Re: [tlug] RE: User-friendliness and windows and Linux

["Stephen J. Turnbull" <stephen@example.com>]

>>>>> "jb" == Jonathan Q <jq@example.com> writes:

    jb> As simple as it is, a lot of people find getting an Internet
    jb> connection working in Windows to be really hard

Yup.  I resemble that remark.

I have had some good experiences with Windows networking where it
"just worked", but they are equalled by the number where Linux just
worked (except in Red Hat over-the-net installs where I'm 0/3 ;-).
The need to reboot (even in NT, no?) to do the equivalent of ifconfig
(or Ctrl+Alt+KP_Plus, for that matter) worse than cancels out any time
saving even comparing best case to best case.  In the worst case,
unfortunately it's often true that the closest thing to a diagnostic
tool available at that stage is the Setup program---but it's borked,
and that's why I'm having problems.

You know, it's funny.  It's easy to come up with a script that will
get a bare-bones PPP connection up and running on almost any Linux or
*BSD box with a working modem.  I suspect it's not that much harder to
write a Python (or Perl, if you must) filter to identify a bunch of
likely problems, including with the modem, that the user can report to
telephone support.

So why is it that ISPs can't get a network Setup program that works on
all Windows boxen, given that the modem driver support is much more
complete and they spend so much attention and resources on them?  And
why don't they do something minimal for *nix ppp?  It really wouldn't
be that hard to come up with a one-page-cheatsheet for pppconfig,
would it?  (Honest question, jb is likely to have an answer.)

    jb> Microsoft is trying to address those problems, but they
    jb> haven't gone away yet.

Really?  As far as I can tell, their idea of "address" is "push the
price of incompetent MCSEs to zero by flooding the market with them."

The "Help" that came with Windows and the IBM Windows drivers
addressed problems like "is the computer plugged in?", but not even
"is the PCMCIA card fully seated in its socket?", let alone real
issues like how to configure a network, what the semantics of the
"network name" is, or what the "signal level" really meant.  No theory
of operations, either, just "programmed learning" style "if yes, goto
page 23, if no, goto page 52" "diagnosis".

[The IBM manual for configuring the "advanced capabilities" of the
gateway, OTOH, looks to be written by somebody who actually read
Cheswick and Bellovin and hangs out on comp.risks.]

It's always been possible to diagnose and usually fix Linux networking
problems within 10-20 minutes, while I often find myself booting Linux
to diagnose Windows problems.  Windows "Help" always assumes that the
used has hosed herself, so it's useless in diagnosing real problems.
(Now that's really "friendly": "Hi!  We both know that _you_ are an
idiot, and _I_ am here to protect you from yourself!"  Howcum it's OK
when a talking paperclip or iyarashii iruka says that, but ESR gets
reviled for http://tuxedo.org/~esr/faqs/smart-questions.html?)

Certainly, many users would prefer to have Tech Support come down and
fix it for them.  But many others (including my mother!) would
actually look up a theory of ops if it was written in terms they could
understand.  My mother now understands that it's not easily visible
when a computer cable is too loose (cf a power cord, where if you
can't see 2/3 of the prongs you're probably OK), and that she can ping
the ISP gateway to see if it's her connection (which justifies a
reboot or a phone call to one of the kids) or the ISP.  The latter
took a couple of pictures to explain.

I really don't see any effort in that direction on Microsoft's part.

    jb> (we know the installer won't do that for you; it'll install
    jb> it, but not in a best-accepted-practice way) and properly
    jb> locked down,

How is this the reverse of Microsoft?  Sure, if you don't provide any
real services, you have nothing to lock down.  But when they do
provide services, hoo boy, look out!  With the exception of those damn
Apple printers that grab random IPs on the LAN, _all_ of my bad network
neighbors are Windows boxes.  There's one right now filling my logs
with UDP packets source 127.0.0.1 dest 255.255.255.255.  My God!

"And now for something completely different":

    jb> but after that it will just work.

:-)

    >> would depend on age as well, 15 years ago we didnt have
    >> Windows). Windows will help raw newbies to grasp a lot of
    >> concepts

    jb> Usually, they just use it.  The concepts go totally unnoticed,
    jb> for the most part.  In fact, the whole point of click-drool
    jb> GUI is to keep people as far from concepts as possible, is it
    jb> not?

Of course not.  At least half of it is to get them as close to the
concepts _they_ choose to work with as possible.  In some cases those
"concepts" leave me wondering if a permanent vegetative state can be
called human life, but I suppose that's my deficient sensibility.

The problem is not that zombies can use click-drool software; that's a
good thing if it makes them productive.  Not to mention that some
highly intelligent and otherwise very productive humans become even
more productive if the software gets out of the way.  The problem is
that the click-drool software typically pollutes the environment.
When installed on a box connected to the Internet, even only for
email, it can play an important role in creating havoc.  And it does
so equally well when wielded by an HIAOVPH.

Whether the HIAOVPHs belong on TLUG is a different issue.<wink>


-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
              Don't ask how you can "do" free software business;
              ask what your business can "do for" free software.