[KLUG Advocacy] Re: Advocacy digest, Vol 1 #11 - 1 msg

[Adam Williams]

>>The cross examining of Microsoft's "expert" witness Stuart E. Madnick:
>>When government attorney Kevin
>>Hodges asked him to name an operating
>>system besides those made by Microsoft in
>>which the Web browsing software could not
>>be removed, Madnick immediately offered up
>>KDE as an example.  But KDE is a computer program designed to run
>>on top of the Linux operating system, as Hodges pointed out.
>>Madnick conceded that was true, and instead suggested GNOME as
>>an example.
>>http://money.cnn.com/2002/05/02/technology/microsoft/index.htm
>This reminds me of a saying attributed to Ghandi: 
>                First they ignore you
>                    Then they laugh at you
>                      Then they become angry at you
>                         Then you win
>It seems that while the rest of Microsoft is at stage three, this fellow is 
>still more or less at stage 1.

Stage 4 does require that one survive stage 3.  I can think of several
great packages that have died at stage 3:  OS/2, Lotus 1-2-3, Word
Perfect, MacOS (at least before OS/X,  lets face it, effectively the Mac
was dead, OS/X was a brilliant coup).

All these products (at least IMHO) were gunned down using quite focused
marketing and feature-tweaking tactics.  I'm certain I could think of
more if I spent some time meditating on it.  

There are others I could think of: Informix, OpenMail, etc... that were
great products that suffered from targeted marketing but in reality
probably died as a result of hopeless marketing and corporate
cluelessness.

I mention this as I think there is still a realistic chance that M$ will
successfully crunch Open Source.  Not via technology, but through
"mandated" proprietary services (ala PassPort, Active Directory,
etc....) and whacked licensing and pricing policies (paying for use of
M$ software by computer count,  irregardless of whether M$ products are
loaded on the machine).  I do not think academic computing alone can
support the Open Source movement,  corporate entities and academic
entities are now much too closely intertwined.  If corporate America
moves away from Open Source, it will wither.

>On a bit more serious note, my read of the transcript seems to show that the
>testimony covers some interesting points; notably that rendering information 
>at URLs is built into what are nominally several file/disk management applica-
>tions (in both Windows and KDE, for example, and that the lines between the act
>we know as "browsing", GUI apps in general, and OS utilities is indeed blur-
>ring. 

Yes.  Witness gnome-vfs:  the clearest expression currently is
Nautilus.  It will show you what resides at a URL (and many protocols
even beyond http and ftp), a UNC (ex: \\servername\sharename\filename),
a standard file/directory, and I've seen alphas of even more bizarre
things like news groups and data base tables.  Does nautilus know
anything about these things?  No,  it just hands them off to gnome-vfs. 
And does gnome-vfs know anything about these things?  No, it uses a
little logic and a search of its registry to find a dynamically loadable
module that does, and then summons that module.  Does the module know it
is displaying something in nautilus, or even where (local computer or a
computer in the Ukraine)? No.  The module just gets a component handle
(probably some thing derived from a CORBA IOR) and passes the
information to the widget set.  The widgets draw themselves in some
window (oh, it happens to be nautilus) drawing upon the users
preferences/themes,  which by the way are acquired from another module
that knows about the users preferences.  Oh wait! That preference
module.... probably used gnome-vfs and gnome-xml to acquire and parse
your preferences.

So where does the functionality of nautilus/gmc/galeon end and the
functionality of GNOME (oaf/bonobo/corba/gnome-vfs/gconf, etc...) begin?
...insert long protracted philosophical argument here...

But does this mean the web browser (or any application) and the OS are
one?  No.  All the above could happen on Linux, FreeBSD, AIX, Solaris,
or (E! gads) M$-Windows. (Yes, there is a Win32 port of GNOME).  The OS
simply empowers the hardware to "run" and exports some (hopefully)
consistent and flexible interface/expression of the hardware.

Something else sits on the OS (a component layer, desktop manager, user
interface, all it what you will).  In cases such as Linux it is pretty
easy to conceptualize where one end and the other begins.  IMHO,  with
Linux, the point of OS-endage and platform-architecure-beginage is best
described as "glibc".  But on something like Win32 it is quite a bit
more muddy,  although there still is an underlying API, upon which
things like OLE, DDE, COM all rest.

Mostly I think the "browser bundling" arguement is spurious, and a
really dumb tactic.  Did M$ use monopolistic practices to brow beat
competition?  Have they deliberately created egregious barriers-to-entry
for new competition?  That is the real question at hand.  I suspect M$
is very happy the more things focus on the whole IE issue.   Did they
prohibit OEMS from offering other products?  Did they design products in
order limit interoperability?  To me those are the real questions.

>A technical point that the proceedings seem to have jumped over is the
>degree of modularity inherent in modern OS'es, to the extent that a utility
>can call on particular .dll's (or .ld.so's, as the case many be) to bring
>large new claases of functionality into play. Also missing is the notion that 
>one can run a GUI without a broswer at all...