[Nettime-bold] Re: <nettime> Review of the CODE conference (Cambridge/UK, April 5-6, 2001)

[Florian Cramer <paragram@gmx.net>]

Am Sun, 22.Apr.2001 um 20:04:43 +0100 schrieb Benjamin Geer:
 
> Perhaps I'm simply misunderstanding, but it seems to me that there's no
> conflict between Free Software and standardisation; in practice, they
> depend heavily on each other.  Later in your message, you mention open
> standards such as HTML and JPEG; one could also mention PNG images (the

I guess my wording was not precise. While Free Software, as you
mentioned, clearly provides the best support for open standard
protocols, interfaces and file formats, the APIs - i.e. libraries and
component models - are quite diverse and provide quite instable (i.e.
constantly changing) interfaces on the same operating system, often
duplicating functionality (imlib vs. gdk-pixbuf, gimp-print vs.
gnome-canvas vs. display ghostscript vs. X11 print extensions, to name
only a few). 

An application written for KDE 1.1 cannot be compiled for KDE 2.x while
the win32 and MacOS APIs have been stable at least for a couple of
years. The standard interface on Free operating system is POSIX, the
lowest common denominator of shared code the glibc.  Everything else is
in flux and provides inconsistent interfaces to users and programmers. I
personally see this as an advantage (as I prefer choice, configurability
and avoidance of backwards-compatible cruft), and think that it reflects
the openness and freedom of Free Software on the level of structure.  

Proprietary operating systems are more unified and consistent in
themselves simply because they are controlled by central entities, and
because open standards tend to describe low-level rather than high-level
stuff (with the exception of OpenStep, where the Free Software community
made the mistake of rushing KDE and Gnome out of the door instead of
properly finishing GNUstep and end up with a MacOS X-compatible GUI
framework). It is unlikely that Free Software would have produced
anything similar to the classical MacOS or PalmOS both of which are
strictly engineered from a non-technical user's perspective. 

> of a standard GUI is not as much of a problem as you'd think.  The
> different GUI environments are not mutually exclusive; applications
> written for one environment can be run in the other environments, as long
> as the user has the necessary libraries installed.  Which isn't a problem,
> because they're free.

They are not mutually exclusive, but don't interoperate very well. For
users, this means to set up and configure the same things twice or more,
for developers to settle on specific combinations of programming
languages, toolkits and libraries at the risk of running into license
incompatibility and producing code which becomes obsolete once it's
released because the underlying APIs have changed. And it's simply a
drag for a digital artist working with, say, the Gimp, jMAX, PD and
KDevelop to cope with three wholly inconsistent user interfaces (GTK,
Java, Tk, Qt/KDE). 

This is a problem for high-level multimedia and GUI development only,
not for daemons, scripts and console applications which can settle on
the historically quite stable Unix/POSIX interface and nicely
interoperate via pipes, sockets and child processes. (Although still
being inconsistent in keyboard control and configuration syntax.)

Florian

-- 
http://userpage.fu-berlin.de/~cantsin/
http://www.complit.fu-berlin.de/institut/lehrpersonal/cramer.html
GnuPG/PGP public key ID 3D0DACA2 


_______________________________________________
Nettime-bold mailing list
Nettime-bold@nettime.org
http://www.nettime.org/cgi-bin/mailman/listinfo/nettime-bold