On Mon, May 03, 1999 at 03:21:21PM -0700, ferret@ricochet.net wrote: > Anyway, wouldn't the existance of a client implimentation imply sufficient > information to design and/or code a compatible server? > > I would be interested in information concerning the truth or falsity of > the second statement. :) The X protocol is defined in a freely available document (install the xbooks package). Using no other resources, code a working X server using that document. Reference to the source code of existing X server implementations, under any license, is forbidden. Client/server system design goes far beyond the specification of the wire protocol. For simple protocols, it may be feasible to code a client or server knowing nothing else. As the protocol complicates, the complexity of the implentation goes up in what I would suspect to be an egregiously non-linear fashion. I'm not saying ICQ or AIM are quite as complicated as X, but I am saying that feasibility in principle is a far cry from feasibility in practice. This is why I am concerned about risk of enslavement to "practically proprietary" protocols, especially when proprietary interests try to lure you into it by distributing the client side as widely and freely as possible, but keeping the server under lock and key. -- G. Branden Robinson | Debian GNU/Linux | Never attribute to malice that which can branden@ecn.purdue.edu | be adequately explained by stupidity. cartoon.ecn.purdue.edu/~branden/ |
Attachment:
pgpugvQbT8X96.pgp
Description: PGP signature