf, doc, html)||++||++||+||++||
||Discussions / Blogs||++||++||-||-||


=== Extensions / Development ===

|| ||Magnolia||mediaWiki||eXist||eSciDoc||
||Predefined extensions / templates||++||++||-||-||
||Programming||++ ^[#note1 1]^||++ ^[#note2 2]^||++ ^[#note3 3]^||+ ^[#note4 4]^||
||Index / Query system||++||++||++||++||
||Many document formats (xml, pdf, doc, html)||++||++||+||++||
||Notes / Annotations||-||-||-||-||

Notes: [[br]]
[=#note1 1]: Freemarker, Java Server Pages, XSL/CSS, limited XPath (works also in JCR 2.0 ?)[[br]]
[=#note2 2]: php, Java over JavaBridge [[br]]
[=#note3 3]: Java, XQuery / XPath, XSL / CSS [[br]]
[=#note4 4]: Java [[br]]

=== MPDL software ===
|| ||Magnolia||mediaWiki||eXist||eSciDoc||
||Get XML page fragment||-||-||++||-||
||Language technology||-||-||++||-||
||Document upload||++||++||- ^[#note5 5]^||+ ^[#note6 6]^||
||Web viewer||+ ^[#note6 6]^||+ ^[#note6 6]^||++||+ ^[#note6 6]^||

Notes: [[br]]
[=#note5 5]: already implemented but not stable enough [[br]]
[=#note6 6]: has to be implemented (relative easy) [[br]]


=== Magnolia ===
* URL: [http://www.magnolia-cms.com/ http://www.magnolia-cms.com/]
* Advantages
  * scalable
  * stable
  * performant
  * is used at MPIWG: see [http://hartmutkern.mpiwg-berlin.mpg.de/ here]
  * user management included (incl. LDAP)
  * version control system with easy usable history functions (commercial EE version)
  * notes/annotations: blog comments for each web page
  * many installations and users worldwide
  * many extensions / templates
  * programming: Freemarker, Java Server Pages, XSL/CSS, limited XPath (works also in JCR 2.0 ?)
  * fulltext search contained (Lucene)
  * multimedia support (image, sound, video)
  * free software (CE version)
  * MPDL software
    * document upload: stable and mostly included
* Disadvantages
  * user interface: all web pages have to be implemented
  * XQuery support has to be implemented in Java (relative easy)
  * MPDL software
    * get XML page fragment: has to be implemented in JSP (relative easy)
    * language technology: has to be integrated by JSP's (relative easy)
    * viewer: has to be implemented by templates and JSP/XSL/HTML (relative easy) 

=== mediaWiki ===
* URL: [http://www.mediawiki.org/wiki/MediaWiki http://www.mediawiki.org/wiki/MediaWiki]
* Advantages
  * scalable
  * stable
  * performant
  * is used at MPIWG: see for example this [https://it-dev.mpiwg-berlin.mpg.de/tracs/mpdl-project-content TRAC]
  * user management included (incl. LDAP)
  * version control system with easy usable history functions (not free EE version)
  * notes/annotations: discussion of each web page possible
  * many installations and users worldwide
  * user interface
    * already included (Wiki web pages)
    * simple (e.g. for data uploads)
    * user know the Wiki system
    * no extra web server
  * many extensions / templates
    * footnotes: [http://meta.wikimedia.org/wiki/Help:Footnotes internal], [http://www.mediawiki.org/wiki/Extension:Cite Cite]
    * [http://wikisource.org presentation of old books]
  * programming: php, Java over JavaBridge 
  * fulltext search contained (Lucene)
  * multimedia support (image, sound, video)
  * free software
  * MPDL software
    * document upload: stable and mostly included
* Disadvantages
  * XML/XSL/CSS support has to be implemented in Java (relative easy)
  * XQuery support has to be implemented in Java (relative easy)
  * MPDL software
    * get page fragment in XML document: has to be implemented in Java (relative easy)
    * language technology: has to be integrated by Java programs (relative easy)
    * Viewer: Zope programs have to be implemented in Wiki/Java (relative easy) 

=== eXist + Zope ===
* URL: [http://exist.sourceforge.net/ http://exist.sourceforge.net/]
* Advantages
  * performant
  * XML/XSL support
  * is used at MPIWG: see [http://mpdl-proto.mpiwg-berlin.mpg.de/mpdl/query.xql MPDL system]
  * some installations and users worldwide
  * XQuery/XPath support
  * programming: Java, XQuery/XPath, XSL/CSS
  * fulltext search contained (Lucene)
  * MPDL software
    * get page fragment in XML document: already implemented
    * viewer: available as Zope programs on another web server
    * document upload: already implemented
  * free software
* Disadvantages
  * not scalable for many documents (> 1000 XML documents)
  * not really stable
    * many system crashes at startup time (up to 2 days)
    * sometimes system crash when a document is uploaded
  * no usable version control system included, no history functions
  * notes/annotations: has to be implemented
  * user management not usable / has to be implemented
  * user interface: all web pages have to be implemented 
  * not many extensions / templates
  * MPDL software
    * document upload: not stable through eXist

=== eSciDoc ===
* URL: [https://www.escidoc.org/ https://www.escidoc.org/]
* Advantages
  * scalable
  * is in test use at MPIWG: see [http://pubman.mpiwg-berlin.mpg.de/pubman/ Pubman]
  * user management included (incl. LDAP)
  * version control system included
  * XML/XSL support
  * programming: Java
  * fulltext search contained (Lucene)
  * free software
* Disadvantages
  * not in use by many users
  * performance of many queries not tested
  * very complicated: much work for administration and development (API etc.)
  * version control system: no easy usable history functions
  * notes/annotations: has to be implemented
  * not many users, installations and developers
  * user interface: all web pages have to be implemented 
  * not many extensions / templates
  * XQuery support has to be implemented in Java (relative easy)
  * MPDL software
    * get page fragment in XML document: has to be implemented in Java (relative easy)
    * language technology: has to be integrated by Java programs (relative easy)
    * viewer: has to be implemented in Java/XSL/HTML (relative easy) 
    * document upload: has to be implemented (relative easy)