Re: SEUL: Fwd: Linux (potential) end-user research.

[Roger Dingledine <arma@mit.edu>]

This was a document that was sent to me a while ago. I figured I'd
forward it to the independence and laet lists, in case they can make
better use of it than I, in terms of actually taking these issues into
account.

It needs some serious work if we (seul) are going to publish it as an
"end-user example" document, but it might still be very useful in terms of
ideas. It gets better in the second half, so don't stop reading early. :)

I've interspersed comments, along the lines of things we'd want to work
on if we want to get it into publishable shape. (I don't have anywhere
near a clear idea of what it's supposed to look like at the end, else
I'd probably have tried to get it there.)

I've gone through and fixed up typos and other small things. I tried to
leave the tone intact, because I got the impression that was core to it.

Please direct followups/replies to seul-pub@seul.org.

>------- Forwarded Message
>
>Date: Tue, 26 Jan 1999 16:54:30 +1100 (EST)
>From: "dscott@acs.itd.uts.edu.au" <dscott@acs.itd.uts.edu.au>
>To: seul@seul.org
>Subject: Linux (potential) end-user research.
>
>Hi,
>
>After writing this out, I looked for the most helpful place to send it,
>and I guess you're it. It seems you're way ahead of me on my main point
>(about researching potential end users), but this piece might still help
>you to identify some issues for consideration -- issues that might not
>readily occur to highly expert computer types, such as I assume you to be. 
>
>I'm (very) new to email, so I couldn't figure out a more useful way to
>send this than to post it directly to the SEUL organisation. It would be
>very kind if you could forward it to whoever there was most likely to find
>some use for it. Thank you in advance
>
>Best wishes, and thank you for being interested in the non-elite
>multitude. 
>
>=====================================================
>Why I Am Interested In Linux, and FUD As It Really Is
>=====================================================
>
>by: a "newbie." (Maybe. I still haven't decided.)
>
>I wrote this out to begin with, to sort out my thoughts on the usefulness
>or otherwise of Linux, a topic on which I could go either way. 
>
>I finished this piece, because I looked at some discussion lists on Linux
>and it seemed to me that people were speculating or even inventing images
>of these strange, stupid non-technical users and their FUD. No one seemed
>to be asking the people concerned about anything. Because I am basically
>sympathetic to the Linux project, it seemed to me that I could do a small
>service by explaining what one of these dullards who don't "get" Linux was
>actually thinking about, and what FUD, as a reality, not a slogan, is all
>about. 
>
>My individual problems are of not the slightest interest or importance. 
>But, if Linux is to achieve "world domination," someone whose say-so
>counts for something had better be thinking damn hard about these problems
>in general. It is not possible to beat Microsoft while Microsoft retains a
>"lock" on nine out of ten non-technical home users across the world. I'll
>say that again: it is not possible to beat Microsoft while Microsoft
>retains a "lock" on nine out of ten non-technical home users across the
>world.
>
>If I can leave people with just one thought, I would say: "Stop speculating
>about the motivations and fears of non-technical potential users of Linux,
>and ask them. Do research." 
>
>
>Why I Am Interested I Linux
>===========================
>
>
>(1) I crave stability in an operating system.
>=============================================
>
>By "stability," I mean:
>
>Linux won't be "killed" by its owner or the market. For example, I loved
>OS/2, but IBM "killed" it as a personal operating system. They won't do
>anything more for it, and because it's proprietary, no one else is allowed
>to run with it. No-one has authority to "kill" Linux like that. It has a
>"bullet-proof" future. 
>
>Linux is properly constructed so as not to self-destruct. Windows 95 has
>near enough to no self-repair capacity, so when it gets itself screwed up,
>you have to back up all your data, wipe your hard drive, and start again.
>(At lest that's the only 100% solution that a non- technical user has the
>skill to implement.) I can't use Windows 95. I have my hard drives
>partitioned from C to G, with everything pertaining to Windows 3.1 backed
>up on C, and separate installations of Windows on D, E, F and G. During
>semester, I don't have time to fix anything that goes wrong, but If
>Windows doesn't get screwed up more than three times, I'll still have a
>good installation to work with. This works, but I feel there must be a
>better way, and I believe Linux is this way. 
>
>Linux is properly constructed so as not to crash. Windows 3.1 is easy to
>crash, and Windows 95 is much worse, because it crashes almost as often,
>and it takes the whole computer with it, instead of just sending you back
>to DOS for a little while. I feel there must be a better way, and I
>believe Linux is this way. 

How much experience have you had with windows 95? I always found win31 to
be less stable than win95 (which in turn is less stable than winNT). Can
you provide some more details here?

>Linux makes sense. I don't have confidence in Windows, because its design
>makes no sense to me. Broadly speaking, every program can do anything to
>anything, anywhere, at all times, and the windows\system directory, which

Actually, protected mode in theory fixes a lot of these issues. Specifically,
winNT is better about this sort of thing.

>contains the "guts" of Windows, and ought to be a fortified stronghold, is
>actually a garbage pail for any kind of miscellaneous file, and it never
>gets cleaned. I don't understand anything about programming, yet, but I do
>have enough common sense to see that this is "garbage." Anyone can
>understand this. I understand that Linux is built logically, with key
>files segregated in protected areas, dangerous actions restricted to the
>"super-user mode," and so on. I feel much more secure with this kind of
>system. 
>
>Linux is constantly peer reviewed and debugged. Windows is secret, and in
>the case of Windows 3.1 (the only kind of Windows I would run) the secrecy
>seems to serve the function of hiding what a shoddy job of design and
>programming it is.  (Compatibility is king, otherwise I would prefer
>GeoWorks.) Peer review is impossible, there is essentially no ongoing
>debugging, and Microsoft won't fix it because they want to force everybody
>to upgrade, and upgrade, and upgrade. Naturally, I trust only the OS that
>gets debugged, and not the OS that is swarming with uncontrolled bugs and
>viruses. 
>
>Proper login security, with an option for restricted permissions so that a
>visitor can't do anything that could fundamentally destabilise the system,
>is highly desirable. This goes from "stability" to "security," but many of
>my friends share house, with several personal computers to a house. People
>really hate having their personal computers used without their permission,
>particularly as the "visitor" can do absolutely anything on a Windows
>machine, even by accident.

Again, don't generalize this to all windows environments. NT is much better
about this. But of course, NT still has its bugs, and indeed it's designed
for much stronger hardware than you're likely to be using.

>Why Stability Is Key
>- --------------------
>
>I don't need an operating system with a lot of features. At least 90% of
>what I do is just word processing, and the main "features" I care about
>are those that are imposed on me as standards, for example: "You must use
>typographic quotation marks." But I absolutely, positively do need a
>stable platform under my feet when I am on the ragged edge of a deadline,
>and this is exactly what Microsoft non-peer-reviewed, closed source
>software will never give me. I don't care if my computer's screw-up can
>easily be corrected in time. There are frequent occasions when I don't
>have time to deal with such problems at all. Pass or Fail might not look
>"mission critical" to Redmond, but by G-d it looks critical to me. 
>
>(2) I want the contents of my hard drives to be works of art, whether it's
>my own work or someone else's. 
>==========================================================================
>
>I feel that Linux is a work of art by many hands, and so I am happy to
>have it on my hard drive. Guys like Torvalds and Rasterman aren't saying:
>"Screw quality.  We'll toss it out there, win market share, and fix the
>bugs later, if we have to." They're saying: "Let's do the best job we
>can." I have a really good feeling about having the work of such artists
>in my home. 
>
>I feel the opposite about Microsoft Windows 95, which seems to be the only
>"live"  alternative on my hardware. This is not art. This is a trick to
>gouge money out of people. I am viscerally hostile to Windows 95, which
>defined the Start button as my primary interface point, and then dumped
>the menus there full of advertising and dross that I didn't have the skill
>to get rid of. Even in Windows 3.1, which is mostly sort of OK, Help About
>leads to a "gotcha!" where the software company reminds you who's boss. 
>Help About should lead to Help About that program, or I should be able to
>change it so that it does. 
>
>I want to control every visible aspect of the OS, to make it beautiful in
>my own eyes. OS/2 gave me a lot of that: absolutely everything was an
>object, with an extensive property sheet for me to define. Windows 95
>takes that away: I couldn't even rearrange the items in my Start button. I
>understand that Linux is the best: it will be possible for me to choose,
>and customise, my desktop program to get the effect I want, and to get
>things to work the way I want. (Ideally, the way I'd like to do things is
>to have a directory tree showing my files (with "long," unmistakable file
>names) here, and labelled boxes of tools and programs (with unmistakable
>icons and maybe "balloon help") there, and drag this file to that tool (or
>vice versa) to do everything. Real "toybox" stuff. I assume there'll be a
>way for me to set that up in Linux, if things are even half as
>customisable as people say.) 
>
>Linux is open source, and I hope to do a computing major, so in time I may
>hope to contribute to this art in a fundamental (though probably tiny)
>way. 

My general impression is that this section is not 'universal' with
end-users -- specifically, they don't much care who wrote it or how or
whether it killed trees or dolphins to do it, as long as it works well
enough for what they're doing. In addition, few of them are intending to
major (or did major) in computer-related fields.

>(3) I can run Linux on my hardware (probably?), and I can afford it.
>====================================================================
>
>Suppose I thought that Windows NT/2000 would solve all my problems: on my
>hardware, and with my budget, so what? Linux is a "real world" solution. 
>
>(4) Linux will run applications that will enable me to do my work.
>==================================================================
>
>I read that Corel is committed to WordPerfect on Linux. That kind of thing
>is critical to me.  Now if someone would just write an equivalent for
>Microsoft Works for Windows, Version 3.0, but on Linux, I would be really
>happy. 

While I don't know what MS Works is, I bet Staroffice (www.stardivision.com)
has a substitute that works pretty well.

>The Internet is becoming important to me, as a student and as a guy with a
>network of friends who are all easiest to reach by email. I believe Linux
>is totally superior for anything related to the Internet. Nothing else
>comes with a remotely comparable set of (free!) tools, or, above all, has
>the "life" and future on the Internet that Linux does. 
>
>Even just having a ""kill"/stop doing that!" button would be a Good Thing,
>and I know that Linux has one. Many things that are impossible with
>Windows are automatic with Linux. 
>
>I read that SCO Unix is free for personal use. Joygasms. What kinds of
>word processing and works programs are written to run under SCO Unix on a
>486? If I need strong tools (and I do), I need Linux. 
>
>(5) I believe that having good open source software will help to teach me
>computing.
>=========================================================================
>
>The only way I learned anything about chess was to study the games of
>great masters like Capablanca and Fischer. It makes sense to me that to
>understand software, I'll have to study the programs of great masters like
>Torvalds. 
>
>This kind of beauty, based on appreciating an underlying logic, won't
>yield to superficial study and "once over lightly" treatment in lectures.
>I'll need authentic, top quality source material, working in my own home.
>Maybe (and "maybe" is not good enough here) I can pass without it. But can
>I learn art without it? The idea is ridiculous and offensive. No
>worthwhile art can be learned without earnest study of the great masters. 

As above, in my experience most people don't want to learn how their
software is written.

>(6) There are -- or soon will be -- enough Linux people around to trade
>ideas and files with. 
>=======================================================================
>
>It would be easier for me, personally, if there were more end users, and
>not just hackers who obviously can't be happy to be disturbed by newbies.
>(Even help files and FAQs tend to say: "Don't ask in public. You'll only
>make a fool of yourself.") But the people exist.  Saying "OS/2" now is
>like disturbing a graveyard. 

There are many Linux users around these days. If you're looking for more
'standard' support, check out linuxcare.com.

>Fear, Uncertainty and Doubts about Linux
>========================================
>
>
>(1) Once bitten, twice shy.
>===========================
>
>I tried to install Red Hat Linux, supposedly the "easy" package, a couple
>of years ago. It was a nightmare, and completely useless. Effectively, I
>turned my computer into a paperweight, and myself into a nervous wreck. I
>tried re-installing, over and over. No dice.  As for my "free support," it
>was online, and of course "I had no computer." Things stayed that bad for
>over a month, after which I wisely gave up. "Never again." 
>
>Some minor points in the maelstrom:
>(1) As it happens, I do not carry around the chipset and exact screen
>refresh rate of my monitor. And anyone who can move as often as I have
>while never losing any documentation is a genius. 
>(2) Why should I guess what partition sizes I need? Hard drive sizes +
>packages = exactly the sort of task a computer is supposed to automate.
>The installation program should ask questions in the right order, ask
>permission, and then "just do it." 
>(3) Cylinders? At any time, but now with a hard drive that is damaged in
>some way I don't understand except that 70MB of it is now useless? "You
>jest." 
>(4) The important task of determining whether you have just done the right
>thing, and if not, how you should back-track, is radically more difficult
>on a kind of system you have never used before. 
>
>I can't help contrasting this with my experience of OS/2's installation,
>where everything worked perfectly every time -- no fuss, no tears, just
>results. And this was with all the "difficult, advanced" options, with
>Boot Manager, HPFS and so on. If I had found OS/2- native software to do
>what I needed done, the way I needed it done, I would still be using OS/2
>and pitying the suffering masses. 
>
>I read lying "booster-talk" about how easy Linux is to install, and I just
>laugh. You'd better believe I warn my friends about the truth. If luck
>does not favour you, installing Linux can easily become a species of
>computer Hell. 

Which version of redhat were you trying to install? Have you ever tried
a different type of Linux? Have you ever done an install, from scratch,
of any of the Microsoft operating systems? How did those compare?

>(2) Linux may not work on my hardware.
>======================================
>
>I never did get Linux to recognise my CD-ROM drive. Maybe that's because I
>was doing things wrong. After all, there were lots of things about Linux I
>never got to work properly, starting with LILO, the start-up program. But
>maybe Linux just won't tolerate my CD-ROM drive, or my sound card, or
>something. I've talked to other people who also couldn't get Linux to
>recognise their CD-ROM drives, and gave up. I read the compatibility
>lists, and all my stuff (Panasonic CD-ROM, true Sound Blaster and so on)
>is there, but maybe it is somehow set up wrong, or connected in a way
>Linux doesn't like, or something. I could pay someone to crack open the
>box, and go on a random fishing expedition, at $60 per half- hour, looking
>for "things Linux might not like." Then again, perhaps not.  There may be
>no solution, given that I can't afford a new computer. 
>
>My CD-ROM drive is my only music player, so one of the things I remember
>about my efforts to install Linux is that my flat fell quiet. 

This is a good point -- no installation that I've seen has done that
"last little step" of setting up the right permissions on /dev/dsp --
indeed, the Debian 2.1 installation I recently did totally forgot to do a
'MAKEDEV audio', so I had to point cdrom at my hdc, change the perms, etc.

>(This is the right time to think very seriously about this, because
>"disaster just struck," and I have to get my CD-ROM repaired, perhaps
>replaced anyway. This may be a critical decision "in the dark.") 
>
>I also have only two hard drives, of 470MB (after damage) and 200MB. Linux
>hates multiple drives -- it doesn't even like to read a floppy disk, in my

Actually, Linux deals very well with multiple drives, provided you did
the right thing in the 'fdisk' section of the install. What you want to do
is set up partitions on the various drives, and then during the installation
specify that you want /usr/ to be a certain partition, /home/ to be a
different partition, maybe /sbin/ or /bin/ a third, etc. But this requires
knowing the sorts of directory hierarchies Unix uses, and it also requires
having an intuitive feel for how much "stuff" will go under each directory.
No end-user (or even windows programmer) can be expected to know this stuff
when he's starting out.

>experience -- and it wants its one hard drive to accommodate hundreds and
>hundreds of megabytes of -- "stuff."  It was way beyond me to figure out
>what I should leave out and what I might need.  According to an expert I
>read, the only smart, safe move is to buy a multi-gigabyte drive and
>install everything. I don't regard that as an option. Besides, if I want
>to install hundreds of megabytes of useless fat, Windows can do that for
>me, no trouble at all. There is no "logic,"  no economy of force, no art,
>in this approach. And I haven't got the money. 

Modern installers (with Redhat, at least) have options for "workstation"
or "desktop" installs, where presumably they make most of the decisions
for you in terms of which packages to install. (I always choose 'custom',
so I don't have much direct experience with this.)

>A 486 DX2/66 with only 24MB of RAM may not be nearly enough to get
>acceptable performance out of the graphical interface(s) and all the
>programs that rely on them. I think a well-written, light, lean program
>should go: "Snap!" Very likely, that won't be possible.  Well, it was a
>sexy machine when I went so deep into debt to buy it. Not any more. 

With a light window manager, this machine will still be relatively fast.
It's true that you won't be able to (easily) use the newer desktops like
KDE or Enlightenment, but you don't need them. Perhaps the installer
should take into account the "power" (cpu + memory + drive space) of
the machine when deciding which desktop setups to recommend.

>I also have definite ideas on multi-tasking performance. OS/2 was
>splendid. Windows 95 was grotesque. ("Kalunk, kalunk, wait while I think
>about it, wait ...  Thunk. OK, now I'll give you back your input
>devices.") If, because of my weak hardware, Linux multi-tasks as badly as
>Windows 95, or even worse, that would be -- ugly. 

While I haven't used OS/2, I'm told it's pretty good with multitasking.
Linux is at least on par with it (given that Linux beats the pants off
every other x86-based OS out there, in most benchmarks I've seen). It's
certainly far superior to Windows.

>(3) There is no cheap, safe starting option for applications software (as
>far as I know).
>=========================================================================
>
>This issue can't be wholly segregated from installation. I know when my
>operating system is properly installed, because I do simple tasks with it,
>and they work. I write: "Hey, it worked!" in IBM Works or Windows Write,
>and I print it out. I play a CD, and it sings: "I Feel Good!" And so on.
>For a non-technical user, the possibility to read a "man" page, or start
>"emacs," and in either case be baffled, does not provide the same sort of
>feedback. It's still an open question whether the system can be somehow
>manipulated so as to do a useful task or generate an output. 
>
>One of the things that boggled me about Linux is that there is nothing to
>get started with. Not IBM Works. Not Windows Write and Cardfile. Nothing.
>There's no Microsoft Works to buy. There's no "happy stuff" either: no
>Chessmaster 4000, no Microsoft Encarta, no Dangerous Creatures, no
>Microsoft Home Essentials. Nothing. You sacrifice all that, and lash out
>for your "big" applications suite, because no-one is offering anything
>more "human-sized," and because, consequently, until you get your big
>applications suite, your Linux home computer is an elaborate paper-weight. 

Project Independence (http://independence.seul.org/) is attempting to
address the issue of "no little apps", but they're still a ways from
finishing that goal.

>With the "reputation" that applications for Linux are earning, this is not
>an attractive gamble. Will Star Office or Applixware support my 9-pin
>dot-matrix printer that emulates an Epson LX800? "Do ya feel lucky, punk?"
>As it happens, my honest answer to that question is: "No." And I have no
>plans to rob a bank in order to buy a PostScript printer. 

Oof. Good issue, I don't do non-postscript printers so I'll let somebody
else tackle this one. :)

>Besides, I don't want a high-tech, advanced-technical-feature "office." I
>want Home Essentials for Linux, but I don't think Microsoft will be
>bringing it out anytime soon. 

Independence: do you have a list on your website of the sorts of apps that
you think would cover this, and why they're good/etc? It seems like a good
document to have if you don't already have it.

>If it doesn't work, what is the next move? Configuration Hell reinstalling
>DOS and Windows? Yes, that would be the next move. "Been there, done
>that," and "better the devil you know." 
>
>(By the way: I'm not mad at Microsoft in general. I think they make
>excellent mice. If they were strictly a hardware company, I'd put them
>almost up there with Logitech. But they've virtually "locked out" low-end
>users who need a small, good-quality, reliable operating system, and I am
>mad at them over that.) 
>
>I don't believe WINE is the solution. First, I don't believe WINE works. 
>Second, I don't believe WINE works. ("It's such a big problem I thought
>I'd mention it twice.") A toy that's considered impressive because it runs
>WordPad, usually without crashing, is of no use to me. A WINE that's
>almost as good as the real thing returns me to OS/2:  "Congratulations. 
>You have just installed an enormous Program Manager replacement. And by
>the way, it will cause problems for your multimedia CD-ROMs." A
>better-than Windows WINE, a bullet-proof WINE? I don't believe it. The
>future for Linux is Linux native software, or failure. More to the point,
>I don't believe anything but Linux-native software will give me the
>reliability I want. 
>
>(There is one area where WINE is valid and vital. But that comes later.)
>
> Maybe the only way to get Linux-native applications as reliable as Linux
>is to make them open source like Linux. In which case, I'm out of luck,
>because there's no indication I know of that open source word processing
>is possible, or that an open source works program is even "imaginable." 

Open source word processors are slowly progressing, but you're right that
they haven't been proven truly feasible yet.

>Ideally, I would like programs to work the way they do in GeoWorks. Along
>with your choices about new document/old document or template, you get to
>choose a complexity level from (say) 1 to 3 (and you can customise what
>each includes and leaves out, if you want to). So, if you're making a
>"difficulty of one" document, (or if you're a "difficulty of one" user),
>you never have to deal with buttons and choices that are useless to you. I
>know lots of Word users who would like to be able to do that. But it's
>only too obvious that no-one is about to write this kind of simple
>software for Linux. 
>
>As far as I can tell, the people who write stuff for Linux only care about
>impressing each other with advanced technical features, in a spirit of
>"can you top this?"  It seems irrelevant whether the people who make the
>software charge money for it or not:  Applixware, which is commercial,
>seems to have the same attitude. I think that if that's your attitude,
>it's impossible to write the kind of good quality simple software I need.
>If someone insists on making a screwdriver advanced and technical and
>complicated, he'll make a lousy screwdriver. 

You should check out a lot of the applications written for the K Desktop
Environment. Unfortunately (there's always an unfortunately), it's kind of
a resource hog and might not work very well on your hardware.

KDE supporters: you guys have support for a light window manager and
light applications......right?

>(4) Compatibility is king, and it is in question.
>=================================================
>
>Can I get the university computers to "eat" the files I would generate
>with Linux? I don't know. Can I print out on the superior printers at the
>university? I don't know. When I get an assignment that requires 100%
>Microsoft Word compatibility (and I know I've got at least one coming up),
>could I do it on my Linux home computer? I don't know.  Internet Service
>Providers generally want to dictate what software you use. (They also want
>to take your money and provide no service, in my experience, but that's
>another problem.) Will an ISP who would be acceptable to me on other
>grounds tolerate a Linux home computer? I don't know. 

Good concerns. Staroffice and Corel have some good filters (for converting
files from word format and back, for instance), and there are some other
applications out there designed to convert. But I don't know if anybody
is claiming 100% compatibility yet.

>My guess is that all of this would work out, somehow or other, after some
>trouble. But guessing is not knowing by experience. I know what I can do
>with the software I've got, defective as it is. 
>
>I do know that a whole lot of ISPs want/require you to run Win95/98 and
>Explorer. Maybe the whole Internet will divide, with the ISPs keeping
>Linux for themselves, and forcing their clients to use Windows 2000 and
>Explorer? 

Presumably they do this because it's easier to support a single operating
system, and Windows is currently the most common (and probably the simplest
to support as well, given the relatively small number of things you can
do with it). Many ISPs that "want" you to run windows actually just set up
a PPP connection for you, and then you use their software to communicate
over it. You could just as easily use any other software, including the
Linux PPP support.

>I do know that using Linux puts me at loggerheads with the Australian
>Government. Telstra is partly privatised, but still the government
>telecommunications company.  It is providing a basic email service for
>free. (And a good thing, too. This should be part of the basic phone
>service by now.) This service is provided free exclusively for Microsoft
>Windows 95 and Microsoft Windows 98 users. The Microsoft Australian
>Government software is written exclusively for those platforms, and if
>you're not running the Microsoft Operating System that Microsoft wants you
>to run, you're
>(a) about to "upgrade" as required, or
>(b) not entitled to equal services.
>As the man on the phone patiently explained to me: there are no
>work-arounds, there aren't supposed to be any work-arounds, and there are
>no plans to support any "obsolete" (= "non-Microsoft") platforms, ever. 

Wow. Yuck. Check out http://www.anatomy.usyd.edu.au/danny/freedom/ip/aidfs.html
and contact Danny; he knows more about this than I do (given that he does
free software lobbying for Australian organizations), and might have already
done a lot of research on this topic (I have no idea).

>This is a disturbing straw in the wind. When a key Government company goes
>into the business of diminishing personal choice and market competition,
>for the benefit of Microsoft Corporation, then it may become very
>"inadvisable" for small, weak individuals to buck the system. This is
>where WINE is vital. If it works well enough, it may enable Linux users to
>put up a "mask" of Microsoft conformity when the Australian Government
>requires it, as the Government may in future do on far more important
>questions than whether or not you get equal phone services. 
>
>Maybe other governments will also standardise on Microsoft, because it's
>easiest to write for one platform. When masses of ordinary people need to
>be Microsoft customers in good standing, in order to talk most easily with
>their governments, then the competition about operating systems will be
>over, as far as I can see. 
>
>Anyway, let's keep the paranoia to a dull roar. No-one is going to shoot
>me for not going Microsoft all the way. They're just going to make life
>inconvenient. But my capacity to buy my way out of inconveniences is
>limited to say the least. 
>
>(By the way: I feel this is really unfair. I am a Microsoft customer: 
>Microsoft MS-DOS, Microsoft Windows, Microsoft Works, Microsoft
>"everything." I just don't count as one, because Microsoft has dumped my
>operating system, and, effectively, me.  They didn't even maintain
>backward compatibility for Word. This does not engender goodwill.) 
>
>
>Why I Am Still Strongly Interested In Going To Linux
>====================================================
>
>
>Fear, Uncertainty and Doubt are justified, but they don't get me anywhere. 
>I can limp along effectively enough with my present system, and I would
>probably be wise to do so.  However, the down side of that course is that
>my present difficulties will never diminish.  They will only increase, as
>time and Microsoft-controlled standards leave me further behind. 
>
>If I want security, stability, beauty, strong tools (at least in certain
>specific areas), and the satisfaction of supporting people I like instead
>of people I no longer like much, I want Linux.  This would be the (only)
>way to solve my computer problems rather than putting up with them. That
>is, it would be if Linux worked for me, and I don't know whether it could
>or would do so. 
>
>- -----------------------------------------------------------------
>
>I guess this is where the file ends. I've laid out my pros and cons, and I
>still have no conclusions. 
>
>I wonder how many there are like me? Quite likely there are millions of
>us, inflating the figures for "people who've installed Linux." 
>
>If I can leave people with just one thought, I would say: "Stop speculating
>about the motivations and fears of non-technical potential users of Linux,
>and ask them. Do research." 
>
>And if that's already agreed, I would say: "Do qualitative research."
>
>- -----------------------------------------------------------------------
>
>Best wishes,
>		David Blue
>
>------- End of Forwarded Message

Thanks for the notes. Hopefully I've given you some more things to think about
from my comments, and hopefully other people will address other issues.
Please take a look at http://www.seul.org/pub/fsib.html for a draft document
that might convince you that maybe free software is the more rational choice
for companies as well as individuals, and therefore maybe Microsoft is going
to continue to lose influence rather than gain it. (Comments on that greatly
appreciated, as it's still in draft form.)

--Roger