Hello Jena Community,

I write regarding the recently retired 'jena-maven-tools' sub-project.  It 
provided
a maven plugin used to cleanly execute the 'jena-schemagen' tool during a maven 
build.
I found this tool very useful for generating .java constants from domain 
ontologies.
Example:  I edit a file like GoodDomainOnto_owl2.ttl in my (maven) project 
source tree.  Then I configure the jena-maven-tools plugin to run during my 
maven build,
to cleanly generate updated java constants (in an output .java file).  These 
constants
are in turn used during later stages of the build.  Very tidy!    

But there is trouble in the wind.  I summarize what I have read about the 
apparent 
demise of this maven plugin in items 1A, 1B, 1C below.  Then in item #2 I 
describe 
my own situation, listing some possible courses of action in 2A, 2B, 2C.  

1A) There have been two mentions of 'jena-maven-tools' on the jena-users list 
in the last two years:
https://lists.apache.org/[email protected]:lte=2y:jena-maven-tools

1B) One of those mentions is this announcment from Andy S., on 2018/12/14 
entitled 
"Retiring Jena modules", where he said "we are looking at the status of:  
jena-maven-tools - command line schemagen is in jena-cmds." 

Obviously this was my BEST chance to speak up, but I tarried too long:

https://markmail.org/message/v6i3d5ktqx5j6teb

1C) The other mention is the 3.13.0 release announcement on 2019/09/29, 
which included this fearsome notice:

JENA-1760: Retire jena-maven-tools    
https://issues.apache.org/jira/browse/JENA-1760

https://markmail.org/message/tidlnyfivbr4xumj

==========

2) Currently I use jena-mvn-tools at version 3.6.0 (due for upgrade!) despite 
several bugs in 
the maven plugin implementation.  I have used the jena-maven-tools in several 
past and current 
projects.  I have a pragmatic interest in seeing jena-maven-tools continued and 
improved, or 
replaced with something better.  Upon request I am happy to report what I know 
about the extant 
bugs (in the jena-maven-tools plugin).  But as noted in 1A,1B,1C above, the 
jena-maven-tools 
are currently on their way out of the main jena codebase.  So perhaps I will 
wind up fixing these 
bugs myself, in a fork, or...?  

Option 2A) I formally report the bugs, and we collaborate to get them fixed, 
resulting in an 
eventual pull request that re-establishes an improved version of 
jena-maven-tools.

Option 2B)  If the jena team prefers to be rid of this troublesome plugin, it 
may be 
best for me to publish a new artifact from one of my own projects.   I could 
fork the
code from jena-maven-tools, and eventually publish it somewhere.  I am guessing 
it would
take 6+ months, depending on how much support + interest we get from the 
community.   
As illustration of how I think about this tool, here is an example where we 
used a previous 
version of jena-maven-tools (0.8) together with 'rdf-reactor' (which is also 
falling into disuse by ... 
everyone but me?).  

https://app.assembla.com/spaces/cogchar/subversion/source/HEAD/trunk/maven/org.cogchar.lib.onto/pom.xml

Option 2C)  Perhaps there is some better way to organize build integration for 
jena-schemagen 
(or something similar) which I am not aware of.  I do prefer to remain in the 
maven orbit, but 
if someone has a really snazzy config for some other build environment (like 
sbt or gradle or ...?), 
then I will be happy to hear about it, and maybe I can find a way to switch 
some of my builds to 
use it.  (Not looking for a general discussion about build tools; just the 
specific experience with
invoking jena-schemagen, or comparable alternatives, from these build 
environments).  

I am eager to hear any comments and suggestions from the jena-users community.

peace,

Stu

Reply via email to