openSUSE Strategy Discussion: For the productive poweruser

English Open Chat
1

Hi all!

Okay, you know the drill. This one is the “openSUSE - For the productive poweruser”.

—8<------8<------8<------8<------8<------8<—

== openSUSE - For the productive poweruser ==

=== Statement ===

openSUSE should strive to be the productive distribution for powerusers on modern PCs (workstation, laptop, netbook, server) and having a healthy balance of innovation and stability.

We cannot compete with Ubuntu for the übernoob segment, and we shouldn’t compete with Fedora on being experimental bleeding edge - instead we should pick the middle ground.

This strategy would be nicely in line with SLE and what (open)SUSE has historically been, and what existing users expect from openSUSE.

The main purpose of the strategy in my opinion is to help developers, contributors and marketers all pull in same direction, and to clarify for users what openSUSE tries to be and do.

NOTE: In my mind you don’t have to be a kernel hacker or a guru sysadmin to be a poweruser, in my estimation powerusers cover:

  • ~10% of all PC users
  • ~50% of all Linux users
  • ~75% of existing openSUSE users
  • ~100% of existing openSUSE contributors

== Activities ==

==== We need to be excellent in the following ====

  • Making sure as much as possible just works out of the box
  • Having good and sane defaults so the user can do what ‘‘he’’ wants to do
  • Focus on providing tools for being productive/creative (IDEs, editors, authoring tools, graphics manipulation, office productivity, etc.)
  • Providing admin tools that are powerful yet (reasonably) easy

==== We will try to do the following effectively ====

  • Deliver a strong, general purpose distro that anyone can use without too much effort
  • Innovate and keep up with latest upstream developments

=== As project, we will not focus on the following anymore ===

  • Dumbing things down for Aunt Tillie
  • Going out of our way to support old hardware and non-mainstream architectures
  • Supporting form-factors that are not workstation, laptop, server or netbook
2

Great idea. One thing, though,

  • Going out of our way to support old hardware and non-mainstream architectures

how old is “old?” 11.3 runs with 512MB and a Pentium. It runs smoother than 11.2 on my old desktop with 1GB and P4. Will we expect that kind of support in the future?

3

I wouldn’t consider that old. I think they mean things like old ISA sound cards, Sparc workstations, that sort of things.

4

Ok, my two cents is that most of the time the needed software is available, but not as easily accessed as could be. Say I need phpMyAdmin. I find that it’s in auxiliary repo. Ok, so it’s just a one-click-install or zypper ar followed by zypper in. But still not as convenient as the one-stop-shops of other well known distros.

Quality control seems to be also an issue with these auxiliary repos, sometimes you’ll find 5 different packages of slightly different versions provided by well-intentioned builders. This is confusing to users.

Could there be a community effort to gather most of the wanted poweruser software that is not in the main repo under one auxiliary repo?

5

Thanks for the confirmation. Pentiums go a ways back, but I guess they are pretty standard.

6

I’m not sure I’m a power-user (it’s a vague term after all), but I do agree with the goals stated here - and am already pretty satisfied with openSUSE. Due to the buildservice it is quite easy to decide which parts of the system should be stable, bleeding edge or plain nuts. The software collection (when taking the buildservice and the usual 3rd parties into account as well) is huge and up to date, advanced software management can be handled easily and flexible via zypper priorities, locks and the like. There will always be a limit to the goal of a system running out of the box (exotic or brandnew hardware etc.), but I personally have never really had any hardware issues (might be just luck, of course). I also find openSUSE to be one of the most innovative distributions, just look at what has changed and improved during the last two years.

So obviously I am missing some point here - what exactly should be adapted to those goals? There are some ideas I’d like to see implemented, but those would be things like not using the same password for user and root during the initial install by default (sorry Aunt Tillie) - nothing earthshaking really.

As I said, I do agree with these points. But isn’t that pretty much the direction openSUSE is taking for some time anyway?

7

On 2010-07-30 19:36, prusnak wrote:
>
> Hi all!
>
> Okay, you know the drill. This one is the “openSUSE - For the
> productive poweruser”.

Well, no, I don’t know the drill.

…

> The main purpose of the strategy in my opinion is to help developers,
> contributors and marketers all pull in same direction, and to clarify
> for users what openSUSE tries to be and do.
>
> NOTE: In my mind you don’t have to be a kernel hacker or a guru
> sysadmin to be a poweruser, in my estimation powerusers cover:
> * ~10% of all PC users
> * ~50% of all Linux users
> * ~75% of existing openSUSE users
> * ~100% of existing openSUSE contributors

In my mind a power user is one that can adjust the computer to do things on his own to do tasks on
the computer, that don’t come completely prepared for “play” out of the box. An admin, even a user
that setup a linux box for other people, is a poweruser.

A power user, for example, would be one that creates openoffice calc sheets with formulas. Or one
that setups postfix for doing “things”. Or one that reads the docs and adds amavis to postfix. Or
one that asks the proper questions in a mail list or forum, and answer other people questions,
properly, and successfully.

> == Activities ==
>
> ==== We need to be excellent in the following ====
>
> * Making sure as much as possible just works out of the box
> * Having good and sane defaults so the user can do what ‘‘he’’ wants to
> do
> * Focus on providing tools for being productive/creative (IDEs,
> editors, authoring tools, graphics manipulation, office productivity,
> etc.)
> * Providing admin tools that are powerful yet (reasonably) easy

  • Providing a distro that continues working, without having to reinstall every year (in a hurry).

Meaning that 18 months is too little for an admin. I know admins (some of which which contributed a
lot) that have left the distro because of this, migrating to another with a longer life, and thus,
they have stopped their useful contributions.

Sometimes it happens that a version doesn’t work well on some hardware. For example, it happens that
I had to skip 11.1. I will probably have to skip 11.3. A power user that needs to do useful things
with a computer, an admin of several machines, will probably wait a month or more after release to
upgrade shop, till the GM becames really stable. Then use it for a year. Then almost no time left to
upgrade to the next one (3 + 12 + 3), no safety margin. And no way to schedule a yearly upgrade at a
fixed month.

No, don’t suggest SLE. If I were to use SLE, I would stop contributing
(because if I pay, why should I contribute?). Many of these admins I’m
thinking of can’t afford it, anyway, or their bosses allow them to “play”
with Linux because it is free.

Otherwise, a short life for a distro means that the upgrade has to be totally safe and very easy to
do, with no problems. And this does not happen, upgrade is a pain. Ie:

  • Provide a distro which upgrades reliably and easily, on every single release.

> === As project, we will not focus on the following anymore ===
>
> * Dumbing things down for Aunt Tillie

No idea who this aunt may be? :-?

Guessing a little, I know what you mean, but use English that non english people can understand in a
paper or proposal. This is an international project.

> * Going out of our way to support old hardware and non-mainstream
> architectures

Which may mean evicting from our community users and power users in poorer communities. Do we wish that?

> * Supporting form-factors that are not workstation, laptop, server or
> netbook

Like a desktop? I don’t understand.

Yes, I’m being thick on purpose :slight_smile:

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))

8

I think of anybody with more than a passing interest in the details and management of their system. Most people want to do a task and move on. They don’t care about, or in some cases are even horrified by the thought of learning what makes this work and how to improve it or how to make it last longer. They would rather let somebody else do this, or upgrade every year and not worry at all. Power users would be “other” users: Not necessarilly an administrator or a developer, but somebody willing or even desiring to go deeper and “get dirty.”

9

Hey, I’m a poweruser then! :slight_smile:

And I agree that these are the potential users openSUSE should be aimed to. Users who are interested in their system.

10

I consider anybody who has added software beyond the standard install (whether from install media or online repos) for the sake of software development, system maintenance, or just plain tweaking, to be a poweruser. As opposed to someone who just uses the computer as an appliance. (Which is not pejorative, it’s fine if that’s all they expect of the computer.) As such, many people in this forum would be powerusers.

11

On 2010-07-30, prusnak <prusnak@no-mx.forums.opensuse.org> wrote:

Great speach. I don’t like speaches, sorry. Every time I hear a buzzword
such as strategy (when did the military release that for use in just about
anything?), I’m reminded of all those wasted hours in meetings with
marketroids and pointy-haired manglement people.

Sorry again, pet peeve…

Anyway, just one thing really. The following statements:

>==== We need to be excellent in the following ====
> * Making sure as much as possible just works out of the box
> * Providing admin tools that are powerful yet (reasonably) easy

>==== We will try to do the following effectively ====
> * Deliver a strong, general purpose distro that anyone can use without
> too much effort

… to me seem to contradict the following, in many ways:

>=== As project, we will not focus on the following anymore ===
> * Dumbing things down for Aunt Tillie

Dumbing things down is exactly what you need to do. Plenty of help, plenty
of intput-checking and simple, clean presentation of choices. (currently,
the printer config interface is a good example of a bad interface).

Just do it in such a way that more advanced user can still access their
configurations in a more ‘hands-on’ fashion.

The two approaches aren’t mutually exclusive.

–
When in doubt, use brute force.
– Ken Thompson

12

Rikishi42 wrote:

>>=== As project, we will not focus on the following anymore ===
>> * Dumbing things down for Aunt Tillie
>
>
> Dumbing things down is exactly what you need to do. Plenty of help, plenty
> of intput-checking and simple, clean presentation of choices.

Developing UI (user interface) is a much more work then coding some
functionality.
Do we have enough interested to read about usability, psychology, analyze
workflow, develop graphics, write manuals and so on.

–
Regards,
Rajko

13

Good point. We could focus more on development, but who would be doing that? I don’t see too many mockups of better/improved interfaces on the planet or in other places…

So if one wants to focus on powerusers, you have three different kinds of those. The systemadmin type (long term support is important as Carlos mentioned, and needs good sysadmin tools), the office worker (needs focus on good office suite, scriptability, poweruser tools for eg filemanagement, and long(er) term support) and the developer (needs FAST release cycle, easy package installation and management (esp up-to-date packages), otherwise things must be easy).

IOW developers seem to kind of clash with the other two, needing a shorter release cycle and such. So this strategy focuses on powerusers and thus should have a long release cycle, the ‘home for developers’ strategy focuses on the developers, obviously. With good repository management you could combine the two…

14

Developers don’t necessarily need a short release cycle. Not all developers are chasing the latest release. Sometimes they want to stick with a known version of the compiler tools, for example, to reduce one variable, and worry about making it work with all compiler versions later.

However providing an up to date repo doesn’t preclude stability. The onus is on the poweruser to make choices about which versions they want. An install of a package doesn’t necessarily trigger an update of a dependency if the rule specifies >= and it is already fulfilled.

The hassle of the openSUSE poweruser repos is that packages are chopped up into many many repos. Every person and his dog seems to in the I’m providing a package game. I’m arguing for a more unified and coherent repo for powerusers.

15

Hmmmm… while I get your point, “chopped up” repos have the advantage that the user can decide which elements of his system are bleeding edge and which are stable. I for example like to use the very latest KDE releases, while I prefer my Mozilla apps to be rock stable.

16

Isn’t that what version numbers are for? Also if you still want to categorise by bleeding edge/stable, then that’s two repos, not 2x(number of packages).

17

You have to be more specific if you want me to understand - right now I can decide exactly which apps (or group of packages) shall be stable and which may be more experimental. In what way can that still be provided when all groups of packages are divided into stable and bleeding edge only? They’d contradict.

18

Well if you want to classify, then you have to apply some criteria. Generally the dev group has one stable release and one dev release, that could be the first cut.

The way it is, there aren’t even any labels on the personal repos by default. Joe releases 5.4. Is that the dev release or the stable one? Have to do some research. Maybe John’s 5.3 is the ticket. But he hasn’t updated to 5.3.6 that came out a month ago with bug fixes. Bugger. (Not real examples, but the sort of thing that could happen.)

See the problem? No organisation. If classifications are to be useful, then contributors have to submit to some discipline and make an effort to keep up.

19

I was not referring to ‘home:’-repos here. I find them to be a pile of trash (only very few are managed with care). I am rather referring to classical buildeservice-repos such as ‘Mozilla’ or ‘KDE4:Factory’ etc.

20

Well, in that case those still could usefully be classified as stable or dev for the benefit of powerusers. As opposed to the one from the install media+update repo which one could call frozen perhaps. You say you might want to mix stable A with dev B. But this situation can be handled with the zypper DB features, lock one of the packages.