# [tex-live] [pdftex] PDF 1.5 by default in TL 2010

James Quirk jjq at galcit.caltech.edu
Wed Jan 13 14:42:26 CET 2010

> bandwidth, and thus polar bears (although it would be interesting to
> see if the additional power consumption due to the compression is
> lower that the average consumption of the saved bandwidth)... So I
I can't help but wonder how many polar bears have been adversely affected
by this thread. But the purpose of this message is not to dwell on such
imponderables, it is to point out an important fact that has escaped the
current protaganists: as of PDF1.4 a document's /Catalog can have a
/Version entry that *overrides* the version given in the file header (PDF
Reference 1.7, Appendix H, p. 1096). In other words, for some documents
the version given in the file header is moot. For example, try processing
the following:

\documentclass{article}
\begin{document}
\pdfcatalog{/Version /2.0}
In this example the version in the PDF file header
is overriden by that in the document's /Catalog.
\end{document}

and then view it using AR9.3, which was released yesterday.
Even though the resultant PDF has a file header, %PDF-1.4,
AR9.3 will complain that the file is a newer format than it can handle.

The ISO-32000 spec extends this idea further with the use of
/BaseVersion and /ExtensionLevel. Of course, many of the "free"
PDF viewers do not recognize these /Catalog entries (e.g.
xpdf 3.02). Therein lies the problem, what's the point of having
an international standard if individuals choose to use
non-compliant browsers?

I do understand why people choose to do so (acrocrap and all that),
but if I author a PDF, I'm going to have no qualms in using
a feature that's in ISO-32000, especially as Adobe Reader is
freely available. And if I lose some readers in the
process, so be it: they're probably not the sort of people
I'm aiming to communicate with, anyway.

> guess the whole point is to know whether it's better to have small
> PDFs or PDFs readable by a 9 year-old software... does TeXLive have
> something like a voting system?
>
A vote would be fine, so long as the electorate are properly
educated to know the pros and cons of what they're voting on.
And as I've tried to show here, in this particular instance
matters are a little bit more involved than the current debate
suggests. The good news, however, is that the /Version entry
in the /Catalog could be leveraged to allow macro packages
to fix the effective level of a PDF. This would naturally
require a standardized interface to allow an individual
package to stipulate its specific requirement, so that the
highest requirement can then be output in the /Version name.

James