On Wed, 2 Mar 2005 14:15:33 -0800 Steve Langasek wrote: > > Are you implying that a 2-clause-BSD licensed manual can be > > distributed in main in PDF format, if the LaTeX source (preferred by > > upstream for making modifications to it) is kept secret and not > > available? > > I think it's sucky and we're better off distributing the LaTeX source > as well if we can get access to it, but I'm not convinced that this > should be a release-critical bug. I simply do not believe that LaTeX > -> PDF conversion constitutes a technical barrier to modification to > the same degree as compilation of C/C++/Java source to native > assembly/bytecode, because the amount of higher-level markup > information that's lost differs by an order of magnitude.
So it's fine if someone distributes C code after stripping all the
comments and renaming all variables to non-significant short
identifiers?
I call it obsfuscated code, not source code.
Note that, to the best of my knowledge, no LaTeX comment ends up in the
PDF. Nor any user-defined LaTeX command. Nor any LaTeX label.
> You can
> take a PDF and usefully extract the entire text back out of it (even
> if people set cheesy "no copy" flags in their PDFs, thanks to
> non-crippled readers), and all that's missing is the typesetting
> markup;
I don't want to extract plain text: I want to modify the manual because
I noticed a typo at, say, page 11.
I do not have any free tool to modify PDF files directly. Nor does the
upstream author (he/she uses pdflatex !).
I know how to write LaTeX code and have pdflatex. So does the upstream
author.
Why can upstream fix the typo the easy way, while I cannot (without
rewriting all the LaTeX markup by reverse engineering)?
Do you think that figuring out the LaTeX markup by looking at the
resulting PDF is easy?
> but decompiling a binary gives you none of the text of the
> original higher-level source.
I can extract all the strings contained in the binary executable with a
decompiler (or even with the strings command), but no source comments,
nor variable/constant identifiers.
The same applies to the PDF: none of the original source comments,
label, user-defined commands are recoverable.
Interestingly enough, you also cited Java -> bytecode compilation: I've
been told that Java decompilers can recover even variable identifiers
from bytecode (but I don't know if this is actually true).
For this reason, proprietary programs written in Java are often built
from obfuscated code (so that their actual source code cannot be
obtained too easily from the distributed bytecode).
--
Today is the tomorrow you worried about yesterday.
......................................................................
Francesco Poli GnuPG Key ID = DD6DFCF4
Key fingerprint = C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
pgpoy3ZiWz1TC.pgp
Description: PGP signature
