[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: The Direction of Debian



On Fri, 2002-04-12 at 18:39, Paul 'Baloo' Johnson wrote:
> On Fri, 12 Apr 2002, Dimitri Maziuk wrote:
> 
> > Basically, Debian project seems to be moving towards
> > more feechoorz, more luser-friendly helper apps etc.
> > Most software engineers believe that this exactly the
> > wrong thing to do: more complexity => unforseen
> > interactions between the parts => more bugs, less
> > stability.
> 
> And people wonder why I think both KDE and Gnome suck.  8:o)

Choosing to stagnate is never a successful option. The beauty of the
Debian system is that you don't have to use the progressive tools if you
don't want to, but they're still available. Just because it cramps the
style of some people is not a valid reason to eliminate major categories
of a general purpose distribution such as Debian.

> 
> > Besides, there already are newbie-friendly
> > distros, why try to muscle in on the niche that's already
> > been filled?
> 
> It's newbie friendly.  Debian won me over on price, but I stayed for
> it's features.

Since when has Debian tried to muscle in anywhere. Offering an option is
not "muscling in", instead of is giving choice. I find it baffling that
a Debian proponent would actually be suggesting that the removal of
choice is the best path.

> 
> > 1. I've nothing against helper apps. as long as they *can
> > be turned off*. Dexconf cannot be turned off (don't get
> > me started on general idiocy behind dexconf. Or alsaconf).
> 
> What is dexconf?

I second that ... what is dexconf? To my knowledge I've never used it.

> 
> > 2. Current practice of building packages with "commonly
> > used options" and adding dependencies to packages that
> > provide those options.
> 
> I thought it set packages at "recommended" when a particular feature had
> the dependancy.
> 
> > 2.1. Who decides what's "common"? Which idiot decided
> > that analog's libgd should depend on X11 libraries?
> 
> I've noticed this a lot.  It bugs me when I try to install something on
> my older box (that I don't have X on) and it insists on installing xlibs
> and whatever other x stuff it can find, when I know perfectly well it's
> not an X prog.

This is a vaild point, in my mind. Implementing a BSD-style ports
structure would help to get around this in my mind.

> 
> > 3. Potato is too old for many real-life uses. Woody
> > is unsuitable for many production systems -- not only
> > for political (managers have every reason to distrust
> > "officially beta" OS), but also for technical (Woody
> > doesn't get security updates, Woody breaks) reasons.
> > That will change once woody is released, but then the
> > cycle will begin again. On top of that, 2.2. makes
> > Woody unsuitable for certain class systems, and that
> > will not change with Woody's release. The point? --
> > Debian is becoming less versatile with every release.
> 
> That's always bugged me about stable:  It ages quickly.

Potato is also just fine for many real-life uses. Sounds like an empty
complaint ... I dont' see a suggestion for a solution, just criticism.

> 
> > Oh sure, RPM format was really inferior back then. Now,
> > Woody is still better that Seawolf (RH 7.2, that's the
> > one I had in mind). What I'm afraid of is that Woody+1
> > will not be any better than Seawolf+1 -- it'll suck as
> > much, only in different ways.
> 
> RPM *still* puts files in the wrong places, and you *still* have to find
> packages and resolve dependancies by hand.
> 
> > Murphy must have been asleep, then. As for dist-upgrading a
> > 4-yo potato, lots of updated packages will come with impoved
> > and slightly incompatible config files (X v.3 => X v.4, for
> > obvious example). If you're upgrading a fully-configured box,
> > you have every reason to expect it won't come back up after
> > the upgrade -- so expect at least 10 hours downtime (just
> > like you describe: Friday, Saturday and maybe Sunday, too).
> > Obviously, this won't work on a 24x7 e-shop webserver: you
> 
> Sure it is.  Round robin DNS, upgrade one server at a time.

Once again, lots of complaining and little offered in the way of ideas
for solutions.

Somebody once told me that if you come to the table with complaints and
no ideas for solutions for your complaints then you're part of the
problem, not the solution.

Just my 2c.

Sean

-- 

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: