
Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [tlug] How revolutionary is M2?
- Date: Mon, 01 Mar 2004 12:25:06 +0900
- From: "Stephen J. Turnbull" <stephen@example.com>
- Subject: Re: [tlug] How revolutionary is M2?
- References: <20040229010646.GI18439@example.com><opr3306nqe0fabl5@example.com> <20040229031050.GL18439@example.com>
- Organization: The XEmacs Project
- User-agent: Gnus/5.1006 (Gnus v5.10.6) XEmacs/21.5 (celeriac, linux)
>>>>> "jb" == Jonathan Byrne <jq@example.com> writes:
>> up-in-arms-over-using-linux-at-home wife. It's funny to watch
>> her open open-office and try to drag the link into her mail or
>> copy it so she can paste it there. I know those are Microsoft
>> tricks and the things she
jb> I think that may work between Koffice and Kmail.
On right-press, Mozilla brings up a menu which allows you to choose
the content-type (technically, the X11 ICCCM selection target) of the
selection. This works fine with XEmacs, although it's not DND.
We made a stab at getting DND going in XEmacs about 5 years ago, and
basically it's been bitrotting since about three years ago. The
problem was that we had to support a bunch of different, occasionally
incompatible, protocols, and things that would be working on CDE would
stop working for programs that used Offix, and MS works well enough
for the primary developer so if anything has changed there since 2001,
he hasn't added it, but of course the low-level implementation is
completely incompatible with X11 mechanisms as well as CORBA.
Has this Tower-of-Babel situation changed since then? I would expect
that the K crowd and the G version have as usual failed to agree on
this, but would love to be pleasantly surprised.
Berlin had this pie-in-the-sky idea of using CORBA as the (virtual)
wire format, but I haven't heard anything about Berlin in several
years, have to wonder if anything will ever come of it.
>> financial documents but I somehow think mutt is a little too
>> much to ask her to do - even thought it may be superior.
jb> Superior always requires the question "Superior for what?"
jb> Mutt is superior for my needs, but it may well not be superior
jb> for what she does. We like to think of Linux as being
jb> superior to any version of Windows, and for most/all of us for
jb> most/all of our needs, it is. But there are certainly people
jb> in the world for whom Windows remains superior. My wife, for
jb> example. She neither knows nor wants to know anything about
jb> computers, she just wants to use them as a tool.
I dunno, Jonathan. AFAICT for people with those needs, Windows and
the Mac are adequate, and Linux could be adequate, too, if it weren't
for the need for bug-compatibility with MSFT Office on MSFT Windows.
Mac users here regularly apologize for lack of compatibility of MSFT
Office on Mac-generated files with MSFT Office on Windows. It's not
that the most commonly used features are that hard to implement.
As far as I can tell, though, email itself is not one of those, and
there are plenty of email clients (including ones based on Emacs) that
would satisfy 50% or more of the innocent users with a little training.
Of course, Windows does it with basically no training once you grok
drool'n'drip. I've always found it curious that something as
obviously winning as an audiovisual menuing system (every file should
be allowed to have _both_ a bitmap thumbnail and a 3 second audio clip
for menuing purposes) has not attracted hacker attention. Maybe Linux
will get that when extended attributes for file directories become
common. (Maybe I should patentleft[1] the _multi_media thumbnail idea.)
Unfortunately, actually collecting all of the needed audio and image
icons seems to be a proprietary job, at least if you want a fairly
comprehensive collection in real time.
jb> People who are at, and who wish to remain at, that level
jb> shouldn't be using Linux without the supervision of an
jb> experienced user to take care of the system for them. Of
Try a Sharp Zaurus. All the opacity you could want; the notepad tool
can't even find a README without the .txt extension. vi, of course,
has no trouble. :-)
jb> course, we can make the same argument about Windows or even OS
jb> X, but it's more true about *nix than it is about those
jb> systems.
If you take the "should" as an ethical imperative, it is _less_ true.
When was the last time you got spammed by a Linux virus? When was the
last time you had to reboot because an app got wedged? And I don't
know any organization which pays people to run virus scanners to
filter Linux viruses.
It's actually much easier to provide a socially responsible *nix
distro than one based on Windows or maybe even the Mac. All that's
really needed is one that comes with a firewall that defaults to
letting nothing in and nothing out, plus a simple script that enables
a short list of commonly needed services as a package:
Basic "out" package:
HTTP and HTTPS globally
(passive) FTP globally
DNS globally
POP, IMAP, NNTP to servers specified by the user, as usual
- new protocols for service discovery are in development and can
help automate this
SMTP to servers specified by the user, as usual
- ditto
SSH globally
Basic "in" package
nothing? maybe identd?
Plus a few options to enable additional services:
"Active" FTP out globally (requires _incoming_ connections)
sshd
Incoming X connections (localnet only)
Outgoing X connections (localnet only)
File-sharing package (checkbox selectable, off by default):
SMB in/out
NFS out
NFS in
Coda out (personal fave sneaks in here)
Coda in
Peripheral-sharing package (checkbox selectable, off by default):
Printer
Fax/modem
Add an /etc/issue and *dm login screen that informs you of what
service packages are currently enabled and how to configure that, and
what more do you need?
OK, maybe I've only got half of what's needed for a personal
workstation. It should be proof-of-concept: a personal workstation
distro that's fairly robust to most of the exploits that cause MSFT
boxen to emit pollution that puts an untuned diesel engine to shame
would not be that hard to put together.
Note: webservers and stuff like that are a different matter. The
target audience ("Windows is good enough for them") doesn't want them,
at least not yet. People who need more than the list above also need
to learn enough that basic iptables shouldn't faze them (and of course
the script is open source---it should have a "cat me" option and be
heavily commented with pointers to applicable dox so users with the
interest can see how easy it is to do).
The problem with "Linux for the masses" is that doing it right
requires immense effort emulating the convenience (and occasionally
the advanced features) provided by the MSFT and Apple platforms, and a
fair amount of armored footgear. The social responsibility end is
actually fairly simple.
Rather, social responsibility is a real problem at the level of geek
wannabes, who wannarun their own mail host and wannarun their own web
and ftp servers and wannarun their own DNS, etc.
Footnotes:
[1] Really patenting an invention is fraught with expense and
annoyance. Although in the US at least a patent application is by far
the best documentation (because the Patent Office in practice only
searches their own database when checking for prior art), a
SourceForge project is a reasonable and very cheap way to publish an
idea, and should be sufficient documentation to kill a patent if that
later becomes necessary.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
Home |
Main Index |
Thread Index