Too many dependencies? was: Is Linux always so frustating?

[M. Fioretti]

(Jef's reply rearranged to explain myself better)

On Wed, Apr 07, 2004 14:50:38 PM -0400, Jef Spaleta
(jspaleta at princeton.edu) wrote:

> > In general, careless packaging like this means that custom installs,
> > selecting packages by hand, are much less effective, since one has to
> > install (and later upgrade/maintain) much more than he will ever use
> > or need anyway
> 
> Why do you assume that this is CARELESS.

I sure owe you and the list an explanation. I lumped in one single
definition, careless packaging, different problems. Not explaining
what I meant by that is obviously my fault, please excuse me.

By careless packaging I meant *anything* that may force the *end* user
to install (maybe buying new hardware) applications or libraries that
might never be used. This goes from building an RPM with unnecessary
Requires to, as you pointed out very well, a suboptimal architecture
of the initial package.

Let's assume, as you suggested, that OOo is made in such a way that it
now *pretends* another application instead of having a plugin or other
open, general interface to *any* address book. If so, OOo itself is
packaged/architectured badly, at the source.

Of course, yes, in such cases cases where the fault is at the origin,
there is almost nothing that distro package builders can do. In that
case I can certainly see why they have to end up compiling in just
about everything.

As a matter of fact, you have just given one more excellent variation
to my standard answer to all who ask me "if RULE is aimed to low end
systems, why base it off Red Hat/FC instead of Debian/ Slackware/
etc..."????" " Because many problems are upstream, so it doesn't really
matter which distro we start from".

Said this, I am still (even more) interested to find out how many
applications are "carelessly packaged", as defined above, and at which
level. The OOo+Mozilla issue in the other thread just prompted me to
ask. Sorry if my sleepy wording made it look harsher than I meant.

Said this, and with respect to this comment:

> what you think the end product should look like to make your niche,
> low memory, project need easier to build.

I'd like to point out that "low end" systems (not necessarily RULE)
are not niche at all. That might have been true 15/20 years ago, when
5+ years old PCs were unavoidably much less than new ones and almost
exclusively in the hands of professionals who could afford top dollars
for them.

Today is the computers with big, "cheap" HW, less than one year old,
which are the niche (as in "the minority" of active PCs): and if it
isn't so yet, it should, and will probably be so anyway. One month
ago, just to say one of many, many proofs, the UN university
recommended to *extend* as much as possible the useful life of PCs to
reduce the related pollution:

http://www.wired.com/news/technology/0,1282,62562,00.html?tw=wn_tophead_6

Ciao,
	Marco Fioretti

-- 
Marco Fioretti                 m.fioretti, at the server inwind.it
Red Hat for low memory         http://www.rule-project.org/en/

There is hopeful symbolism in the fact that flags do not wave in a
vacuum.                                        -- Arthur C. Clarke