features for integration project

[Stefan Krieger]

Hello,

I am planing to use babeldoc for a mid-size but mission-critical integration project. (FYI, Babeldoc is competing very well against a BEA Weblogic Integration solution).

Anyway, this will require to build some new components for or based on babeldoc:

+ Integration into J2EE. This means:
Decoupling of stages through JMS, starting of Pipelines by Message Driven Beans

+ FlatFileWriter: Takes the document and a format description to creates a flat file (similar to the existing conversion stage, just the other way round)

+Scheduler: starts pipelines at fixed times. (batch processing)

+ Logging / Journaling to database
+ Some web-based Admin-GUI

Now, I am wondering if any of the above issues is currently in development. Otherwise, we will provide it. But therefore, we probably need active support from the community. I guess, I can count on you ;-)

Best Regards,
Stefan

[Dejan Krsmanovic]

Hi Stefan,
I am very glad you found this project interesting.
Bruce will probably tell you more information you need but here is what I can tell you:

1. JMS
I think you should talk to Erik Klain (one of developers). I think he have worked on MDB pipeline feeder for Bank of New York. We are also planning to implement JmsScanner which should scane JMS queue or topic.

2. FlatFileWriter
Currently does not exist. Sounds interesting!

3. Scheduler
Currently you can schedule scanners using cron syntax. Maybe this could be interesting for you.

4. Logging/Journaling to database
I am not sure what exactly you plan here. There is SqlJournal for writing info into database.

5. Web-based Admin Gui
There is no real Admin application yet. There is application for scanner monitoring and web application for journal inspecting but I have never been too satisfied with them. On of plans we have for the future is to enable JMX support for all Babeldoc components. That would made administering and monitoring much easer and standardized.

I would suggest you to subsribe to our developer mailing list. You can also send these features to our RFE tracker so others can discuss about them.

Dejan

[bruce mcdonald]

Stefan,

Glad to have you aboard.  Please join the developers mailing list and we can address your issues.  The big things that are missing is a GUI frontend.  All the rest are smallish and can be done quickly with help from us.

Bruce,.

[Anonymous]

Starting of Pipelines from an MDB (or any other class), is quite simple.  You will want to look at the Javadoc for com.babeldoc.core.pipeline.command.PipelineFeeder.process( ) ... but here's a code snippet:

Hashtable props = new Hashtable();
byte[] ba = xmlString.getBytes();

try {
   PipelineFeeder.process(pipelineName, ba, props, false, false);
} catch(PipelineException e) {
  throw new MyException(e.toString());
} catch(JournalException e) {
  throw new MyException(e.toString());
}

In this example, <xmlString> represents XML that Babeldoc gets sent from the wrapping program.  Additionally, the Hashtable <props> can contain any "arguments" you want the Babeldoc engine to have access to while running your pipelines.

[Stefan Krieger]

Thanks for the fast replies.
I have a good feeling with babeldoc. Hopefully, the customer does so as well ;-) The decision will be made next week.
Anyway, you are doing a good job.

Stefan