--- Ross Gardler <[EMAIL PROTECTED]> wrote:
...
> There is a preprocessing step to fix related links and > linking etc.
This could be a concern, although we can do
preprocessing too. Can you explain a little more about what this preproccesing
step is and what it does.
From the DITA Toolkit docs:
"A pre-process is done before the main transformation. The input of Pre-process is dita files and the output of Pre-process is also dita files. But the output is in temp directory. Pre-process is the basic for the main transformation, it handles several different processing before the main transformation. Without pre-process, dita topics and map can still be transformed into different outputs, but the features in pre-process such as resolving conref attribute are not available."
So the dita files get transformed to new dita files in a temp directory, and those are the dita files that get processed by the rest of the xsl files into html, pdf, etc. The preprocess xsl files are located in xsl/preprocess/ in the toolkit. The ant build.xml file is used to automatically call the preprocessing and then the general processing using the correct xsl files for each output.
As long as all this preprocessing is XSL then this will be just another step in our pipeline, so no worries there.
...
First off, a few questions:
- what stylesheets are used to create the HTML output at present - what are the processing steps to create HTML at present - can you make available some of the output of the existing process
The stylesheets for html are called in the ditatargets.xml file (called by build.xml)...and ant is something that is relatively new to me...However it does call the general processing xsl file for html, which is xsl/dita2html.xsl (which I believe also calls the xhtml processing file, dita2xhtml.xsl because they use similar templates). That xsl file calls the rest of the xsl files it needs, all located in xsl/xslhtml/. Before it gets to any of that, it uses one or more of the preprocess xsl files. I am guessing this process makes duplicating it with Forrest a mighty task...
Actually, it sounds really easy. I'll look at the build file since you say Ant is new to you.
I have included an html file created from this
process.
OK, these all looks nice and simple (so far)
If you'd like the entire set of dita files for derby, as well as instructions on building them, go to:
http://incubator.apache.org/derby/manuals/dita.html
So what you have the dita documentation already embedded in a Forrest site. I assume you are doing this by copying the HTML generated files over to the site and having some static links in site.xml to create your navigation menu;s.
In another mail Jean T. Anderson said that you don't want to generate the docs from the DITA files every time you generate the site (we do the same thing here in Forrest, we have our site, and we have our documentation.
How do you want the integration to change? What is it about your current way of working that you don't like? How do envisage Forrest improving this situation?
Whilst you are thinking about that I'll be having a look at your build system.
Ross
