Some of my Kubuntu-Hoary tweaks

Sat Apr 16 15:44:15 UTC 2005

Jonathan Byrne dedi ki:
> On Friday 15 April 2005 13:31, Abdullah Ramazanoglu wrote:
  --8<--

> Ubuntu is Debian Made Easy (not just [Ff]ree), and its explosive
> popularity is a testament to how much this was needed.

Really, people are weary of too commercial distros, and Debian way is not
only [Ff]ree but it's also [Oo]pen (as in OSS and as in open to join and
contribute). IMHO Debian is a server class OS, but many Linux users
want/need a desktop class distro with Debian way. This is one thing which
makes me think that Ubuntu is getting its users basically from already
established Linux user base: It's a boom with bounds. But if Kubuntu can
be tweaked into something which would appeal to the masses without making
it too unfriendly for the erstwhile users (see Firefox analogy below),
than it would be a boom without bounds!

BTW I guess I seem to found a wishful slogan for Kubuntu... ;)

Ubuntu is Debian Made Easy (attracts current Linux base),
Kubuntu is(?) Ubuntu Made Easy (attracts the whole world).

Obviously I'm ignoring the question of how to build such a different distro
upon Ubuntu foundation. That's another issue.

  --8<--
> I've been using Linux since 1997, and I've seen excellent forums nearly
> destroyed when they became so newbie-focused that they drove away most
> of the people who had been helping the newbies.  Tilting totally toward
> beginners isn't the way Ubuntu will continue its rapid advance; it will
> continue it the same way it has gotten this far: by being accessible to
> both newcomers and old hands.

True for forums, but also it is true that there are many OSS projects
either technically non-exciting by their nature or geared specifically for
casual users yet they have good developer and community support. A good
example is Firefox. They've found the optimum balance where it is perfect
for both casual and advanced users.

E.g. "about:config" is genious of them. Till Firefox I used to think that
UI should be hierarchically layered (e.g. "Advanced" button) so that only
those who are interested in advanced setup/features would enter there,
hence the software would be both usable by the clueless and satisfactory
for the expert. But the "Advanced" tab was still there to confuse the
users. Firefox carries this a step further and "hides" the advanced
options (about:config) in such a way that those who know about it is those
who would get excited with it. Those who would be confused by it, wouldn't
be aware of its existence in the first place. In this way it's both
perfect for the casual user, and also perfect for the expert. The UI
designer must be a genious.

I believe there are lots of lessons to be taken by dissecting and
inspecting their success. From UI to marketing to fusion of expert and
newbie appeal to flexibility to whatnot. And they're succeeding in the
world at large, not just in a certain community. Firefox is a good study
case for OSS.

  --8<--
> Pretty much all development packages are excluded from the CD, which
> while I support that, it is a bit of a PITA.

I agree. Upon my first Debian installation, I had found it quite
frustrating to miss even ubiqutious system tools/utils that I had
accustomed to having in any default Linux installation. But then I was not
aware of meta packages.

> If vi were to be kept out too, I'd like to see it as part of a
> meta-package called developer-tools (actually, this is a great idea
> anyway, IMO), which would in a single apt-get install give you the
> development kitchen sink: all the languages, gcc et al, IDEs,
> development libs, the works.  This would be along the lines of what you
> got with the old Red Hat (I've never installed Fedora, but I imagine it
> does that, too) Development Workstation installer option.
>
> Or, put different profiles in the installer: Default (what we get now),
> Development Workstation, Server (already available), maybe one or two
> others.

Actually there's no problem as of now. But when the CD gets filled up, we
will all start asking the question of whether to install vi or kdeedu over
the 'Net (an over-simplification). I agree with your concerns and I try to
find a wayout in my mind. Here is a simple analysis and a compromise:

Let's simplify it as a vi versus kdeedu case, and let's assume vi appeals
to 100K experienced Linux users and kdeedu appeals to 100M casual users.
There will be 100.1 million CDs shipped. 100M of the audience probably
won't know, or care for, or have the means of installing additional apps
over the 'Net. But the 0.1M of it would both know, and care and most
probably have better internet connection means. For those with limited
internet connection, a very simple howto (rather a step by step cookbook,
even a script) would be quite satisfactory for them to prepare a secondary
CD crammed with their favorite apps, so that they won't have to download
the same packages over and over again for repetitive or LAN-wide
installations. An experienced user is more prone to do repetitive
installations, or being in charge of many PCs to install Kubuntu. But a
casual individual user wouldn't have incentive for repetitive
installations, nor would he likely be in charge of many PCs. So, even if
we -incorrectly- presume that a casual user is equally well equipped to
prepare a custom (secondary) CD, it would still be illogical for him to do
it.

It looks quite convincing to me, to have a pristine Kubuntu CD (crammed
with apps appealing to the masses) and a not-so-pristine Kubuntu CD-2
crammed with everything appealing to an experienced Linux user. Personally
I wouldn't mind at all (on the contrary it would be more appealing to me)
having to download two CDs: Second CD can be an iso download or a
CD-preparing script somewhat along the lines of Jigdo.

How's the idea? Would it be a good compromise for everyone?

>> While we're at it, (fasten your seat belt, don't fall down :) I also
>> promote any and all console tools totally excluded from the CD except
>> the ones cruicial for system maintenance for which there's no GUI
> 
> Better to have two CDs and throw in the kitchen sink :)
>
> OK, we probably don't really need that,

But the idea seems increasingly warmer to me. Working further on the
suggestion above, it would be really exciting (for me) to have not one,
but two additional CDs, with a total of 3:

CD-1: Pristine Kubuntu CD.
CD-2: Kubuntu-Power CD (filled with everything you'd have dreamt)
CD-3: Kubi+ (filled with everything a casual user would have dreamt)

CD-2 & CD-3 don't have to be released by Kubuntu core team, though it's
more desirable. Also there may be many "CD-2" and "CD-3" flavors (even
other CDs focused on various areas) built and distributed by several
people. Either directly as iso images or as CD building scripts.

I was just thinking aloud.

Best regards
-- 
Abdullah Ramazanoglu
aramazan ÄT myrealbox D0T cöm