Re: lynx-dev proposal: rendering tables

[Larry W. Virden]

From: Chuck Martin <address@hidden>

> If all tables were really tables, I would agree with you.  Howver, the
> vast majority of "tables" are abuses of the TABLE element, intended to
> make web pages "pretty."  For example, it's very common to see a "table"
> with some variation of the following format:

Perhaps trying to find just _1_ algorithm for table display is the problem.
Perhaps the initial approach is to figure out how to implement a modular
approach to tables.

For instance, what if there were a compile time option that indicates that
lynx should call an external program when encountering a <table> tag,
passing the table tag _and everything to the matching </table>_ along as
input.  The stdout from the external program would be HTML without tables
in it.

This might allow lynx to remain relatively one pass and then to defer
table handling to some other piece of code which might do better.

For that matter, with this construct one could, in theory, write one
multi pass function which is called with the <table>...</table> as
input and which itself did recursive processing I suppose using the
other functions.  I suspect this would be more difficult.

As for this 'line at a time' approach, perhaps if column headers are provided
they should be displayed each time a column is output?

-- 
Larry W. Virden                 <URL: mailto:address@hidden>
<URL: http://www.purl.org/NET/lvirden/> <*> O- "No one is what he seems."
Unless explicitly stated to the contrary, nothing in this posting should 
be construed as representing my employer's opinions.