The state of UNIX in general and PLD in particular

[Cezary Krzyzanowski]

Dnia 05-03-2007, pon o godzinie 03:42 +0100, Jakub Piotr Cłapa
napisał(a):

> 
> I'm not trying to limit this. I know why I use PLD and why I'm a 
> developer (unactive recently but still). It's just that maybe our policy 
> of breaking everything down to almost a one package --- one file state 
> is not the best thing for the desktop. Thinking about a better idea 
> shouldn't be a bad thing.
> 
+ all other arguments

Thinking isn't prohibited, is is *usually* needed. All what's been said
above comes to one end - how to use these bundles and if we want them.

My personal opinion is, that it is totally impossible to put together
bundles + The Ultimate Package Fragmentation. I think I'm done reading
arguments about possible usage of bundles - I acknowledge them. The main
thing is - how to put that into PLD not breaking what we have now.

I'm not sure if there is a possibility to draw a clear line between
server (fragmented) apps and desktop (bundled), as things like VM-ware,
firefox, xorg are used in mass server-like environments, like campus
networks or some 'embeded' systems to use in libraries or during
students recruiting (vide Undefs work). IMHO metapackages could bring
PLD closer to the desktops. For me poldek is the ultimate tool for
applications acquiring, so I don't see any need for bundles, as NU won't
have the need to install anything from the outside via bundles (in a 15k
SPECS distribution).

I think, that statements presented in the referenced paper are only in
power on desktop machines and they do not apply for server distros like
ours. The usage of libraries on a 'typical system' (understood like in
the paper - ubuntu desktop) is truly limited, but then again we don't
aim at this part of market. I'm sure the percentage usage of shared libs
varies from system to system and most important - from system type to
system type (vide multi-desktop environment) and there is no clear way
of drawing lines between bundles to fit every need. In contrary of
metapackages, where every possible imaginary system could be set up out
of millions of tiny cut packages of PLD.

I suggest showing some real implementation of the bundling idea in PLD
or think of ways to incorporate this into existing PLD. Before that
there is no need to make a discussion, as fighting the whole idea is
wrong, but fighting a particular implementation is not.

Cz w rny