Re: Flash Player 9 beta released

From: Dylan Griffiths <dylang_at_no.spam.please>
Date: Sat Oct 21 2006 - 16:42:42 CST

Steven Kurylo wrote:
> Or that you're trying to use a tool not aimed at the desktop end user.
> Of course I don't know what would have happend if you had clicked on
> that deb on the ubuntu desktop. When you're using a beta third-party
> program, its at your own risk. Last time I was on the ubuntu desktop
> and clicked on an official deb, everything just magically worked.
>
> I bet I could find/make a package which would cause DnD to fail too.
>
> You have to comapre the same things, if you want me to let you get
> away with a smug Apple is so cool rant. (Not that Apple isn't cool,
> and doesn't generally have a better interface)
>

Yeah, but this is an example where the tools aren't supporting my help.
  People talk about how awesome Debian is, and I'm sure it is if you're
in that headspace.

Let me compare some same things. On Slackware when I want a new kernel,
I freely grab the source from kernel.org, compile it, and install it. I
can do it with LILO and GRUB easily. I'm not inexperienced with kernel
stuff. I would not expect a mere mortal end-user to do this.

How does Ubuntu support this? Through a Debian method, which is
completely different than the kernel development model. However, it's
not simpler! They've just traded the complexity of understanding where
to put kernel bzImages and system.map files and talking to the boot
loader, with a complex package system built on implicit and explicit
assumptions that only the Debian maintainers seem to be fully aware of.

Do you think that's a good idea?

I don't. I think it'd be a lot better if the tools were redesigned to
support being pointed at a kernel directory in a fire'n'forget manner.

Debian's toolchain is built around nerds. And not mere normal nerds who
happen to like D&D (not drag and drop, the D20 kind), the kind of nerds
that memorize eigen values for fun. I've used Ubuntu for as long as
I've used a Mac (maybe more), yet I can say that in daily use of a Mac
toolchain, I've learned more about how the system works (much like I
learned a great deal of how Linux worked by starting with Slackware a
decade ago).

Ubuntu proves that throwing a lot of people who want systems to be
useable to at Debian is a mostly-ok affair, except when you want to
actually do something like this where assumptions fail. That's why you
should move to a package standard that doesn't make such assumptions!
Apple made these standards when they moved to a UNIX base in OS X. .App
files will not break -- they will install easily via drag and drop. Why
has Debian not moved past 1995 package standards? I know there are
enough skilled, talented, and intelligent people behind Debian that they
could do it if they were motivated.

Ubuntu, unfortunately, is limited by its base and its own developer's
personal time, like any other project.

Let me summarize my perspectives:
* I'm frustrated as an end user when the software fights me. Software
is a tool, not a battle.

* I'm frustrated as a programmer because I know there are better
abstractions and designs available. These should have greater adoption!

* I'm frustrated as a Linux user because I can't get the latest flash
easily. Policy is not clearly dictated by anyone.

* I'm frustrated as an Ubuntu user because no Ubuntu people have put it
into the official non-free UNIVERSE repository yet. You would think
they'd have a quick turn-around-time since it's been a few days now, and
YouTube is something that seems popular.

I honestly thought about getting involved in Debian politics to fix the
situation starting at frustration #2 and #3, but then I came to my
senses and realized it was easier to use a distribution designed for
end-users (Ubuntu). If I had personal time, I'd address the issue at
point #4, but as point #1 shows -- I shouldn't have to.

You have a great technical point (apples to oranges and what not), but
this is about policy. Apple isn't cool; I sure as hell can't compile my
own kernel (or fix their stupid block-IO layer, which sucks donkey
balls, let alone the memory manager!), but from a policy level they kick
ass.

It's less of a sacrifice for me to live with a shitty kernel and a great
userland than a great kernel and a shitty userland :p
Received on Sat Oct 21 16:43:13 2006

This archive was generated by hypermail 2.1.8 : Sat Oct 21 2006 - 16:43:19 CST