Marc, himself, his blogs, and you reading them.
Old wine in... new poems.
Just reformulated some earlier thoughts... lazy me just got tired of trying to figure it out myself, so hoping that bouncing off my vague ideas in the group will at least deepen my own insight. There is some comfort in knowing that I'm not alone on some of this... (if not in detail, then at least in principle)
I'm a bit afraid though that the embryonal state of my thoughts is very much causing me to fall in my own trap of elliptical and near-poetic communication style for which my colleagues like me so much...
In fact there is even more embryo's fighting up under the skull here, so more verses for future poems:
- Can we have an ObjectResolver like we have the SourceResolver? Euh, Do they need to be kept separate?
- Or is this already been called an input-module?
- Where is the fundamental difference between how sitemaps look at subsitemaps, actions and flowscripts? Why the different contracts between them? (Why the sometimes strong feelings for or against one or the other?)
- What is (detached from the notational wearies) so different between the transforming pipes that use xslt with document() function, jxpath, and even xinclude to some extend.
- Why can our sources be xmlized (and thus plugged in just about everywhere), while it requires a specific jxpath-aware pipeline element to extract smaller pieces out of an object-graph.
- And reversely why can we only xmlize the full source, and not extract jxpath like strings out of it to be used in e.g. attribute-values.
- Java objects? You mean the things that weren't serialized to XML yet? (There exists a wealth of standard, loose coupled data-handling techniques that work on pure self-describing XML-data structures)
- XML? Ah, that handy notation to communicate object states with? (Java objects have methods, you can never grasp those dynamics inside XML)
Post a comment

