Fwd: Etch

Kristian Hermansen kristian.hermansen at gmail.com
Mon Apr 9 16:00:48 EDT 2007


On 4/9/07, Matthew Gillen <me at mattgillen.net> wrote:
> I found a local Ubuntu install, and found an example of what I was talking
> about regarding lots of small packages vs. one big one:
> For X fonts, Ubuntu has 82 packages (aptitude search ~n | grep "xfonts-" | wc
> -l).  Fedora has 19 (yum list | grep xorg-x11-fonts | wc -l).

Ubuntu breaks up the Xorg packages into many subcomponents, yes.  This
is designed to deal with not having to download the entire package to
update just one part that changed.

> Basically what I'm getting at is that I'm not convinced that 23489 packages
> have more useful software than 6797 packages.

I still maintain my hypothesis :-)

> I agree that years ago the number of officially supported packages was rather
> small, and I'd inevitably have to go hunting for useful software to install
> manually.  But you'd be surprised at what's in the default repos nowadays.
> The only software I typically manually fetch and install anymore is commercial
> stuff like Acroread (because evince still gets confused by some of the busy
> pdfs I need to read, and xpdf is just ugly).

Ubuntu offers these packages in their own repositories, including
Sun's Java, Adobe Flash, and Acroread.  This is beginning to prove my
hypothesis above...

> Since I don't think the package-count is a good metric, and I can't think of a
> better one, I think the only way to prove you wrong would be for you to get a
> Fedora install and try to 'yum install' all the software you normally use.

I agree that it is not a great metric for the reasons you mention.
The real metric should be, "how often do you have to circumvent your
repositories to get the software you want".  The answer with Ubuntu
for me is, "almost never".  For Fedora, I already mentioned three
cases above for which Ubuntu wins here, and my hypothesis grows
stronger still...
-- 
Kristian Hermansen

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.




More information about the Discuss mailing list