Re: <pre>

Brian Behlendorf (
Thu, 6 Oct 1994 23:30:09 +0100

On Thu, 6 Oct 1994, Brian Gaines wrote:
> > It should be noted that people who use desktop publishing software heavily
> > (Quark, Word, etc.) have it drilled into them to *never* put line breaks at
> > the end of the sentences, only at the ends of paragraphs. Now, conceptually
> > asking them to end their lines with a hard break and beginning and ending
> > paragraphs in <P></P> might not sound too hard to us, but to them it's
> > like asking them to use a Dvorak keyboard. This actually came up in the
> > realization by people here that typing long strings into TEXTAREA fields
> > without line breaks is aesthetically pretty ugly and contrary to what
> > most people are used to.
> >
> The logic seems back to front. Putting a <P> at the beginning of a paragraph
> (the </P> is optional and rarely used) marks a para without line breaks
> eaxctly as in a word processor. The original use of <P> at the end to break a
> para was probably more natural but bad sgml.

Really? I thought the concensus was to encapsulate paragraphs in <P>,
and that it *was* good SGML (and allowed things like <P ALIGN=CENTER>,
which we really REALLY want).

> TEXTAREA fields should always word wrap exactly as a word processor, and
> are aesthetically, and in human factors, exactly what a word processor
> user would expect.

What browser does this for you? None of the ones I've tested do, I don't

> The nice thing about html and forms is that they DO conform to what most
> word processor users expect.

I wish.

> The semi-binary interpretation of <pre> is fine now that it is well-defined.
> It will continue to confuse new users of html who naturally assume that
> 'preformatted' actually means simply 'preformatted', but will always be
> essential to read the specifications carefully to use html correctly. It
> is a precise programming language not a set of intuitively grounded
> suggestions.

Which is a good way to intimidate content providers away from HTML.

Again, maybe I'm making a bigger deal out of this than I need to,
expecially when doing a s/CR/CRLF/g is so easy to fix this problem.