Subject: Re: XSL Requirements (was: Microsoft extensions to XSL) From: "Oren Ben-Kiki" <oren@xxxxxxxxxxxxx> Date: Wed, 18 Nov 1998 14:40:20 +0200 |
Guy_Murphy@xxxxxxxxxx wrote: >Hi. > >Being new to the list (although I have followed the archives) I did find >myself wondering at the objection to persuing transformation within a style >language, but assumed I was missing something. It would appear that I'm not >the only one seeing transformation as integral to style formating. > >Surely at a fundamental level, styling raw text is applying a >transformation to that text, the question merely being to what extent this >is extrapolated, and how complex it becomes. If I wish to render a simple >series of characters it multiple columns, this requires significant >transformation of the structure, but is most definately a style formating >issue. This turns out not to be the case :-) Let us take a simpler formatting issue: controlling the font of the text. It is obvious you need some language in which to express the notion of text font. You might have a choice of several languages. Now, formatting text into columns must depend on the font (since that controls character width). It is therefore expressed in the same language used to express fonts (and colors, and border types, and backgrounds, and positioning, and ...). On the other hand, reordering text (say, sorting a list) is a structural transformation which does not require any special output language. So does pruning some parts of the data, as in creating a summary report. Generating TOC and indices is also independent of formatting .Of course, the generated TOC and index needs to be formatted, but the data in it is independent of such formatting. >Maybe I am still missing something, but with my initial dabbling with XSL I >cannot see how transformation can be regarded as less than vital to style >application. A great many things are vital to for style application. Specification of syntax, resource locators, character encoding... This does not mean they are part of the specification of a stylesheet language. Instead we rely on XML, URIs, Unicode... Each standard deals with one thing. The question is "are style application and transformation _the same thing_?" - and I feel the answer is "no". If it was the same thing, we wouldn't have two camps with regard to XSL features, target audiance, and so on. >It would initialy seem to me that arbitary transformation ie., that which >is "hard-coded" within a flow object, or through the application of inline >CSS to well formed HTML, is often regarded as style, and self specified >transformation as some how "other". I personaly can't see the distinction, >at root we're still transforming text/data at the end of the day. A transformation is from some "thing" to another "thing". CSS (as it stands today) does not transform HTML into a well defined something else. It just annotates HTML entities with a well defined set of attributes. XSL, on the other hand, both specifies how to transform an input XML tree into output text, _and_ how to annotate it with (a similar) set of attributes. Maybe I'm a die-hard UNIX user but I prefer two smaller tools, each doing one thing well, then a large one which compromises both. Adding formatting attributes to HTML in a stylesheet instead of in the tags is a welcome extension, and CSS does it well. Transforming an XML tree to something else is an industry need and XSL may do it well - given it goes off the stylesheet language track. >If I am amiss, I'm sure it'll be pointed out to me :) Any time :-) Oren Ben-Kiki XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
RE: XSL Requirements (was: Microsof, Guy_Murphy | Thread | RE: XSL Requirements (was: Microsof, Lawton, Scott |
RE: XSL Requirements (was: Microsof, Guy_Murphy | Date | RE: XSL Requirements (was: Microsof, Lawton, Scott |
Month |