[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

Re: [xsl] Building Dynamic width Cals Table

Subject: Re: [xsl] Building Dynamic width Cals Table
From: Wendell Piez <wapiez@xxxxxxxxxxxxxxx>
Date: Tue, 6 Nov 2012 17:03:34 -0500

Hi DT,

You should probably go with Option 1.

This isn't to say that Option 2 might not be a good one for particular
systems with particular types or families of tables. But unless your
publishing system is so highly constrained (and always will be) that
you know you can manage, over the lifetime of your system, designing
and developing new semantic descriptions for every new kind of
information that you wish to present in tabular form -- this is true
for some kinds of publishing, but not most -- it isn't a general

If you don't trust your staff to encode your tables, or don't want to
burden them, then let them encode the tables in minimal form, without
optimizing them for display. Then someone else (a specialist) can do
that later, possibly with machine assistance at least for the simpler

But the reason table formats exist in XML (and have made their way
even into general-purpose semantic document description formats such
as Docbook, TEI and NISO JATS) is that even if the world of possible
information types has not been shown to be infinite, it is large
enough to be practically so.

Tables show where the fractal edges between controlled and
uncontrolled information types become evident. They, like most
presentational encoding, are an effort to extend some minimal control
(at least control of presentation) into areas where stronger semantic
control is not practical or cost effective.

The doctrine of separation of presentation from content -- as a
corollary of a deeper doctrine about regularization and automation --
is an important one to understand. But it's a mistake to take it to
mean that all presentation of documentary data is, or should be,
automatable, or that semantic description of content (with the burden
it imposes of building a bridge from semantics to their application)
is always practical.

On the other hand, if you know you'll be making the same kind of table
to show the same kind of data over and over again (or discover you
are), you can look at abstracting this (and just this) information
type into some richer semantic description, and converting back down
for presentation. At that point, your implementation questions become
relevant again ... and their answer depends on the particular
circumstances and needs you face.

Good luck,

On Tue, Nov 6, 2012 at 1:24 PM, Mailing Lists Mail <daktapaal@xxxxxxxxx> wrote:
> Hi All,
> I have a question regarding the building dynamic width cals table.
> We have an editorial system, and a publishing system (this creates
> intermediate XMLs, which will be then translated to HTML through
> XSLTs.) We have two options of how to carry out our design editorial
> DTD.
> Option1
> ======
> Have Cals Table in the Editorial DTD model, so editors will be able to
> define their content in a table , and adjust the widths of the columns
> etc,, This gets into the publishing system and the tables are
> translated to HTML and we are fine.
> Option 2
> ======
> We decide against having the tables in the DTD, because we would want to
>   a. Not have presentation specific information in the XML
>   b. Have the XML purely with the Semantic information...
> Option1, is straight forward and there is no real thinking about it..
> I have doubts about the Option2..
> 1. Is it correct design to *have* the presentation specific markup (
> Tables etc..) in the XML.. There are people arguing for it and against
> it.. So would like to know from the elite members of this group, what
> they thiink about it..
> 2. If we go ahead with *not* having the tables. We face issues with
> the table column width. So the stylesheets creating intermediate XMLs
> should be creating cals table, and  should be able to work out the
> correct width based on the content of the column. Is there anybody out
> there who thinks it is possible to do this conversion without having
> too many issues..
> 3. If we ignore "evaluating" the column width, then can we fix the
> column width to some intelligent guess on a fixed length and assume
> the HTMLs to automatically wrap the content, when the content length
> is more than the length of column.. so if I have <table width="700"
> style="table-layout:fixed> or some such thing, I should hope that the
> HTML should be able to autowrap, without me doing anything to fix the
> table width?
> Your help is very much appreciated in advance
> DT

Wendell Piez | http://www.wendellpiez.com
XML | XSLT | electronic publishing
Eat Your Vegetables

Current Thread