[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
State of the distribution
At this point we cannot deploy a complete distrib due to lack of disk
space. What we have is an upgrade kit allowing to go from RedHat 5.1
to Independence 0.0.0.1. It allows testing, proof of concept, CD
building but nothing more.
What does Independence has Redhat does not have?
1) ADEQUATE NETWORKING: People accessing the Internet with modems have
ever been treated as second class citizens in most distribs.
Consider the fact that RedHat installation only provides for LANS,
that in the manual the chapter about PPP configuration
For a home user networking is not just a nice thing. It can be his
only source for help. Provding him with adequate tools for mail,
news and IRC can make the diifference betwween success and failure.
While curses based software can be necessary for emergencies (in case
X does not work) for day to day use it will give a bad impression to
people used to GUIs. In addition we must not assume the user lives
in a country where phone is cheap: connections must be made short.
We get this by providing offline readers for mail and news and a
cache ftp and http server (wwwoffle) who is in our opininion superior
to squid for home use and simpler to set up. Small companies could
want servers intead of offline readers but for them software like
sendmail or INN is an overkill. That is the reason we provided qmail,
leafnode and Newscache. I am not too happy about providing qmail and
I think in replacing it by Zmailer in future releases. As a news
server leafnode is simple but far too rigid so we have to provide
NewsCache in addition. This is still an area I am unhappy (that
makes us three news servers if we count INN) I hope we will find one
combining Leafnode and NewsCache qualities: could that be Leech?
Shortcomings: We still don't have a half decent X-based Ftp client, we
still have not integrated PPP into the install, we still have not
integrated notifying the servers when PPP is up in order they ship
traffic: the cron based mail and news traffic does not work in
dial-up sites. Of course we don't have our own manual. :-(
2) GUIs: We have a plug and play KDE integrated to the distribution.
If you uninstall KDE (it is pretty heavy) you find the classic
FVWM/Afterstep user interface you had in RedHat 5.1. In addition
we have Enlightenment although it is not really integrated (E
cannot be made the default WM due to its weight). In addition
installing fonts, font servers and toolkits needing resources like
neXtAw is transparent in Independence.
Shortcomings: The plug and play system is far too rigid. Only way
to switch between KDE and Gnome is unistall one of them. That
precluded putting both in the distrib. In addition the Gnome we
have in the site is old (will try to replace it by the version
shipped today. I want we keep neutrality on GNOME versus KDE and
in addition I am disapointed by the fact X+KDE is nealy 32 Megs. I
hoped KDE would thin for 1.0 due to removal of debugging code and
inted of this it fattened still more. Matt Ehrlich saying
yesterday in linux-center that KDE is lean must have been joking.
We haven't solved the problem of bad fonts making xdvi, gv and xpdf
providing nearly unreadable ouput.
3) "REAL LIFE SOFTWARE": Normal people are not mad hackers whose life
centers around their C compiler. They have non computer-related
hobbies, pay taxes, have moms or children. They need different
software. We tried to include a sample of sotware for real life: a
small and attractive database for managing collections or a wine
cellar, an UI for elmectronic cameras, software for Astrology,
personal finance, WYSYWYG word processing. Unfortunately we don't
have a half decent Spreadsheet we can put on CD (there is WingZ in
the NoCD section).
Shortcomings: We need more of it.
4) ARTISTIC SOFTWARE: Here is the software for fascinating people
specially those with artistic tastes.
We include a "Maximal" Gimp. One compiled with every option turned
on. We have a (bad) RPM of experimental plugins. I wanted to
modernize and improve it but this week I learned a new GIMP is
imminent so I am waiting. We also have a Modeller (moonlight), and
a ray tracer (povray). I have heard of other modellers we could
use but could not review them. We also have some sound
composition and playing software.
Shortcomings: There is a site for 3D software for Linux and one for
sound. They have material but I cannot explore them and still less
try what they have.
6) GAMES: What other use for a computer?
Mirrormagic (small facintaing Amiga game), xkobo, FlightGear (Flight
simulator) but we need more of them.
5) THINGS NEEDED:
-Doc. The best software in world is useless if the user is using
other software. If we don't document then people will use the user
hostile but traditional sofware they heard about in nesgroups.
-More Doc: It is diificult to find what you need in the general Linux
doc. Also at times people use other systems because they don't know
Linux has software for the job. That is why I want to put Woven
Goods in the dsitrib. It is more attractive than classic doc and it
also includes a database of available software.
-Another attitude to the doc: RTFM is a brain damaged concept. It si
based on the paradigma of the person having had Unix and system
administration before being shot in anger by the system. But for
spreading Linux we must think in people who have to manage _all_ the
box from minute one: they don't know where is the doc or how to look
at it, perhaps deadlines to meet and they have real problems to
solve. In addition people whose interest in life is not computers.
When there is a FAQ intead of flaming the poor newbies we should
realize that in fact there is a FEP (Frequently Encountered Problem).
That means a problem to be solved or at the very least having the
info popping under the user nose. Many persons ask why X works bad
when the X config software could have flatly told them: "Your card is
not supported. This server is old. Look for an upgrade".
-A manual: The RedHat one is not comprehensive enough and in addition
sometimes it is completely unadequate (see above for the PPP example).
-A better install: The importance of installation is far too
exagerated. Real problems begin after that. However I would like:
the user being assisted for partitionning, sofware selction according
to use (home, server, hacker), PPP networking
-Different user interfaces for different users.
-Translations: Outside English speaking countries, most people are
not fluent in English.
Jean Francois Martinez
Project Independence: Linux for the Masses