[Tm4j-developers] TM4J 0.7.0 Features (Long)

[Kal A. <ka...@te...>]

Hi,

Now that 0.6.0 has been released, I am starting work on collecting the 
features list for 0.7.0. Below is a list of some of the features I have in 
mind. I would like to hear from anyone that has suggestions for additional 
features. I would like to make the release cycle for 0.7.0 much shorter 
than we had for 0.6.0, but that does mean that , without a lot of help, not 
all of these features will make it in to the release. What I am proposing 
to do is to collect together a list of the features proposed and create a 
survey on SourceForge which we can all vote in and then try and partition 
the voted on list into 0.7.0 and post-0.7.0 features.

Here are the features I have so far:

Locator resolution interfaces [REQUIRED]
   The current Locator interface provides no way to retrieve the data at 
the end of the locator.
   This is *required* in order to be able to properly support the mergeMap 
feature (see below).
   This would probably take the form of a separate LocatorResolver 
interface which would return
   a data stream for a given Locator. The architecture for this would be 
flexible and support
   plug-in extensibility for different locator notations. This feature 
would include the following
   sub-features:
   o  Pluggable URI resolution implementation
   o    http: URI resolution
   o    file: URI resolution
   o  Catalog-based locator resolution implementation

ID round-tripping [REQUIRED]
   Support the direct mapping of -id- attribute values in the input XTM 
file into the value of the
   ID property of the Java objects created by the XTMBuilder. This would 
enable complete
   round-tripping of -id- attribute values in most trivial cases. When 
loading multiple XTM files
   into a single TopicMap object, some additional attribute processing may 
be required to
   prevent DuplicateObjectIDExceptions getting thrown (alternatively, this 
could just happen...)

Merge-map support [REQUIRED]
   This is the last major piece of XTM support required for TM4J.

External topic reference processing [REQUIRED]
   This is the *other* last major piece of XTM support required for TM4J ;-)

More flexible indexing architecture
   Split the indexes away from the TopicMap object - this would make it 
easier to interface
   with persistent storage systems which provide an alternative means of 
query (e.g. Ozone
   or an RDBMS). It would also significantly simplify the TopicMap 
implementations.
   Elements of this feature include:
     o  Basic index interface
     o  In-memory Index implementations
     o  Ozone Index implementations

Rework DOM implementation to work on XML-DB (http://www.xmldb.org/index.html)
   XML-DB is a standard API for XML repositories, which is based on DOM and
   XPath. The work already started on the DOM implementation should be pretty
   straightforward to port to XML-DB. The original intention of building a 
TM4J interface
   on an in-memory DOM structure does not need to be abandoned - the two
   could probably be supported by a single interface.

JDBC/RDBMS implementation
   Apparently some people still like relational systems...

New TMNav architecture
   This would include:
   o  TMNav browser implementation
   o  Touchgraph visualisation
   o  TMNav editor implementation

Improve consistency of use of Log4J
   o  Ensure only Log4J is used for logging warnings/errors
   o  Ensure command line apps and other apps initialise Log4J correctly
   o  Create default Log4J initialisation object

Support for ISO Topic Map Syntax
   o  Support at least import of topic maps in ISO syntax (with the
       HyTime links replaced with URI references.

Dynamic web publishing application (based on Cocoon)
   o  Cocoon generator for topic maps
   o  Some XSLT templates for viewing topics
   o   Additional utility classes for extracting and sorting topic names,
       occurrences associations etc.

Static/dynamic web publishing application (based on Velocity)
   o Utility classes for extracting and sorting topic names, occurrences,
      associations etc.
   o Configurable framework for using Velocity templates to view topic map
      objects
    o Wrapper application to allow deployment in a servlet container
    o Wrapper application to allow invocation of the processing from the 
command line

Support Canonical XTM output
   This could form the basis for conformance testing.
   See http://www.ontopia.net/topicmaps/materials/cxtm.html for a description
   of Canonical XTM format

Support LTM input
   This is a simple text format for topic maps (without the XML verbosity)
   See http://www.ontopia.net/download/ltm.html for a description of LTM.

Please note, this is not a closed list ! If you have other suggestions, 
feel free to add them. And just because something is not initially 
suggested or not voted in for release 0.7.0 does not necessarily mean that 
it cannot be included in that release - as long as some one is willing to 
take on the work.

I look forward to hearing your thoughts!

Cheers,

Kal