Ok, first of all, alot of the things that make Windows so "easy" to use
and Linux so "hard" are things that involve superuser acces to do. Alot
of the things you ask would lower the security of a system. However, if
we can afford to saccrifice some security for user friendlyness (or in
this case, user Pampering!), or at least build some workaround where the
regular user has a good deal more power over the system, I believe that
we can indeed build that version of GNU/Linux (or maybe not even GNU,
but certainly prefferably so).
> My thought on this is: what would be included in a Desktop Distro? I mean,
> stripped down to the bone and then built back up. I realize that the overall
> community has had some not-so-good experiences with this, since we're used
> to Linux doing everything.
Yeah yeah... we've gotten Lazy. Actually, I prefer to think of it as
"We've expanded to the point where we can afford the luxury of not
having to know everthing ;)
> But I'm thinking about a Linux distro that
> wouldn't even look like Linux. I would do all that Windows currently does
> (except crash! :-).
Or, hopefully, get viruses. But then again, that would depend on whether
or not we gave the user write privilages to the excecutables. This is
one of the biggest security flaws in a run of the mill windows system.
> You would walk in to someone's home, and they would show
> you the $300 1.5 Ghz system they bought including monitor (it's so cheap
> because the OS is free!) and how they go on the internet (You open KDE's
> browser, and your computer knows enough to start kppp to get you connected
> to the net, all you have to supply is the username and password).
Actually I believe that this can already be done if you set the web
browser icon to give off a specific sequence of commands like
"kppp | netscape http://www.slashdot.com"
(at kppp, they would have to press enter. I believe that there is
already an option of KPPP to keep the password. By lowering security a
bit more we could re-write kppp so that it can be started and initiated
by a simple double click of the icon that it's attached to). Ofcourse,
you mean a prepackaged system that can do this. Heck, it could even be
made up with a script and distributed via DEB or RPM. But we're taling
out-of-the-box here.(I'm currently working on a script that, when
finished, will allow a user using GNOME to setup their interface in a
totally user-pamper fashion. I'll send you a screenshot of the model, if
you'd like). This should be included
> They fire
> up their Corel WPO or kOffice (since I'm doing the KDE theme here, but any
> GUI/WM office suite would do) and type up a paper in support of the local
> PTA. Later on, they get an email saying that DSL/Cable Modem service is
> available in their area, and they jump right on it. The company sends a self
> install kit, and our computer "expert" plugs in the DSL/Cable modem to the
> USB port (yup, USB) and the computer recognizes it right away and now you're
> surfin' at 40 Kpbs or more...and recognizes that the telephone isn't
> necessary to get on the net.
Ok, so what you're talking about is, again, alot of automation which
would require some lowering of security. A the time, I'm not sure that
the kernel has the capabilities to simply pick up a USB device and run
with it (though I'm sure that we can modify Kudzu to listen to the ports
and when something tries to talk to the port it configures and
autostarts the device). But this is getting over my head on this subject
and is general speculation on my part. You will have to forgive my lack
of knowledge (see above ;). Now, as for testing whether it uses the
phone or DSL is a matter of speed in booting. Obviously you can't
install hardware while running the system (at least not us who are faint
of heart), so obviously you would reboot. What you are suggesting is
that the OS pick up the new device, configures it automatically and sets
the connection automatically (personally I would have it ask the user if
they would like this to be made their default connection. There's
nothing like the power of choice). This would be very realistic if
distros included dirvers that manufacturers would make for Linux for
their hardware (which, unfortunatelly is not generally the case).
However, this then would get into the subject of initializing an icon
with the kppp command. Instead of that "kppp | netscape" command, you
would just have "netscape" as something of a keyword (keywords would be
anythign related to outside communications that could be initiated by
the user, such as Netscape, or gFTP or Lynx...). When that keyword was
put forth, a script would be initialized to see what picks up that
connection, whether it be DSL, DHCP, Static Network, PPPD, whatever.
This script would ask the OS "what device should I use to connect this?"
if there have been any changes to - say - a log file that it keeps of
how connections change. The OS answers back "Well, this is the default
(now DSL), so here's the path to the connected device. This daemon
should take care of your talking to it." The daemon, of course, would
now have been initiated when you did the install and been set to
initialize every time the computer starts up (every 9 months or so,
after an upgrade). Now, this system is already in place, at least the
basic idea (I just click on my Opera icon and go on the net because I
use a cable modedm with DHCP as the connector). But what you suggest is
that these daemons be started and stopped automatically without the user
telling it anything (like when I unplug my ethernet card and then plug
it again. it should stop the daemon and then start it once more when I
plug it back in. But then again, this would also mean that this scrpt
wpould try to detect whether you were connected or not BEFORE the prgram
gets to know this.
Does this make any sense?
>
> Our friend then goes to work the next day and sees something completely
> different, but his company just saved $5000 on new systems because the OS is
> free! :-) All the machines at work easily connect to each other with a very
> minimum of fuss. All you need to do at work is file sharing, some
> spreadsheets, keep up with putting new contact information into the
> database, write a few emails to his co-workers in Ireland...general office
> stuff...
>
Alot of this stuff can already be done, but in involves a sys admin
making it pamper the user. You suggest doing the out of the box. Again,
that can currently be done, provided everything is set up right and
companies - who would use an open standard in their protocols - set the
scripts and daemons to be initiated correctly.
> Our friend doesn't need most of the things Linux provides. So, based on this
> scenario, what would go into a Desktop Distro?
>
ALOT of daemons. It would KILL booting time, but everythign else would
be automated :)
I know it's not much, but I hope this gets a start. Alot of what yu
describe can be sone by daemons. Some of it can't yet be done by daemons
because of hardware incompatabilities and you might take a hit in
performance if daemons are looking for devices that don't exist (perhaps
erroneously).
I hope this get you a start on what would go into something like this
and hopefully will spur some discussion.
Norb
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 20:38:08 EDT