[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
HOWTOs for the new newbies [was: procmail and kmail]
Despite my disagreements with Martin, I think he's formulated the
issue and described the need very well.
>>>>> "Martin" == Martin P Holland <m.holland [at] noether.freeserve.co.uk> writes:
Martin> On Fri, 18 Feb 2000, Stephen J. Turnbull wrote:
>> My advice is to ditch the K environment,
Martin> What on earth are you ranting about Stephen? ;-) What has
Martin> a desktop enviroment got to do with setting up
Martin> sendmail/fetchmail properly? Nothing.
Oh, I'm confused. So all those k* utilities really have nothing to do
with their Internet access and mail systems?
That's what I thought, but people seem to keep expecting them to set
up their PPP and mail systems for them. :-)
Martin> BTW, I wrote some simple (I think) notes at
Martin> http://www.noether.freeserve.co.uk on setting up fetchmail
Martin> and sendmail for my ISP that might help people who are
Martin> struggling. (If the emails I get are any guide it has
Martin> helped quite few people.)
This is exactly what I meant to suggest by "get rid of the X
environment". I meant get rid of the Fisher-Price busy-box config
tools. Even MSFT doesn't do them very well, and Heaven knows they
spend enough money on them.
The problem is that understanding both the syntax of the config files
and the limitations of the config tools is not easy, and not worth it
IMHO. If the tools work, good. Use them and forget about it---Linux
will do the rest. But if the tools don't work, you are best advised
to start from the beginning.
Martin> Def: A Home User has a linux box at home (possibly not
Martin> even on a LAN) which is periodically connected by dial-up
Martin> to the internet and is being used as a desktop workstation
Martin> as a drop-in replacement for win98/98.
I have to quarrel with the "drop-in" description. Nobody in their
right mind would use Linux as a "drop-in" replacement for Win98. What
Win98 does acceptably well, it does far better than Linux can. The
only advantage Linux has in competing with Win98 on its home ground is
that the applications, where available at all, are often far cheaper.
I don't think that adding /usr/doc/HAND-HOLDING/ to the doc tree would
keep these users in the Linux fold, no matter how well they were
written. So home users really have to be something else.
Sure, they expect Linux networking to be as easy to configure as
Windows networking. But if that's all, it ought to be possible to do
it in a script (maybe with a Fisher-Price interface). What Windows
does for configuration is not very complex.
What we really need here is a list of common functionality that users
(have a right to) expect that Linux will provide a "dumb Windows
compatibility configuration" for. PPP. Somewhat more powerful mail
service shouldn't be hard. ISDN. XFree86 still comes out of the box
with 8bpp (256 colors) standard, so you have to configure TrueColor
specifically. (Dunno about Britain, but in the US and Japan all home
users have true color.) XF86_SVGA makes extremely conservative
assumptions about what is needed to run at TrueColor depth. Printing.
(This is a crime AFAIK; I'm always using multiple "decapitated-edge"
versions of Ghostscript and typically a witch's-brew of TrueType
libraries and funky filters, so I haven't let anything like
"magicfilter" near my printer in over 5 years. But what I hear is
distressing.) WWW browser preloaded with Linux links. (Netscape will
do, Mozilla ain't up to it yet. But none of them is great in a
non-Latin-1 environment yet.) Desktop themes (window manager-
independent! Debian is working on this but I haven't tried it yet, I
think the Gnome people are collaborating with them?).
But the DWCC ought to be flexibly extensible, and not break when you
add bells and whistles that Windows couldn't handle (filtering and
caching WWW proxies, servers, procmail, etc.).
And then there is the "apps" problem. Menu-driven WYSIWYG word
processor (AbiWord looks good, though), spreadsheet (Gnumeric is
getting good reviews but is clearly still not ready for prime time),
presentation software (MagicPoint is very cool but not WYSIWYG, and
some of the best effects like swallowing other X apps and generating
dynamic graphics inside the presentation are still buggy; I dunno
about the K thingy). Too many media players (xv, imagemagick, rsound,
esound, nas, Real*, xmms, etc, etc).
That's a bit out of bounds, the relation to configuration that I see
is that those FAQs are often closely related to (ie, same poster) the
config FAQs, and second, maybe rather than writing HOWTOs for the "new
newbie", one might make more contribution by working on docs for the
apps.
Martin> In fact I would guess that the majority of linux users are
Martin> Home Users in this sense. (This is all presumption on my
Martin> part of course, any one know any hard information?)
How would you get it? But I think you are on safe ground in assuming
this is a substantial demand.
Martin> There is a distincy lack in the HOWTOs of concise
Martin> infromation _targeted_ at these particular users.
Of course. That concise information is all in k* scripts already,
where it belongs. :-)
But back to your point. Who's going to write those HOWTOs? You don't
want newbies writing HOWTOs ... where do you find authors with
expertise who sympathize with newbies who don't want to be experts?
I, for example, am working on a couple of HOWTOs related to
internationalization. But surely you don't expect me to hose my book
royalties by targeting my HOWTO at people I don't know, will never
meet as peers, and who will never contribute anything except revenue
to any projects I care deeply about?
Sure, I'm an arrogant arse; that was pointed out by my 10th-grade
English teacher. But I'm also willing to answer questions on Sheflug,
and a couple of other LUGs, because I know the people here, because
some of you have answered my questions, and because I like the company
anyway, even if I can't smell the curry from over here. What I'm not
willing to do is take those specific answers and turn them into a FAQ,
let alone a HOWTO, for people I don't know and whose main concerns
often leave me completely cold. I'm mostly concerned about admin
types, and working on free stuff for them doesn't leave me a lot of
free time.
I don't think I'm untypical of HOWTO authors in those aspects.
Martin> The info is all there of course, but you have to read
Martin> carefully, interpret, have some experience (certainly an
Martin> unreasonable expectation for new users).
(1) I didn't find it so when I was a newbie.
(2) It's not clear that having high expectations of "newbies" when
they are using an industrial strength OS is unreasonable, any more
than having higher expectations for lorry drivers than for bicyclists
is unreasonable.
Be that as it may, the "new newbie" does not have the background, as
you say, and there is a demand to be addressed here.
Martin> Since (I claim) Home Users are in the majority this means
Martin> that there is a gap.
Which I suggest is probably best bridged with pounds (sterling).
Certainly, there will be a few dedicated people like yourself writing
HOWTOs for the new newbie. But are you sure that your effort wouldn't
be better put into putting that knowledge into a script that only Perl
would ever read? The Home Users, as you describe them, don't _want_
to read HOWTOs. They want things to "just work".
So they should buy commercial distributions with hand-holding support
contracts.
Martin> I think that there is need for a dial-up-user-HOWTO that
Martin> addresses issues like sendmail, fetchmail, named, ipchains
Martin> concisely based on the needs in this particuar case (or
Martin> perhaps a sequence of mini-HOWTOs).
But how do you do this "concisely"? The parts that can be addressed
concisely are already in the main addressed concisely: in Perl and
Python. The fact that the new newbie can't read those doesn't matter
if they work, and it probably wouldn't matter if they didn't.
This is not a jab at you; I haven't read your page yet. This is a
statement of personal experience. I've participated in writing a
whole book on installing a Japanese environment on Linux at this
level. I'm not proud of it, but I don't see how we could have done
better. My coauthors are proud of the fact that the bugs in other
books' procedures don't arise on their systems using ours, and we do
provide a lot of technical background that other books do not. We
make some attempt to bridge the gap between the cookbook aspect and
the structural concepts. But (IMO) it doesn't work very well. There
are plenty of potential bugs in our procedures, and we don't (can't,
AFAICT, I've been worrying about this from Day One) provide the means
to debug them to the level of user we're targeting.
Martin> I would welcome any suggestions to make things more
Martin> distro-independent.
Well, we went to the trouble of installing various base distributions
(Red Hat, Debian, and Turbolinux are the popular ones in Japan---this
was one reason for the coauthor organization, it meant we could get
more coverage and better depth by having people run the system as
their day-to-day system), and then added several "Japanization"
distributions on top of them. We also installed 2 native Japanese
distributions (based on Red Hat and Debian respectively).
I wouldn't ask anybody to do that kind of scutwork if they weren't
getting paid for it. Not having done it once myself. But I think
that's the way to do it....
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word
"unsubscribe" in the body of the message.
GNU the choice of a complete generation.