Re: Opinions and experiences with ReportBro

Hi Marco,

Over the weekend I considered the issues you raised regarding your current
reporting setup. In summary:

The tools you're using for designing reports (MS Office, Open Office, Libre
Office) provide nice WYSIWYG interfaces. But they don't adequately support
Freemarker template language. For example, there's really no environment
for Freemarker intellisense prompting, previewing, testing, or debugging
erroneous Freemarker syntax in Office products. There's no intellisense
prompting for the XML data models that users may be attempting to
reference. Perhaps hierarchical XML data models haven't yet been designed
prior to users needing to reference data values in Freemarker statements.
Lacking such interfaces and tooling can be a hindrance to template design.

The Java based xDocReport utility may provide basic support transforming
Office documents to say PDF files. But it too often fails. When it fails,
the debugging tools are inadequate, which shouldn't be surprising given
that xDocReport is trying to interpret the ponderous, horribly verbose ,
horribly documented XML that gets generated by Office products. Plus
xDocReport is binding to other Java APIs and engines such as Freemarker,
which the developers of xDocReport may not understand well.

I imagine that xDocReport failures might also be attributed to a disconnect
between the oftimes hierarchical and complex XML data models that are
generated by your RPG developers vs. the Freemarker references that are
generated by your users.

I imagine that some failures might be attributed to your Java daemon
that wrapps xDocReport and Freemearker engines.

It appears to me that there are altogether too many potential failure
points in the process and essentially no way to fix them other than perhaps
tediously iterative trial and error. Moreover, the errors might be
attributed to users or RPG programmers or Java Programmers, and who between
the three (3) might be responsible?

Nathan.