[Linux-aus] Computerworld: Linux 'not ready' for enterprise IT

[David Lloyd]

Brent,

> > I have had to downgrade a good number of RedHat servers or hold them
> > back because of RedHat's rather broken RPM dependency issues. 
> 
> Same here, up until 18 months ago where I took a decision to stick with
> one distro with a reasonable packaging system, namely RH.

Fair enough.

> RedHat Network seems well suited. 

It is...and it's not altogether different from what the Debian security
folks attempt to achieve.

> Yep, installing individual RPMS for upgrades can be a bastard, but why
> bother when RH do all that via RHN?

Vendor lock in.

> You see, laziness is a virtue. I could choose to run each patch
> individually and suffer the consequences of a never ending dependency
> loop, but why bother when it's all there and done for a paultry yearly
> cost?

See above.

> 4)Have not a productive bone in their bodies and spend their time smoking
> up the room to hide their own inadequate approach to their job. 

LOL

> > "That the Debian packaging system--with its configuration tools,
> > priority levels and in-built dependency handling--coupled with a tightly
> > controlled, peer reviewed, open source repository is flawed and likely
> > to increase a system administrator's workload."
> > Now, go prove it...
> 
> You see,this is where I reckon the FOSS paradigm works so well. 
> I don't have to prove anything, it's the demand on a particular package
> which matters most. IMHO , low demand for a FOSS product equates to there
> not being a need for it, or that there is a fundamental flaw in it's
> approach. The latter often addressed by a code fork if the "demand" is
> there.

True but that has nothing to do with my null hypothesis.

> It's not what I or anyone else does (or postulates...;-) that matters,
> more so, it's the general acceptance of what is being done with it that
> matters most in terms of one distros acceptance over another.

That may be true but again it's not addressing the hypothesis I have put
forward.

> To that end, a convergence of distros would be disasterous, and would slow
> down development by moving away from plural views. The trick ,is it not,
> is to establish differing points of view through which the best solution
> is found. There HAS to be choice, wide sets of choices, otherwise the FOSS
> development model would fail.

A difficult trick indeed.

> The thing that irks me the most about such comments as this CQU admins, is
> that it slates this "choice" as being a downside, rather than a postive
> and enlightening feature. It's a contrary thought in direct oppostion to
> the culture that existed in CQU right up until I finished in 2000... I am
> ashamed of my Alumni...:-(

But it is to a certain extent a downside. What if you don't want to think
about choices...isn't that why you install RedHat?


DSL

-- 
Microbits Linux Technician
 +61 8 8362 9220

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.linux.org.au/pipermail/linux-aus/attachments/20030812/3195cbbd/attachment-0001.pgp