Bowerbird@aol.com wrote:
however, a less-complex subset -- called t.e.i.-lite -- is available, and that is what i recommend...
You do have a curious way of avoiding capital letters... :-) After al my objections against XML and TEI, you may wonder why I still recommend to use TEI lite is that it forms a very decent base to start some structurial tagging with -- you don't need the full 1400 pages of TEI to get started with it, and you also don't need to reinvent the wheel, and come up with some alternative, equally simple scheme. Doing this has the added benefit that for those text that require it, you can easily step up, and work with the full set, if so required or desired. If you're just against using angled brackets, they are simple to use and understand by both humans and computers. You can do more fancy tricks to make marked-up texts look more like plain text, but attempts to do so, both by TeX or SGML add considerable complexity to the reader -- both machine _and_ human. XML has one thing for it, and that is its simplicity (and that some people build complicated things on it, such as namespaces, XSLT, etc., that require a course in computer-science could be quite hidden from most users.) You can ofcourse object out of principle against something 1400 pages thick, but that is unavoidable, given the complexity and wide diversity of books that have been published in the 500+ years since Gutenberg's invention. Since much of the difficult stuff of XML will eventually be hidden from users. Future versions of layout programs will probably be able to read a thing coded in TEI directly (doing an XSLT transform to some internal format), and format it nicely according to some defaults. You can then apply all the required formatting tweaks to it, export to some nice lay-out format (XSL-FO, maybe, PDF, or who knows), and safe all your nice tweaks, linked to your original TEI, so you have best of both worlds. I already have numerous benefits from working in XML, in that I can generate nice HTML files (that often need no touch-up at all) and reasonable plain ASCII for PG, but also have spelling checking on a per language base, extract all fragments in a certain language, create tables of contents, etc. on the fly, extract dublin core bibliographic records, and more. Jeroen.