- From: Patrick Gundlach <gundlach@speedata.de>
- Date: Sat, 9 Feb 2013 16:23:49 +0100
- To: "public-ppl@w3.org" <public-ppl@w3.org>
Hi Dave, > Very interesting Patrick. Now I see where you are coming from. > > The idea of a 'dynamic layout language' is new to me. > Is this the prime area of overlap in terms of formatter feedback? Yes. I think only the dynamic page layout needs formatter feedback. All other areas can be handled by static rules (don't break page/words/table at this place, use this as a table head, place image at the top left, ...). > It would seem that your separation of layout instruction XML > and the source XML is not very great? I.e. they are tied very > closely together? They are only separated until both are read from the renderer (in contrast to XSLT, where they are coupled before they reach the renderer). Then they are highly coupled, except for some declarative instructions. Sample structure of the layout file <Layout> <DefineColor...> <LoadFontfile ...> (master page definitions) <Record element="some_element_name_in_the_data" > (instructions what should be done with the data) <ProcessNode select="child_elements" /> </Record> <Record element="some_element_name_in_the_data" > (instructions what should be done with the data) <ProcessNode select="child_elements" /> </Record> .... </Layout> So the Layout instructions are as closely coupled to the data as in XSLT case. They (except for some trivial global definitions) are application specific. (My slides are online: http://speedata.github.com/publisher/temp/patrickgundlach-speedata.pdf - the place will change) Patrick speedata UG (haftungsbeschränkt) ------------------------------------- Telefon 030/57705055 Mobil 0178/1967142 Mail gundlach@speedata.de Web www.speedata.de Eisenacher Straße 101 10781 Berlin ------------------------------------- Amtsgericht Charlottenburg HRB 135360 B Geschäftsführer: Patrick Gundlach USt-IdNr: DE278023065
Received on Saturday, 9 February 2013 15:24:17 UTC