RE: Antw: reports (clarification)

[Rodrigo Moya <rodrigo linuxave net>]

> [Gerhard, Sep22:IMO the populating-step and the rendering-step are very
> different. Therefore we
> have two related, but different XML-file-formats (=DTDs)]
> 
> <GNUe guys, Sep22 >:
> > > I beg to differ here...  I think the output from the report server should
> > > retain the orgainzational tree structure from the template, mainly because
> > > what if I want to take that report and manipulate it a few different ways
> > > (e.g. strip out the header and footer for a particular report, but render it
> > > a different way in another format) this allows us to only have to populate
> > > the report one time with one template, and gives us a way to use XSLT or
> > > jade or something to alter the document in different ways while having a way
> > > to group the data. 
> 
> In case the queries are cached in the report server, it's not further
> complicated to rewrite the template and have it populated with the same 
> data but with a different format (template). So for every different kind
> of 
> layout-engine you let the report engine do the job for 
> you thus saving complexity. I hope I didn't miss something here.?
> 
well, you're right, but the idea behind having a unique DTD (= XML
format) is that there will be less things to implement in the report
engine. If you do 2 XML file formats, we must repeat some code twice,
and as both formats will be very similar, the easiest way for the
populating step is to just fill in the <data> parts in the XML file.

> 
> Btw. Someone might be interested in playing around with diagrams I've 
> produced trying to follow  the report discussion, they can be grabbed at
> http://www.hot.ee/mvali/GDA
> 
I'll have a look at them right now, and if you don't mind, I guess they
could go into the libgda documentation.

cheers