I installed ghostscript on a unix server, and use it to create pdf files for windows clients ('pdf printer').

When I send true type fonts as outlines to Ghostscript, the pdf document created looks great.
However, the text search tool of Acrobat Reader does not work in these pdf files: never finds any text that is apparently in the document.
Likewise, the text extract tool of Acrobat Reader does not seem to work either: it extracts only unreadable garbage.

This is an old problem of mine, since I had the same problem 3 years ago with the windows version of Ghostscript.

What causes the problem and is there a workaround?

it seems like it is not creating "font" items, just curves...

maybe that gives u a hint, i have never worked with ghostscript though

When using the text extract tool of Acrobat Reader, I found that:
- the extracted, unreadable text has the same length as the original one,
- each character is "mapped" to an other ASCII character like this: if I extract a certain (constant) number from the ASCII code of the extracted character, I get the ASCII code of the original character.

I think this proves that Ghostscript creates the fonts, only these fonts have different ASCII code mappings, and that is why text search does not work.

The above is based on my finding with the Windows version three years ago, but the unix version produces very similar output, so I think the shift in the ASCII code causes the problem there, too.
But why?

P.S.
Using the printer fonts instead of sending the TrueType ones produces unacceptable output: incorrect accented characters and overlapping characters.
Sending TrueType as bitmap produces very ugly documents (rough pixels), and the text extract tool of AcroRead will not work either.
Only Open Office 1.1 creates correct pdf files, but I need a server solution, for which Ghostscript would be ideal.