The two real artifacts from this project are the XSLT style sheets in directory src/main/resources/xslt :

sortOnyxVariables.xsl
sortRefinedMetadata.xsl

At present they get embedded in a jar file, and have to be extracted.

The main allied java program is XMLSortTest.java, which is a series of unit tests aimed
at flexing the style sheets. Look particularly at the file config.properties when setting up.
(All of this can be improved upon if you want to try).

These style sheets are required for a proper semantic compare of metadata coming out of
Onyx (sortOnyxVariables) or going into i2b2 (sortRefinedMetadata). At present the only
project that makes use of these is the onyx-procedures project.

Believe me, sorting of elements as achieved here is required if you wish to compare, 
say, two copies of the RefinedMetadata that go into building an i2b2 ontology.
A change in ontology is a big thing, and needs to be checked whenever data is loaded.
Detection of inadvertent minor (or major) change is important: ontologies otherwise
can fail silently. Each subsequent time you load data, you should compare the ontology
you use against the previous version, or be absolutely certain no change has occurred. 
This is more especially so for nominal ontologies derived from a questionnaire,
where "changed" data can be loaded over time, as different respondents use the questionnaire. 

===> I'll say that again: ONTOLOGIES CAN FAIL SILENTLY. <===

See the allied project compare-xmlfiles, which can undertake a meaningful compare 
once the metadata has been sorted. All of this is orchestrated by scripts within 
the onyx-procedures project.

DURING DEVELOPMENT:

(1) Ensure you have the Eclipse library variable M2_REPO set to point to your local maven repository.
(2) Remember to edit the resources/config.properties file