Linux-Advocacy Digest #609, Volume #28 Thu, 24 Aug 00 04:13:04 EDT
Contents:
Re: The Powerful Linux GUI Interfaces - Re: Why is "ease of use" a dirty concept?
(R.E.Ballard ( Rex Ballard ))
----------------------------------------------------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Subject: Re: The Powerful Linux GUI Interfaces - Re: Why is "ease of use" a dirty
concept?
Date: Thu, 24 Aug 2000 07:14:39 GMT
In article <[EMAIL PROTECTED]>,
"1$worth" <"1$worth"@costreduction.plseremove.screaming.net> wrote:
> s"R.E.Ballard ( Rex Ballard )" wrote:
> >
> > In article
<[EMAIL PROTECTED]>,
> > "1$Worth" <"1$Worth"@costreduction.plseremove.screaming.net>
wrote:
> [snip]
> > Actually, Linux has a number of GUIs, which, when the user is given
> > the choice of the "Redmond95" (windows-like) and others such as
> > KDE, Enlightenment, or even FVWM2 with AfterStep, indicates that
> > those interested in Linux are least interested in Redmond95 (the
> > look and feel of Microsoft).
>
> ..and this is a great strength.
Yes. The GUI is very much a matter of taste, of intangible preferences.
Some people love the Mac UI, others loved Windows 3.1, others loved
OS/2, and others loved Windows 95. Windows 98 even gives you a choice
of GUIs.
> Of course in the
> Linux world everything that is under the bonnet is available for
> inspection and repair and fine tuning. However here you have a
> consistent user interface, but one which does not take the power away
> from the mechanics to alter the innards.
> The main point is that all cars
> have the same consistent interface and even the most inept people can
> use. This is not the case in the GUI+config tools world and that is
> where my issue is. If this is not corrected then Linux will just not
> make it to the desktop of most people.
Again, if we're talking about a properly configured, preinstalled
system, nearly everything CAN be configured by a semi-intelligent
user using Linuxconf, which is the equivelant to the Windows "control
panel" but with fewer cartoon characters.
Part of the structure of Linuxconf is that if you start at the top
and work your way down, and get input from the administrator as
you go down the list, you end up with a nicely configured Linux
system.
There have even been times when I did something the old fashioned way
(vi or simple editor on /etc files), and then found the configuration
tool in Linuxconf or DrakeConf. Even better, the GUI tool helped me
find settings I missed skimming through the printed documentation.
In an office environment, these things would be done by the
configuration management team before the machine was ever delivered
to the user's desk. In a retail environment, retailers who can
configure
the Linux system "ready to run", and even provide a floppy with a shell
script that restores the appropriate /etc files will probably get a much
better price than the dealer who simply carries the sealed box to your
car for you.
> > > I propose that they are not mutually exclusive.
> >
> > I would agree. Just as the driver's view of an automobile is very
> > different from the Mechanic's view, the Linux user's point of view
> > can and should be very different from the administrators view, which
> > should be very different from the Installer/OEM's view.
>
> Good.
>
> > In the Air Force, they hire PhDs to design planes flown by College
> > Graduates, that can be fixed by High School drop-outs (if
neccessary)
> > and still leave the country with a functional fleet of planes.
>
> I know quite a few phd's and i would not trust them to build my plane!
> :-)
But when you are trying to design an airplane that can take off
and land on an aircraft carrier (navy/marines) or short runways
(airforce) and still fly fast enough to outrun the enemy and maneuver
fast enough avoid the enemy's missles, you don't want that plane
designed by a college freshman do you?
> > > If you look at the wonderful projects that
> > > are around such as Gnome and
> > > Eazel, and then you look at the first
> > > stages of graphical configuration
> > > utilities and the installer programs
> > > of the latest Linux distributions
> > > then this indicates that ease of use
> > > is actually what people want.
> >
> > Much as that 2000 Lexus was very different from Grandpa's chevy,
> > the 2000 Linux is very different from even the earliest Macs and
> > Windows.
> >
> > The Linux community understood that much of what existed in the
> > UNIX engine was the result of some of the most intense engineering
> > effort in the world. UNIX has been supported, enhanced, and
> > evolved in all 7 continents, over 400 countries, and by over
> > 170,000 engineers, before Linus ever released his first 10,000
> > line version of Linux (on display on the Red Hat annual report).
> > Since then, Linux has become the point of convergence for this
> > engineering effort.
> > Linux isn't just an OS kernel. If it were, it would probably
> > have been userped by freeBSD back in 1994. Linux has become
> > synonymous with one of the most full-featured, robust, and
> > extensive collections of intellectual property ever produced.
> > Beneficiaries reaped the benefits of prior development efforts,
> > and contributed in their own way as their way of returning the
> > generousity. For some, this meant coding, writing new software
> > such as KDE, GNOME, or AfterStep. For others, this meant
> > documentation such as participation on the Man, Howto, and other
> > LDP projects. For others, this meant participating as advocates.
> > For others, it meant 1 to 1 promotion such as Linux "install
parties".
> > Many of the "WinTrolls" have never quite experienced the better side
> > of the Linux community. In some cases, they tried to install a copy
> > of Linux that they downloaded from an FTP site (often using
different
> > dates of releases), and then attempted to install the software with
> > access to minimal documentation. Furthermore, they used randomly
> > chosen equipment, often without regard to Linux compatibility.
>
> Absolutely spot on here. It is a shame that people have been put off
by
> this because I have found the Linux world to be very rich and
friendly.
> It is amazing how many helpful people are available to answer your
> questions provided that you have had a good stab at it first. There
are
> reams of free documentation available and many very well written and
> helpful. This is truly a wonderful thing and this is the reason that
I'd
> like the more "die-hard" members of the community to understand that
for
> Linux to be acceptable for people who are not like us, there must be
> some changes in attitudes. There are already many fine efforts to make
> things better, faster and more effective and I hope that the momentum
of
> Linux-hype will spur for-profit companies to produce software and
> drivers for the platform. This in turn will produce the snow-ball
effect
> that windows has benefited from.
I thing there have been huge leaps of progress in this direction in
the KDE, KGNOME, Enlightenment, and AfterStep communities. This doesn't
mean that the GUI developers can turn in their coding cards and go
back to "real work". In fact, just the opposite. GNOME and KDE, among
others have created a very good framework for creating "quick and dirty"
applications that look pretty polished. I would expect to see this
technology plugged into new technology such as sattellite/sideband
newsfeeds which no version of Windows has successfully been able
to parse in real-time. Even Java has a hard time with this.
Linux has been able to parse the feed since 1994, but hasn't been
on the desktop to display it in real-time.
Expect to see other real-time features like bid/ask/confirm trade
systems for everything from stocks to Jumbo Jets.
I'd also expect to see a move from e-mail to NNTP news-feeds, private
news servers, and corporate chat-rooms. I currently have a "buddy"
system connected to the private network. I can see who is connected,
and get in contact as soon as they get connected. This is very nice
for urgent information (deals that are ready to close, deal-killers
that the executives can handle, and last minute negotiations on sticking
points that are being "snuck in". at the last minute.
> The subject that most Lin-advocates should
> ask strong questions about is
> why M$ is so successful?
Microsoft combined mediocre technology, with the guts to say yes
when they didn't know HOW they could deliver, and the ability to
make the largest PC makers in the world feel that their survival
depends on Microsoft's good will.
MS-DOS was based on CP/M 80. It had almost no memory management,
(what it had was used very poorly), it had almost no storage management,
and almost none of the features of CP/M 86 OR of Version 6 UNIX.
Windows NT was based on VMS technology. In fact, Microsoft picked
up most of the VMS programmers who had been laid off by DEC because
VMS was considered a dead product as far as the Alpha was concerned.
There wasn't much demand for upgrades. People who had VMS on VAXen
were either quite satisfied, or they were looking at UNIX, not more VMS.
Mind you this was not "Bad" technology. In 1982, CP/M was liked
by most corporate offices and professional users (who often purchased
their own machines and brought them to the office). In 1992, VMS didn't
have a huge market share (compared to UNIX) but most VMS customers were
very happy with it.
It's a strategy that Microsoft first developed in 1977 with MITS,
hand has implemented very successfully until August 2000, when,
as a result of public disclosures via the DOJ court hearings,
the OEMs independentently and jointly spent an entire year
cultivating and developing technology that would reduce their
dependency on Microsoft.
This year, Microsoft got a lot of push-back, and nearly all of the
OEMs got deals that allowed them to promote Linux products as well
as Microsoft products (depriving Microsoft of veto control
of nearly $4 billion in advertizing revenue). Many even pressed
for the ability to sell dual-boot systems.
We will be seeing a number of Linux for WinCE systems, and a number
of Linux-only systems. We saw them at PC Expo, Linux Expo, and will
probably be seeing them at CES as well. And the press, who has treated
Windows 2000 as a non-event, has been very interested in Linux coverage.
> There are many reasons, some good, some bad,
> but at the end of the day their products have penetrated where Linux
> could only dream to go. So I suggest we start dreaming, and start
taking
> some practical steps. My suggestions follow:
>
> 1. USB drivers (when 2.4 comes out) for the many new exciting
> peripherals
This is being done. USB is supported in 2.2.15 kernels (Mandrake 7.0)
and SuSE 6.4. I'm sure others are supporting it as well.
> 2. LinModems: the curse of software modems will not go away and
efforts
> by linmodems.org are commendable.
Unfortunately, linmodems/winmodems are very processor hungry, require
very high priority service, and have a direct impact on system
performance. OTOH, USB Modem support may resolve this problem as well.
I have noticed that many OEMs are now offering smart-modems. You can
still do things like voice synthisis, but you can also use the modems
as standard serial modems.
> 3. An agreement by the most popular distributions to follow *default*
> standards:
> a) of file system layouts
> b) configuration locations
These are being supported by Red Hat, Caldera, and SuSEe.
Debian still does things differently, but a little pressure
from Corel could quickly change that.
> c) package formats (deb v rpm)
Most distributions no support both interchangably. They also
support BSD packages, and several others. They are standard
applications for KDE and GNOME.
> d) Default GUI (Gnome v kde)
Again, there is support for both, and most systems no allow
users who logon with the XDM interface to choose their GUI
at log-in time. I frequently switch around (enjoying the buffet).
They aren't particularly large, they aren't expensive, and they
consume very little RAM, Hard Drive, or tool-space.
My KDE desktop lets me run GNOME apps, and my GNOME lets me run
KDE apps.
Let the users choose the applications they want, regardless of
desktop and toolkits.
> e) Configuration tools
There's more than one way to skin a cat, and there's more than one
way to edit/configure an etc file. Many administrators love Netconfig
and
Red Hat Control panel, others live Linuxconfig.
> f) Printer subsystem - lets make printer programming easier
More importantly, set up ALL the appropriate driver sets for the
printer. Most people don't understand lpd "filters", they don't
realize they can convert DVI, Postscript/Ghostscript, and roff
files to their printers formats (DCL, PCL). This should be
handled by the configuration tool.
For example, if I want to install a BJC 800 printer, I should have
LP1 as "raw" (for windows shares), LP2 as ghostscript to RAW,
LP3 as DVI to RAW, and LP4 as Text to raw. In some cases, these
filters are as trivial as prepending a prefix configuration code.
> g) Consistent look and feel. This
is mostly governed by the qt v gtk
> debate, but drag and drop and copy/paste
> and ole (bornobo) need to be > addressed.
This may be a combination of training users in X11 paridigms as well.
many programs don't do copy/paste because it's already part of X11.
Even when cut/paste is available as a menu item, it only works
"across the board" if it's an X11 cut/paste.
Drag & drop we discussed above.
> I think that the linux users can do what
> they like and may not choose
> the defaults, but this will make things
> more consistent and give a level
> base with which to work from and provide support for.
In many ways, this is like purchasing a car and thinking that
you can still only go from railroad station to railroad station
because there is a road next to the track.
The automobile freed the transportation consumer of
train schedules, locations near the very loud train station,
and fixed routes that often required going to a hub to reach
a spoke only a few miles from your original destination.
In the same way, Linux gives you a great deal more latitude
in terms of routes to a destination. You want to make sure
that you have BOTH.
This was what eventually happened to UNIX. After a bunch of
posturing and market hype about the 2% of the code that was
different between BSD and SysV, they eventually converged
the two products, with AT&T adopting BSD code, and BSD
getting rights to AT&T code as well.
The general rule seems to have become "code to the lowest
common denominator". Which means that if KDE won market share
through hype and advertizing, GNOME would simple GPL and
"outfeature" KDE until equilibrium was reestablished.
This is the nature of a competitive marketplace. Which
is ultimately a good thing for the consumers.
More important is the continued adherance and establishment
of "cross-bar standards" that assure nonexclusive design.
Since both object sets use CORBA, it's concievable that you
could access a KDE server object with a GNOME GUI/Client stub.
you'd need a trivial bridge, but it wouldn't be too difficult.
The irony here is that commercial vendors use CORBA to isolate
the GNU GPL software from their proprietary/BSD blended software.
This has also had the effect of making it possible to call
commercial skeletons from GPL client stubs.
> If these prerequisites are met then I believe that
> software and then users will follow.
Your prerequisites still illustrate a lack of understanding about
Linux.
Yes, back when you only had 2 meg of RAM and a 40 meg hard drive,
you had to make every byte count. You had room for ONE GUI, one
Window manager, one Desktop, and one set of tools.
Today, a 20 gig hard drive costs $140. A 64 meg SDRAM costs
about $60. Because of the modular approach of Linux,
the two toolkits together take up only a few meg, and most of
that is just prompts and icons.
> > The result was quite predictable. It would be like inviting a
> > 16 year old kid to a junk yard and telling him to "build yourself
> > a car". To help him out you give him a couple of box end wrenches
> > and a couple of screwdrivers.
> >
> > > Indeed is is what some people *need* in order to use Linux.
> >
> > Absolutely. And much of what is needed to provide this for that
> > type of user is a community large enough and committed enough to
> > achieve the kind of "critical mass" that reduces the number of
> > incidents of first-time would-be Linux newbies simply attempting
> > to "download off the web and do-it-yourself".
>
> Yes and I think that resoncibility lays at the door of either easy to
> install distributions or more importantly pre-installed computers.
Probably the easiest to install have been SuSE and Caldera. They
have a knack for "getting it right the first time". But more
sophisticated U.S. users want Mandrake for workstations and Red Hat
for servers. SuSE is the "luxury car" of distributions (install
"everything" takes 6 gigabytes) but many Linux machines are still
running in dual-boot on 1 gig partitions.
> > It can be done. But I recall some of the frustrating experiences
> > of my own that could easily have deterred a neophyte Linux user.
>
> Myself as well. I chose slackware quite a few years ago because it
came
> on the back on a library book. Not the best choice I could have made
and
> I ended up very frustrated. I just chose not to give up. I had caught
> the "Linux bug".
Congratualitons. I actually had to phone order my first copy of SLS,
which was the first Linux distribution on CD-ROM, and their screen
saver displayed "Avoid the 'Gates' of Hell, Use Linux". My immediate
supervisor (who was a Microsoft bigot and proud of it), didn't
appreciate that very much. The Microsoft consultant got a big kick
out of it though. What he thought most amusing was that I was running
a Sun console on an 80386/16 box that was headed for the dumpster.
> > Much of this is getting easier. The OEMs are now offering Linux
> > integrated into the PC/Console. Many Linux server makers are
> > making it possible to configure using a simple web browser interface
> > (Linux or Windows, take your pick). The "fine tuning" can be done
> > with LinuxConf, DrakeConf, YAST, or similar GUI based configuration
> > tool. In many cases, the tool is even easier to use than the
Windows
> > Control panel (since the tabs are sequenced in the order of the
> > dependencies).
>
> Yes. But as discussed in the thread, many of these tools bugger up
your
> existing hand-edited configurations and this is just too poor for
words
> - especially if you've spent some time getting things "just so".
True. OTOH, many of them will actually parse your settings and include
your settings as the defaults.
One of the things I like about text files is that you can tar xvzf
everything to /home/root/etc.tar and have a nice back-up system.
> > > Microsoft clearly stole their ideas
> > > from Apple and Apple clearly stole
> > > their ideas from Xerox Parc and Xerox PARC
> > > created their ideas from some
> > > of the finest scientists of the time.
> >
> > And most of the Xerox PARC scientists including a number
> > of affiliates in Rochester/Webster NY were actively involved
> > in the X11 project as well. The "virtual Window" desktops
> > were a Xerox invention which was given to Athena/X11 but
> > excluded from Microsoft and Mac. (Microsoft has something similar,
> > but not quite the same).
>
> Did they all go back to MIT?
No, but Kodak, Computer Consoles, and Xerox brought in about 400
MIT graduates for UNIX related projects (Kodak Image Management
System -KIMS, CCI Directory Assistance System (DAS03), and
Xerox executives were furious with Apple for not paying royalties,
so they contributed to Athena via the MIT graduates.
In the early 1980s, Rochester NY was one of the biggest hotbeds for UNIX
outside of California. We also seemed to exchange employees a great
deal depending on which phase of a project was being implemented.
> > Many of the developers on NeXT contributed to AfterStep. Many of
> > the concepts and even people on CDE were integrated into KDE.
> >
> > > It is not a dirty word, but a goal
> > > to achieve. It is not Microsoft or Apple,
> > > it is just good computing.
> >
> > Agreed. What I find somewhat tedious is your insistance that
> > Linux is so completely lacking in this domain that it couldn't
> > possibly be as good as ANY version of Windows or Mac.
>
> No I did not say that.
But you seem to lack the understanding of the design principles
behind the Linux/UNIX GUI environment. This whole "one look, one
solution" would make sense if you were employed by makers of Qt,
but it shows ignorance of the UNIX history of failed attempts at
balkanization.
The fact is that diversity has actually saved Linux from a nasty
"crash and burn" several times. When Red Hat put out 5.0, there
were some serious problems, but Caldera, SuSE, and TurboLinux
picked up the ball long enough to keep Linux sales from dropping.
It barely affected the growth rate.
> You are only hearing what you think you want me to say.
True. I apologize.
> Linux should be *better*,
Yes it could. We could have MORE diversity, MORE competition,
and better support for existing standards.
Dictating a single-vendor desktop solution doesn't fix the standards
problem, it makes things worse. The key vendor starts deviating from
the standardds so badly that that standard has to be dropped. Linux
gets a great deal of support from ISVs who were driven out of
Microsoft's
market by a GUI toolkit developer who decided to take over everything.
Many developers create apps in GNOME because Qt costs $6000 per
developer for commercial use (which, from my reading, includes
consulting).
Remember, most of this stuff is still within the framework of
PERL, Python, or TCL. Each is quite trivial.
> but attitudes MUST change from within
> the community. The point of my thread was not to convert the
> unconvertible, but to suggest if Linux is to be on the desktop then
you
> must examine how this is going to be achieved. Linux should find its
own
> path.
I agree with your proposal in principle, I just disagree with your
implementation. It isn't that you're wrong, you are seeing the
world of a competitive software market for the first time.
The fact is that Linux is much like the deregulated telephone company.
If the carriers and telehone makers had suddenly decided to "do their
own thing", the entire phone system might have broken down.
Instead, the CCITT and ANSII committees tried to standardize things a
bit.
When they tried to replace TCP/IP with CLNS/TP4 they also raised the
price of "admission" to nearly $60,000 in 1987 dollars. The UNIX
community submitted proposals to IETF showing how to route IP packets
over CLNS (Frame Relay) and keep all the benefits (and low cost) of
TCP/IP.
Rex Ballard - I/T Architect, MIS Director
Linux Advocate, Internet Pioneer
http://www.open4success.com
Linux - 42 million satisfied users worldwide
and growing at over 5%/month! (recalibrated 8/2/00)
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
