Re: Page breaks when you print formatted html docs

Daniel W. Connolly (
Fri, 03 Jun 1994 12:55:34 -0500

In message <>, Steve Waterbury writes:
>Dan Connolly wrote:
>> On the practical side, I think you can choose either form of markup
>> and get the results you're after -- namely, that it doesn't show up on
>> the screen when you preview it with Mosaic. Quick check... Crud.
>> Mosaic displays &pagebreak; and <?pagebreak>.
>> So much for that idea!
>But it sounds like such a good idea -- what are the chances of having
>that implemented and put into the HTML spec? (I.e., that <? ...> and
>&... forms are not displayed by browsers.) It seems very useful,
>along with being proper SGML.

Here's my view:

The DTD will have a declaration:
<!ENTITY #DEFAULT SDATA "&#38;unkown;" --display the markup-->

which says (in SGML) that arbitrary entity references are legal and
(in application conventions) that if you see an entity that you don't
recognize, just display the markup. So you might see:

When &alpha; > &beta;, it is clear that...

So in the future, &pagebreak; won't be silently ignored -- it'll
either be handled in some way specified by the DTD, or it'll be displayed
as &pagebreak;.

On the other hand, <? pagebreak> should _not_ be displayed. The fact
that Mosaic fails to recognize this as markup is just a plain old bug.

The question of how to manage the namespace of processing instructions
remains. I suggest we don't manage it at all -- we leave it wide open
for private mechanisms (text formatting systems, index generators, and
the like). This leaves the possibility that private mechanisms will
clash, but ...