[rfc-i] A4 and 8.5x11

Bruce Lilly blilly at erols.com
Tue Feb 1 09:55:43 PST 2005

John Klensin wrote:

> Folks, if memory serves me correctly, there is actually an ISO
> Standard that specifies print area dimensions for documents that
> might be printed on either paper form.  It is based, to all
> intents and purposes, on the width of A4 and the 11 inch length
> of the North American size.  My vague recollection is that at
> least the earliest versions of Postscript had provisions for it
> and that drivers were supposed to do the right thing, partially
> because the Postscript itself specified "new page" functions in
> the right places.  While anything is possible, I don't imagine
> those capabilities have been removed.
> So, if we are going to go down this path, it would be good to
> dig that spec out and just use it, rather than inventing a new
> set of norms of our own.

I've seen claimed "It was proposed for an early draft of ISO 216
to recommend the special size 210 x 280 mm (a format sometimes
called PA4)".  From the wording, I gather that it is not in fact
*in* any ISO standard.  The size, by any name, is not supported
as a standard size in ghostscript, though presumably it could be
defined on a system-by-system basis.  Ghostscript specifies paper
sizes in points; 11 inch length is exactly 792 points, but the
210mm width works out to approximately 8.267716235 inches or
595.2755906 points; presumably that would have to be rounded to
595 points.

SO, it's not clear that there *is* such a standard...

The so-called PA4 format seems to be a problem; printing a PDF
document from such a source with "Choose Paper Source by PDF
page size" from Adobe Acrobat Reader 6.0.2 on an MS Windows NT
box yielded a printout in one corner of 11x17 inch paper;
hardly the ideal that is claimed for that format.

I'm not convinced that there is any generic way of formatting
a document so that it can be easily printed on either A4 or
letter paper on any platform by any software; it may be possible
to come up with a method that works for a given combination of
OS, software package, printer driver, and printer hardware, but
I strongly suspect that different methods would be required for
different combinations of OS, software, etc.

More information about the rfc-interest mailing list