Subject: Re: Lack of imagination?
From: Zimran Ahmed <>
Date: Sat, 30 Jun 01 02:23:05 -0400

>I've certainly been one of the people who has complained that the Linux
>desktop vision was too small, but my complaint was not that it didn't
>innovate (hey, in XP, MS is finally copying the multiple desktops idea
>from X), but that they missed the network-centricity of the next
>generation of software applications.  

I have not used any GNU/Linux desktop GUIs for a *long* time, so please 
don't flame me if my information is out of date, but I too have been 
dissappointed in the approach they've taken. By and large, efforts seem 
to be built around replicating the Office productivity suite: Word, 
Excel, etc.

What's surprised me is why they have not taken all those good bit 
(digital information) manipulation tools from UNIX and ported them to the 
front end on the GUI as standard. On the Mac I use at home, I have a 
customized syustem with a core set of 5 apps:
- text editor
- browser
- email client
- file manager
- calender (with day linked to-do list)
which i've tied together through hotkeys. The whole system is backed by a 
key-stroke memorizer (which works as shell script) and text expander (tab 
completion function)

This combo allows me to flip plaintext between the apps just as i would 
in UNIX, except I use cut and paste instead of pipe, BBEdit instead of 
vi, and Claris Emailer instead of pine. Mac finder works essentially as 
well as grep (it can look in .pdfs) and much better than finder in even 
Windows 2000 professional. Most of my bits are kept in plaintext, in a 
rigid directory structure which again mimics the best practices of UNIX.

The UNIX environment--small simple programs working together by swapping 
plaintext between themselves, is *so* much better at handling networked 
information that Windows' large, monolithic software approach. It seems a 
little nuts that the GNU/Linux desktop is trying its best to emulate an 
application-set designed for the pre-networked PC world, while Windows is 
shifting it's inferior monolithic PC software model to the networked 
world. I use Windows at work, and my customized Mac (which actually is 
not a Mac at all, it's "UNIX for the front end") lets me work about 10 
times faster. 

The UNIX interface metaphor grew out of the networked world and is better 
in a networked environment than the Mac/Windows interface meatphor that 
grew out of the PC world. A GUI which embraces the UNIX metaphor is real 
networked computing for the non-technical user, not the PC-net extension 
that characterizes Windows-Internet integration.

(Full disclosure: this customized front end was developed by Mark Hurst, 
not me. also, if GNOME, KDE or any other OS GUI have changes since I last 
used them and incorporate these tools and the UNIX idea behind them -- my 
deepest apologies. please ignore the above).

>One of the key concepts for all of us to keep firmly in the forefront of
>our minds is that open source goes hand in hand with open protocols. 
>sendmail and SMTP, apache and http, etc.  If MS is able to proprietize
>the next generation of internet protocols, the internet is in deep
>sh*t.  Creating open implementations of .Net (and especially of

we could be in deep sh*t even if MSFT does not proprietize the next 
generation of protocols. Proprietary centralized authentication 
infrastructure can be built on open source software and open 
protocols--you don't need to control the code to control the network. 

>So I don't consider open source .Net an issue of innovation so much as
>it's an essential defensive move.  Innovation isn't the only important
>thing in technology.  Standards and level playing fields are also
>incredibly important, and we have to realize just how important a role

agreed. In some ways, not needing to innovate means OS needs to do less 
work. Doing something right (from a tech, business, and strategy 
perspective) is more important than doing something first, as we have 
seen too often in the technology world.