Re: Tables and Charts

Chris Tilbury (C.J.Tilbury@estate.warwick.ac.uk)
Fri, 7 Jul 1995 14:36:24 BST


From: "Chris Tilbury" <C.J.Tilbury@estate.warwick.ac.uk>
To: papresco@calum.csclub.uwaterloo.ca (Paul Prescod),
Date:          Fri, 7 Jul 1995 14:36:24 BST
Subject:       Re: Tables and Charts
Cc: www-html@www10.w3.org
Message-Id: <429606117D@forest.estate.warwick.ac.uk>

On  7 Jul 95 at 13:46, lilley wrote:

> Paul Prescod said:
> > The real problem is not in HTML, it is in the non-existance of a 
> > widely used chart data format. 

I think Paul really meant "non-proprietary widely used" :-)
 
> Actually there is an existing Internet Media Type (MIME type) called
> 
> text/tab-separated-values
> 
> which can have any suitable graphing program as a viewer. I use xmgr

Which, as lilley points out, there is.

Whilst bringing in external standards is one possibility, perhaps 
we need to hark back to one of the "first principles" of HTML, that 
being that it is intended to be a simple markup language. We already 
have an (proposed) element in <TABLE> which is ideally suited to 
storing numeric /data/ - a markup version of 
text/tab-separated-values, if you like - and along with the <TD> and 
<TH> elements and the CLASS attribute (as well as the AXIS and AXES 
attributes in more complex cases) we can sufficiently identify the 
structure of that data such that it can be interpreted by a browser.

I would personally like to avoid a new <CHART SRC="some_url"> 
element, or some application of the <FIG> container, for a number of 
reasons. Chief amongst these are

(i)   {on <CHART>}
      If we can manage it successfully with what is currently 
      available, why add another unnecessary element? The HTML 3 Draft 
      already has many, many tags - of which the content value of some is 
      slightly dubious (<I>, <SMALL>, <BIG>, <B>). We should, ideally, be 
      looking to depreciate some (all?) of these, not add new and 
      unnecessary ones. 

      {on <FIG SRC="data.tab">}
      Likewise, further complicating the HTML standard by using a 
      separate external format for storing numeric data, just because 
      the author has hinted, via the stylesheet, that he/she would 
      /like/ the data to be rendered as a 3d pie chart titled at 45 
      degrees with the second and fifth segments exploded by 10% of 
      the diameter of the pie, is probably not a good idea. If 
      nothing else, it instantly excludes the person /viewing/ the 
      document from being able to take their favourite browser, point 
      it at a page created by someone with enough knowledge to create 
      tables (but not enough knowledge to use the other format) and 
      choosing a browser option which allows /them/ to render the 
      data in a chart after a little bit of direction on the precise
      structure of the data in the table.

(ii)  platform independence. I prefer the original posters suggestion 
      of using the <TABLE> tag, if for no other reason than if a platform 
      is unable to have an external chart viewer it can merely ignore the 
      styles applied to that TABLE element and render in a tabular text 
      format, or read it aloud in the case of a visually impaired or 
      blind user. Yes, we could come up with a standard which could 
      be interpreted in different fashions by a range of viewers, or
      incorporate an existing standard into HTML/Web Browsers), but 
      why do this (see [i] above) if what we have already is 
      sufficient - isn't it this which HTML is supposed to do be so 
      good at?

(iii) Using the <STYLE>, or <LINK> to stylesheet, mechanisms also 
      gives authors the ability to compose different stylesheets for
      different environments if it is known that a certain 
      environment is incapable of succesfully reproducing a certain 
      style. This would keep the platform-specific brigade happy, if 
      nothing else.

In my mind, there is no need to alter the HTML markup elements in 
order to fulfill the charting requirements - leave it as-is.
Stylesheets provide us with all we need, and they keep it out of the 
actual HTML standard, which is a Real Good Thing, IMHO. :-)

Ciao,


Chris
--
Chris.Tilbury@estate.warwick.ac.uk; MIME Mail Welcome
Tel: +44 1203 523523x2665, Fax: +44 1203 524444