I am in the process of trying to create a Cocoon "baby-block" that can be used to generate output from a DocBook source file. I was hoping that the group could help me puzzle through the "right way" to accomplish my goal.
The process that I am envisioning looks something like this... 1. The user visits a web page where they specify the DocBook source file they would like to upload and the various processing options they would like to use. These options are presented as pulldowns, text fields, checkboxes, etc. 2. The uploaded file into a temporary directory somewhere 3. The DocBook stylesheets are applied to the file, using the specified processing options 4. (For now) a PDF document is returned to the user My primary implementation goal is to use the minimum amount of code as possible. Right now the W3, and in my opinion the Cocoon developers, are pushing for a more declarative web development experience (check out the XForms spec for an example) so I would like my implementation to adhere to this vision. Accomplishing part 1 is fairly straightforward, as it is a simple XHTML page, no bells or whistles. I could base it on a transform, but for now that seems like overkill. Part 2 looks to be be handled by the SourceUploadAction that is currently sitting in the scratchpad. The alternative that is most often discussed in this list is to use an XSP page to do the work, which I would prefer not to do for the reasons discussed earlier. Any other techniques would be greatly appreciated. Part 3 looks to be the most complicated. I don't want the user to have to always select all of the processing options. I also don't want to have to keep track of the default values used by the DocBook stylesheets, which would be a maintenance nightmare. My plan is to provide a "use default" option for the pulldowns and to assume that empty text fields mean "use default". Another problem is that the DocBook stylesheets assume that there will be a "driver" stylesheet that looks like: <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"; version='1.0'> <xsl:import href="file:///Users/jafoster/Development/docbook-xsl-1.50.0/fo/docbook. xsl"/> <xsl:param name="use.extensions"/> <xsl:param name="autotoc.label.separator"/> </xsl:stylesheet> I would really like to do things "properly" and use the driver stylesheet approach. This means that I need to first use the submitted form to create the driver stylesheet and then use that new stylesheet to process the uploaded file. At a guess what I will have to do is use two different pipelines. The first will use some kind of "FileWritingTransformer" or "WriteableSource" or something similar to generate the driver stylesheet. That pipeline will then call the second pipeline which will actually do the transformation and generate the PDF. Hopefully someone can provide some advice on this part of the implementation. One issue that I have no idea how to address is the namespacing issues associated with having one XSLT generate another XSLT. Assuming that I can get this to work, part 4 is automatic, assuming that I use Saxon. As of today Xalan can't handle DocBook, so I am using the "specify your own factory" approach and forcing the use of Saxon. I have every intention of documenting the full design process and implementation of this project, assuming that I can ever make it work. I know that the DocBook community would love to have an automated way for new users to play with the DocBook stylesheets in a controlled manner. Thanks for any help you can give. Once this is all documented, you (hopefully) won't have to answer these kind of questions again! Jason Foster --------------------------------------------------------------------- Please check that your question has not already been answered in the FAQ before posting. <http://xml.apache.org/cocoon/faq/index.html> To unsubscribe, e-mail: <[EMAIL PROTECTED]> For additional commands, e-mail: <[EMAIL PROTECTED]>
