RE: How to highlight columns?

> Actually, if I understand it correctly, the script below, instead of
> assuming that a tab is one char, assumes that a tab is 8 
> chars (or whatever
> tab-width is) Since a tab can vary between 1 and tab-width 
> chars, this is
> just as wrong as counting tabs as 1 char.

You're right except that in xemacs a tab is always tab-width. When the mode indents the code, 
it use spaces for less than tab and it puts a tab only when tab-width spaces are needed.
> 
> Since tabs are an issue, PERHAPS it's acceptable to untabify 
> the region,
> then tabify it when done?

That's actually the way I used to work, on loading a c file I'd automatically untabify it, and re-tabify on save.
However, it didn't do exactly what I thought it does.

> 
> Also, FWIW, I don't know what system you are working on, but 
> most systems
> I've worked on that were sensitive to line length did NOT 
> like tabs at all
> (such as MVS and its 80-column-card-image orientation). So, 
> perhaps it's
> appropriate to never even use tabs?

[sigh] I'm doing linux kernel device drivers. The kernel folks are extremely contiencious about 
the code style. It must follow K&R style, it must use 8-space indents, and they must be tabs. 
Lines must not spill beyond 80 chars, since if they do it means "you did something wrong in the code" 
i.e. your nesting level is too deep.

Thanks,
Dan