[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Indy web site help.
> On Sat, 6 Feb 1999, Donovan Rebbechi wrote:
> > On Sat, 6 Feb 1999, Eddie Maddox wrote:
> > > I was thinking of offering to help, also,
> > > but I need the software support to make that happen.
> > I am not clear on what you mean by "software support".
> General, specific and examples:
> General: Software that supports frustration-free productivity.
> Software that is a delight to work with.
> Specific: No bugs.
> No stupid shortcomings.
> Examples: Linuxconf.
> Linuxconf seems to be a vast improvement on
> wading through /etc/* in one vc (virtual console)
> while having man pages, HOWTOs, etc. opened on several
> other vcs just to try and figure out the littlest
> things that need changing in configs. I've spent -months-
> doing that. SLIP. Then PPP. Lynx. Minicom. On and on ...
> Have you ever read the serial HOWTO? I have. Printing,
> and many others. Over and over again. On and on ...
> Not productive. Not delightful. Frustrating.
> Netscape-Composer may be a partial answer to having
> a delightful web site authoring tool.
> I'll know when we can get Indy to run X for me
> (see: Install report 5 FEB 99).
> Voice recognition.
> KDE plans to have it. Others are working on it, too.
> Maybe I won't need to hire a writer I can't afford anyway.
First of all what we _cannot_ do at this time: Make a Linux as easy as
a Mac. Because many problems come from the fact we don't have the
same kind of application software as Macs and because there is
unsupported hardware we cannot make something perfect now.
Also consider that the distrib proper has been done by one person and
that the primary building of packages was done by three persons no
moreso it is out of question to be overambitious while the team is as
tiny as that.
What we can do then?
Three things: make a realistic Linux, forget about Unix traditions and
make an attractive Linux.
1) Realism: Problems faced by the Linux user are far different than
those of the Unix user. To begin with we have to deal with a user
who has to learn by himself and meet all tasks from day one, not to
a user who gets training before having to do real work and manage
for real his station like it happens in traditional Unix. That
means that we have to learn to push the info under his nose instead
of relying in him going to the info. I implemented this for LILO
but this is the only area.
Then you need help. That means accessing the Net. And only in
Mars people have LANs at home. I would have liked to have PPP
config in the install. It will be for next release.
Having PPP configured is useless for getting help until you can send
mail and news. For that reason first thing we did was to add
_attractive_ (no curses based) clients who enable the user to do
most of the work with the phone down. This is for people who live in
countries with extorsionist phone rates. I also would like to do
something for UUCP. Some people say it is dead. Those people don't
know about third world.
Having cleaning tasks at 3am, is obviously irrealistic when the user
is paying the phone bills. A nice crontable configurator is
useless: to begin with we have to assume the user doesn't even know
about cron. There is a substitute for cron in the "additionals"
directory. It will replace regular cron in next version.
Because the user could ignore basic things and have better things to
do than configure we tried to make things as automatic than
possible. For instance the True Type font server doesn't require
user intervention for being added in the font path.
2) To hell with Unix traditions.
Unix was ever too expensive for the mass market. For that reason
vendors never tried to develop a UI for the mass market.
There are people who would shot at you if you dared to say than when
logging in you could use anything else than the shell, use another
editor than VI or send mail with another agent than sendmail.
Well I intend to make a distrib where the user could choose if he
wants the shell or not, where he could ignore the very name of VI, and
use another MTA than sendmail in case he works in a company gor whom
sendmail is an overkill.
Following Unix steps will lead us where Unix went and this was not
3) Make Linux attractive and useful:
Making Linux easy will gain it little users if we don't ship software
for people who are not programmers. I certainly would prefer adding a
program useful for people who like to go fishing than an upteemth
programming language or net diagonostic tool.
I favor "real life" software (like software for managing your
checkbook) and "artistic software" over hacker software. Of course we
At first I vetoed any addition of programming software in order to not
increase the "hackeristic" aura of Linux. I am not sure it was wise:
for one thing a good IDEs could attract former Windows programmers and
in addition libraries for creating games, 3D or sound software could
help to increase available software.
Jean Francois Martinez
Project Independence: Linux for the Masses