Re: [tlug] CLI admin how-to ideas?

[Matt Doughty <mdoughty@example.com>]

On Mon, Apr 08, 2002 at 11:07:30PM +0900, simon colston wrote:
> On Mon, 8 Apr 2002 21:57:23 +0900
> Christopher SEKIYA <wileyc@example.com> wrote:
> 
> > On Mon, Apr 08, 2002 at 05:24:09AM -0700, Marlo Rocci wrote:
> > 
> > > Instead there answer appears to be "there are, but you're an idiot for
> > > being curious about them".
> > 
> > More along the lines of "you know that mixing bleach with ammonia is a really
> > bad idea, yeah?".
> 
> I think it's plainly obvious that some users don't want to have to become sysadmins in order to use (what should be) fairly straightforward functionality if implemented in a more user-friendly way.  OK, maybe you wont be able to do as much through a user-friendly interface, but you can certainly do a lot.  Elsewhere Stephen Turnbull described writing such an interface as a "business opportunity" and there are several projects trying to cater for this need.

At this point it isn't a realistic goal and it really contradicts the spirit
of the community.  Unix is a powerful system for those who are willing to 
learn to use it.  Most people in the comunity have no interest sacrificing
power and flexibility for 'ease of use'. As a result alot of design decisions
are made that don't cater to the 'ease of use'.  There is certainly a market
segment there but I think install and forget is more of the goal there.
> 
> Whether you think it's a good thing or bad I find it difficult to imagine a future for Linux without more _and better_ user-friendly interfaces to sysadmin functions.  The big challenge in the past (I don't know, I wasn't around) seems to have be installation and, from what I have read, most distributions seem to have made this an awful lot easier.  The next step will probably be -  "now the thing's installed let's make it easier to configure and maintain".  Undoubtedly there will be a painful period where these tools break more than they fix but, with open discussion and feedback, they will become as good as other parts of the system.

This is chasing an uncatchable target.  You see the reason why MS has 
succeeded[1] in developing workable gui tools is that they control all
aspects of the system.  They have the ability to make sure that config
tools version match with the software they configure.  In the *nix community
most of the developers of the software you want to configure with your
gui tool of choice have no connection with these gui tool projects.  They
don't care if webmin can configure their software or not. It is the lack of
continuity between the configurator and the target software that makes these
tools so dangerous. It makes it distinct possibility that you could get a 
mismatch of software and damage the system badly[2]. 
> 
> I was interested in how well programs such as webmin, linuxconf, ximian setup tools, etc. were doing, whether parts of them were reliable enough to use, etc. and it seems like a few other people on the list were interested too.  Of course, I wouldn't be caught dead using them... :-)

until these tools are being co-developed with the software they configure as
an integral part of software package they will continue to be more dangerous
than anything else. It could happen in the future but I wouldn't hold my 
breath.  The fact is that the people who do the developement don't see the
need.[3]

--Matt

[1]: Atleast for those that somehow consider what MS has done as a success.

[2]: A mismatch between the bind version installed and the one that 
	  linuxconf configured for circa RH 6.2 is one of the reasons why
     so many people hate linuxconf with such passion.

[3]: When will those pesky hackers see the light emanating from Redmont and
     follow MS down the enlightened path. ;)
-- 
"Take away them collisions and the common channel and it's like Christianity 
 without Christ." -Jim Breen (speaking about "full-duplex" Ethernet)