[Zlib-devel] Re: deflate 64

[Cosmin Truta]

On Wed, 27 Aug 2003, Greg Roelofs wrote:

> Cosmin wrote:
>
> > Probably Winzip obtained the permission from PKWARE.
>
> Or not...
>
> > However, we do need the PKWARE permission to use the deflate64 format.
> > If that's not stated in any public document from PKWARE, then we need to
> > contact PKWARE and have an explicit permission note from them.
>
> And you base that statement on what legal theory?

This is the story:

Gilles proposed to support deflate64 in zlib.
Mark refused, for it is not documented in PKWARE's appnote, hence
he considered this method a closed extension to the ZIP format.
Gilles insisted, saying that Info-ZIP and Winzip are doing the same.

> This isn't a patent issue (that anyone's aware of, anyway), nor
> copyright, license, or trademark.  It may touch on trade secrets, but
> that's explicitly permitted in cases of promoting interoperability.
> Which leaves us with...?

You are right to question my "legal theory" :) IANAL. I know that the
open formats can be supported without legal repercussions (because
they're open...), and I thought it isn't always the same about closed
formats.

I know a concrete example: Adobe clearly says that a 3rd party
implementor can support their Acrobat PDF format, as long as a
PDF-producing implementation does not create PDFs running as a server in
a network. PDF is not really a closed format, but it is a proprietary
format upon which the creator has imposed restrictions.

AFAIK, PKWARE has not published deflate64, nor has it said anything
about the right to reverse engineer and re-implement it. Is it public
domain or not? I thought it doesn't hurt to check.


> FWIW, I'm perfectly fine with whatever Mark wants to do.  64 isn't a major
> win for anything but marketing, and if only one application really needs
> that support for compatibility reasons, then maybe the best place to add
> the relevant routine is in the application itself.

Deflate64 is hurtful for it does more harm than good: it is a weak gain
in compression that diminishes the zip portability in return.

The fact is that deflate64 exists, and I am neither sustaining, nor
opposing its support in zlib. I just wanted to have the legal issues
clarified.


Cosmin