W3C home > Mailing lists > Public > www-style@w3.org > June 2007

Re: Stylings only possible with Tables

From: James Elmore <James.Elmore@cox.net>
Date: Sat, 30 Jun 2007 16:36:55 -0700
Message-ID: <4686E917.1050204@cox.net>
To: www-style@w3.org

David Woolley wrote:
> James Elmore wrote:
>> I know that it's not their job, or business, to produce CSS from PDF 
>> files. But are there any CSS group members working for Adobe who could 
>> 'lean' on their bosses to allow CSS capture the way they allow HTML? 

Should I (or anyone) change the subject, since I seem to have wandered off topic 
from Table stylings to 'capturing styles from other sources'? I find this topic 
interesting, but don't want to burden others if there is no interest.

> The HTML is recoverable because it has been explicitly added in the 
> tagging process (e.g. by copying the structural markup from a source 
> MSWord document).  If you have to generate HTML from untagged PDF, it is 
> likely to be a presentational approximation.

I thought that this was what was said before -- that it is possible to 
explicitly add HTML during document creation. My question was sortof "can 
something like this 'explicit markup' be added to PDF for styles?"

> The style isn't recoverable, because there is no indirection of the 
> styling information, and because its theoretically possible for a single 
> structural element to have multiple styles.  Ignoring the second 
> problem, you could recover styles, but only at the level of id based 
> selectors for every element.

Sorry that this is further off topic, but this brings up another question I have 
considered in the past -- some of the HTML generation tools (at least those from 
Microsoft, that I know of) place style in every element. Mostly, the styles are 
the same. In the past, I have collected this style information (by hand and eye) 
and created a style sheet which captured the basic styles and the occasional 
variations. By creating style selectors for, say, paragraphs, I can list the 
basic information, which hardly ever changes. Then, with .classes or #ids, I can 
create styles for the (usually minor) variations.

I have used unix tools (sed, sort, grep, and the like) to program scripts which 
capture repeated information and produce a unified set of that information, 
without repeats, and I would like to try something like that with styles.

I just re-read that last sentence and almost puked. So unclear! Let me try 
again. -- Has anyone had any experience with capturing style information which 
is assigned to each element and automatically collating it to reduce redundancy 
and create a single set of styles for the document? This would produce a set of 
style rules for the document and eliminate (or almost eliminate) the styles for 
each element.

Such a tool could be useful, if only to demonstrate how powerful CSS can be and 
how to use it correctly to simplify document design. I personally would like to 
have such a tool when I have to work on web sites which have been 'designed' 
using Microsoft tools. It would save me hours of work compiling styles and 
deleting them from each element.

Are there any programmers out there who have used a 'pretty printer' to clean up 
code which doesn't follow the coding rules? This could be similar -- a pretty 
styler to clean up the styling rules.

Is there any experience out there with such a tool? Has someone already written 
one? Is someone willing to start one? Am I the only one who thinks this is 

If there is interest, I would like to start a new thread -- something along the 
lines of 'Capturing Styles from Other Sources'

> PDF is a final form document and only the revisable forms would have 
> style sheets.

But other document types could contain styles which might be mechanically 
collectable. At least mechanically produced HTML (as mentioned above) could be a 
source for such capture tools. Thoughts, people?

James Elmore
22162 Windward Way
Lake Forest, CA 92630
Home	(949) 830-9534
Email	James.Elmore@cox.net
Received on Saturday, 30 June 2007 23:37:14 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:29 UTC