Re: Win32 port

["Anthony E. Greene" <agreene pobox com>]

At 03:26 2000-06-07 -0500, Cory Watson wrote:
>At 09:18 AM 6/7/00 +0200, you wrote:
>>Most computer users just need an appliance that does the one or two things
>>they need without a lot of fuss. Linux can easily do that while hiding the
>>complexities. There is no need to come up with a special distro that
>>reduces the power of the system. It's just a matter of configuring the end
>>user's environment.
>
>No need?  Linux isn't going to make it onto the desktop without something 
>like this (unless someone manages to create some all-knowing graphical 
>config tool or installer), and you can quote me on that.  

I repeat: there is no need for a special distro if you just want an easy
end-user environment.

All that needs to be done is to have someone preconfigure the desktop, just
like what's done for Windows PCs. Even Windows PCs come with tools like
ping, tracert, telnet and command-line FTP. But somehow the existence of
these tools never manages to confuse Windows users. They're there, but not
immediately visible. What's the point of omitting them? What does the end
user actually gain if these tools were omitted? Nothing.

For some reason you seem to view this as an "either/or" proposition. Why?

Using your argument, these tools should be omitted from Windows because
people don't need them. But I've never seen a Windows user whose experience
would have been improved in any way by the omission of these tools. I say
the same thing is true for Linux distros.

Omitting tools that aren't visible in the first place does not improve the
end user's experience.

The point is to improve the accessibility of the system for non-gurus. That
just does not require that things be omitted.

If you want to talk about security, then fine. There are things that should
be done diffrently for a desktop than for a server. But you don't need a
new distro for that. Look at the workstation/server/custom choices RH
gives. We can quibble about the exact choices made for each install option,
but the point is that this can be done within the context on existing distros.

>People don't need 
>Linux.  They need Windows for the exact reasons you stated (one or two 
>thing without ... fuss). If you provide them with a nice, minimal Linux 
>install that is very tailored to running X right off the bat (maybe gdm or 
>something), and the proper GNOME config, you could provide them with an 
>'appliance' as you call it.

People with these needs do not install operating systems. They buy
preconfigured boxes and plug them in. Linux in this situation can be
configured to provide exactly what you describe -- without gutting the system.

>>If the end user likes to tinker, then the last thing you want to do is give
>>them a crippled system. Give them all the parts and let them use what they
>>like and ignore the rest. Sooner or later they will grow into the system,
>>but not if it's missing so much that there is no room to grow.
>
>But why give the user a bag of tools when they only need a 
>screwdriver?  Sure, some of them will happily make use of the extra things 
>provided, but for others they will only collect dust, or even worse, cause 
>a problem.

GNOME on Linux provides a hidden toolbox *and* a screwdriver. The
screwdriver is all most folks need, but if they decide they need other
tools, the toolbox is still there, sitting in a corner, not causing harm.
Just pop in the CD and run your favorite GUI package manager, just like
"Please insert your [DistroName] CDROM". I just don't see a problem with that.

>Don't get me wrong, I'm not fussing about this, I love my Linux box the way 
>it is.  But I know that we won't take over the world (we don't have to I 
>guess, but then theres no motivation) unless we start getting people who 
>don't give a rats ass about gtk+ or GNOME, they just want a nice 
>desktop.  I'd simply love for people to have a _choice_ without having to 
>have this big linux distro... why install redhat just for a desktop, why 
>not just install the desktop?

The end user just does not perceive this bigness you keep referring to.
Just consider the user interface of NT4 vs Win9x. Do you think for a minute
that the complexity of NT4 somehow confuses a Win9x user and prevents them
from having an easy time sending email or printing a document? Of course
not. The interface is essentially the same. All the complex stuff is just
not visible.

Do you just want to same disk space? It's not like the user ever sees this
stuff or has to deal with it in any way. Even if they install the OS
themselves, they're just sitting there watching the progress meter. There
is no additional effort required to watch the install script as it installs
300 packages instead of 200. It may take a few minutes longer, but it's not
*harder*. I guess if you want to say "waiting 5 minutes longer is harder",
then maybe so. But as I said before, the target market does not generally
install their own OS anyway.

We need to work at making some things easier. Spending time and effort
trying to figure out what already hidden capabilities to omit just does not
improve the end user experience. The only thing it accomplishes is to give
the user a system that looks and acts like it always has, except it lacks
that capability to grow with them.

If I go home and delete Apache, PostgreSQL, wuftpd, and whatever else from
my home machine, my wife will not see any difference when she logs in to
check her mail. So what's the point? I've just improved her experience how?

Instead of spending time doing that, if I go home and add diald or
configure pppd for demand dialing, then she gains something. She gets to
connect without a lot of fuss. That's an improvement. In comparison, the
deletion of the other packages was just a waste of time.

We need to concentrate on what we can add, not what we can take away.


 Tony
 --
 Anthony E. Greene <agreene@pobox.com>
 PGP Key: 0x6C94239D/7B3D BD7D 7D91 1B44 BA26  C484 A42A 60DD 6C94 239D
 Linux: The choice of a GNU Generation.