On Fri, 2004-03-26 at 17:40, Daniel Crookston wrote: > I'm curious - did anyone read the article I posted a link to? Because I > think that Eric Raymond (the author of that article) brings up some > wonderful points there and in follow-ups that, applied to > Linux, would make Aunt Tillie (and your sister and dad) perfectly happy with > it.
I've been waiting for someone to bring up "The Luxury of Ignorance" for
some time now. Since the old thread seems to have been hijacked by
politics and distro wars, I'm pulling this one out of the fray.
When I read this article, I cheered. I've been involved in linux for
close to ten years now, and I still consider myself a newbie. "The
Luxury of Ignorance" explains why. (and I will too ;)
I don't want this to turn into a distro/toolkit/editor/whatever war, and
I realize that I use specific names here, they should just be considered
as examples, not as my religion. I've thought about this issue quite a
bit, and I would like to hear what everyone else thinks, but I would
appreciate it if you would think about your replies and not just attack
me without good cause.
What follows are some thoughts that I put down after the Miguel de Icaza
talk on Monday. He made some comments about Microsoft Longhorn that
made me think and realize that we have a lot to do to make Linux better:
He wasn't a Microsoft hater. I thing it's interesting that the people
that really know Linux well have a much more mature opinion of Microsoft
than the ones we usually see at the local level. (RMS and company
excluded)
When I started using Linux, I too was a Micro$oft basher and tried to
avoid everything they made. Since then I've matured some (it's a good
thing I've got eternity) and I have learned to respect much that
Microsoft does. This doesn't mean that I could ever work there; I
disagree with their work environment. Microsoft products (for the most
part,) "Just Work." I think this is very important if we want Linux to
be successful on the desktop, the laptop, or anywhere but the server
room.
I've recently experimented with Gentoo. I tried for weeks and weeks to
get the system (my laptop) to a usable state. (Usable meaning: gui,
Internet, wireless, sound, dvds, 3D, battery status) I was able to get
most of it working, but only after spending hours and hours in the
forums (which are excellent) trying to determine what the next step was.
Finally I stepped back and realized that with all my work, my laptop was
^almost^ as functional as when it had Windows on it. Weeks of my life
gone to achieve what was already provided for me, at a small cost that
was wrapped into my purchase of the laptop. The GUI "Just Works" on
Windows, whereas on Gentoo Linux it took 24 hours to compile KDE to get
a GUI. The Internet "Just Works" on windows, but Gentoo required that I
manually configure everything required for my network card to work in
each of the environments I use it in. (dynamic IP at home or campus,
static IP at work, no IP when I'm not plugged in) Every single thing
that I wanted Gentoo to do, Windows did out of the box. I really want to
use Linux, but I can't devote my life to configuring arcane things that
the computer should do for me (why does Gentoo require that you write
your own /etc/fstab?)
I'm now using Fedora Core 1, and it is working so well, that I will use
it before I use windows for most things. This is the longest that I've
been able to use a distribution and I'm using it more time than ever
before. There are still things that bother me, and send me back to
windows, but they are not coming nearly as often. Last week I was
sitting in a class with my laptop connected to the ethernet link for the
TEC room (The professor must wonder why I sit so close) Someone came up
to me and asked me to share the internet connection, so we could all be
online at the same time. On Windows, this took about 3 minutes, with no
documentation. On Linux, I'm still reading mailing lists. Linux still
needs to be smarter about what it does and hide needless complexity from
the user, while allowing access to all of the complexity needer or
desired.
There are several reasons why hiding the complexity is hard. Writing
documentation and running tests on software is not as “sexy” as the
research that is happening on the front lines. Since most of the work
put into open source projects is donated by people in their spare time,
most of the work gets done in exciting areas where new things are
happening. One of the reasons commercial software costs so much is
because of the extensive testing that must be performed before the
product can ship. The testing isn't perfect, and there will still be
bugs, but they are far fewer in number than in a project that doesn't
have anyone willing to complete it and polish it up.
Documentation is also lacking, and people new to Linux are often told
to RTFM, when the manual may be incomplete or non-existent (oh, well
check the -dev mailing list then.) If the software built is easy enough
to use, requests for the manual should go down. To illustrate this,
compare the size of the manual for MS-DOS to that of the manual for
Windows XP. How can Microsoft get away with using a much smaller manual
when the product is much more complicated and is installed on many more
computers? As much as Microsoft basher would like to deny it, Windows XP
is an improvement to what came before. It is easier to use and configure
than their previous products, and much more so than Linux. My wife's
desktop runs Windows XP and it is so stable, that we may only reboot
once a month (yes, I know that Linux can do better than that) which is
huge progress from the Windows 95 days that required many many reboots
of the system.
I would like to see more Linux projects adopt this idea of software
that “Just Works.” Knoppix is making progress in this direction, and is
being used by people that wouldn't be able to use Linux otherwise. They
have a good model and they use the computer for what it was meant for:
to make things easier for the user. Knoppix demonstrates something that
Linux does better than Windows, instead of Linux trying to catch up to
where Microsoft has placed the bar. I shouldn't have to know what
motherboard or sound card I have, the computer should be able to figure
that out. Likewise for network card configuration; the computer should
be able to make most of the decisions for me, and only ask when it needs
help. This doesn't mean that all the options should be automated and
picked by the computer. It is important that the ability to tweak things
to my content is still there. GNOME seems to be taking this path and is
providing the Gconf tool to allow the fine grained control for those
that want it. Joe User only needs to know that the tool exists, waiting
for him if he ever wants more than the default options. The Windows
Registry worked on the same principle, but has grown too large and
unwieldy to use except in rare cases. We should move all the nifty
gizmos and gadgets to behind “Advanced” buttons and leave just what the
user needs most. I don't use most of the features of Openoffice.org, but
I want the one's I do use to be easy and convenient. Some applications
do this well, and they get more use (at least by me, and probably many
others) than the application that requires detailed study to discover
the right sequence of commands to do what you want.
I've probably worn out most of you by now, and the rest are ^flaming^
mad. How can we as a community work to improve these issues? What can
I do as an individual who wants to contribute, but feels like it's too
hard to do so? I'm not a stellar assembly coder, can I still help Linux
progress? Some food for thought.
Nathan Rackliffe
signature.asc
Description: This is a digitally signed message part
____________________ BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________ List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
