[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: having good success with xinclude's and building modular manuals
based on the advice i've been getting here regarding using xinclude's to create modular manuals, things are going well -- i've got a small, modular manual being pulled together nicely from modules in various subdirectories. i haven't finished the entire processing sequence, but it's coming along. so, for now, some observations and a couple of questions. first, since i'm writing the modules in my pidgin docbook, i don't even bother adding a DOCTYPE to those modules. not much point since there's no DTD for that stuff. what i do is pull together the modules in a single large document, *then* run it through a stylesheet to generate valid docbook, which i can validate later. i'm assuming that there's no immediate drawback to this; in short, i'm using "xsltproc --xinclude" simply to build a large single file out of the module files. thoughts? also, since the module files themselves aren't validated, i can invent a new tag for those modules, so each of those files has a "module" document element, which is how i refer to them in the actual xinclude's. i realize i had the flexibility to call them "section"s, but since those modules could potentially end up as chapters, that seemed possibly misleading. calling them "module"s seemed suitably generic and, again, it's doable since i'm not doing any validation at that level, and the <module> tag will be transformed to the appropriate <chapter> or <section> tag long before validation, based on its inclusion in the document. finally, generating the single, large pidgin docbook file creates a file with the top-level <book> element with an "xmlns:xi" attribute (due to the use of xinclude), as well as numerous lower-level elements with the "xml:base" attribute. given that i'm interested only in treating the resulting file as one large text file from this point on, is there any value in keeping those attributes? i could trivially strip them out in the first processing step, as i don't see what i would use them for later, at least at the moment. (I'm not doing any linking between the modules, at least not yet.) any thoughts on the above, given how egregiously i'm violating most of the rules of proper processing? :-) i realize i could do things more properly if i defined a DTD for my pidgin grammar, but at the moment, there doesn't seem to be a lot of value in that. rday
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]