Donnie Berkholz ([info]spyderous) wrote,
@ 2006-09-11 00:27:00
Previous Entry  Add to memories!  Tell a Friend!  Next Entry
Entry tags:enterprise, gentoo

[Gentoo] Focusing Gentoo without forking it
Mark Shuttleworth posted (Thanks to Steve O'Grady for the link) about how Ubuntu focuses in a few specific areas, but Debian is a more general plateau. One can trivially draw the parallel between Gentoo and Debian, so his points are equally applicable to us. Most Gentoo developers draw the most pleasure from working near the bleeding edge, not from trying to backport fixes and fix old stable software.

Perhaps this is because it requires more creativity and less monotony. I certainly feel more challenged and fulfilled by packaging new software (such as the system-config-* utilities I did last weekend) than by fixing some simple bugs for random stable packages.

But this raises some new questions: Can Gentoo develop specific "peaks" in conflicting areas, without forcing new subdistributions to form that focus on them? If so, how? Stuart Herbert and I threw around some ideas shortly after I started a discussion about whether democracy works for Gentoo, and our lack of goals.

Stuart's idea, which I like, is preparing specific "releases" for certain vertical markets. Yeah, I said "vertical markets." WTH is that? Just a given group of people using Gentoo for a certain purpose, such as a LAMP stack, an HPC cluster or a development workstation. One could create a LiveCD with an installer image tailored to, and preconfigured for, a LAMP server. The key here, as Stuart pointed out in our discussion, is making things "just work," not just installing the packages and leaving the user to set everything up. But we'd need more than just the LiveCD, because clearly people want to maintain the installation. Perhaps adding a series of profiles for these vertical markets could do the trick. Some developers have already tested this concept with GNAP, the Gentoo Network APpliance, but not in a formalized way that pushes into a number of different areas.


(Post a new comment)


[info]smitty1e
2006-09-11 09:06 am UTC (link)
My biggest wish, as a gentoober these two years, is that the internal data model were more tidy.
As for the apparent lack of focus, it boils down to leadership.
You yourself did the modular X work; if you simply and gently point in a direction and say "look, we're going here" the sheep (that's me) will follow. I trust your instinct and skill way beyond my own.
I'd love to have the knowledge to be more of a driver and less of a load, but gentoo needs an O'Rielly-quality book describing the gazintas and gazoutas. I say this having pored over a lot of the available ebuild documentation. People have obviously worked hard on that, and it is good stuff. But the underpinnings remain mysterious.

(Reply to this) (Thread)

[info]spyderous
2006-09-11 09:16 am UTC (link)
Could you describe in more detail these gazintas and gazountas that need documenting?

(Reply to this) (Parent)(Thread)

[info]smitty1e
2006-09-11 11:19 am UTC (link)
Let me answer that backwards.
If someone asked me to give a description of gentoo, I would characterize it as a source distro characterized by portage, a python system managing a tree of bash scripts, with something of an object model tacked upon them, that keeps the system software current.
Where I'm a bit vague is whether the python acts directly against the .ebuilds, or how the python imports the .ebuild information into its cache and does its work from that.
Two things that might help the situation:
1) (I think that this might be possible for the end user) Is just trace the emerge command from a python prompt, and see where it winds.
2) A play-by-play narrative of a simple ebuild, say, one for nano, as portage "do the voodoo that it do" thereto.
In summary, despite having read the .ebuild and portage documentation through two or three times, I haven't yet developed a detailed mental model of what its doing, and couldn't really maintain the code.
Is it safe to ask your opinion on paludis?

(Reply to this) (Parent)(Thread)

[info]spyderous
2006-09-11 11:23 am UTC (link)
You might like http://dev.gentoo.org/~g2boojum/portage.html.

(Reply to this) (Parent)
Vertical releases
(Anonymous)
2006-09-11 01:34 pm UTC (link)
I love this idea.

Saying "This machine is a .... ", then installing and configuring the necessary packages to have the machine work that way out of the box would be fantastic.

Ideally we could apply and unapply multiple vertical configurations to a basic install, so that over a lifetime, the uses of a box could change. However, I guess that's not a "must have" requirement - if it's easy enough to build a machine, then rebuilding shouldn't be too much trouble.


Of course, vertical markets are easy to find:
Router,
Web/LAMP server,
File server,
PVR,
Desktop (gnome/kde/xfce)

Hmm, yep I'm suddenly interested in this idea.

Rcomian (user)

(Reply to this) (Thread)
Re: Vertical releases
(Anonymous)
2006-09-11 02:20 pm UTC (link)
I don't think this makes too much sense. If a person wants to run an apache server, her or she should be able to emerge apache. It's no rocket science.

I think the more valid discussion is really whether Gentoo should have an "ancient but solid"-profile. I'm not reasoning here that ancient means more stable. I am more reasoning that once I've got a running server I only want to change as little as possible.

(Reply to this) (Parent)(Thread)
Re: Vertical releases
(Anonymous)
2006-09-11 04:09 pm UTC (link)
Hi,

There's a bit more to running a web server accessible from the Internet than just emerging apache. A lot of folks won't be aware of everything that has to be done, and they won't have the experience to make informed decisions about the many different ways that this can be done.

Even for the folks who can do it - plenty of them simply don't want to. They want to get on with the main event - which is doing business. Setting up and administering a webserver is an overhead - a cost that they need to keep as small as possible.

So, in that respect, it's good for some folks who already use Gentoo .. and it also makes Gentoo accessible to lots of folks who don't use Gentoo today.

It's also good for Gentoo. It gives us interesting things to focus on, practical things, and the results of that will trickle down to everyone who uses Gentoo for anything remotely similar to these verticals. It gives us focus, but it doesn't take effort away from the core Gentoo tree.

And, longer term, it creates the potential for Gentoo to bring in revenue, which can be ploughed back into the wider community.

(Reply to this) (Parent)
Re: Vertical releases
[info]spyderous
2006-09-11 04:17 pm UTC (link)
You left out the MP in LAMP. =) They also need to install MySQL and PHP, then configure them as necessary to work with each other, set up some databases + proper permissions to them, probably want to come up with a GUI or Web-based administration tool or two that's already working, etc.

Pick something you know less about, then, or something that takes a while to create. How about a person wants a router with full Quality of Service, NAT and Layer 7 routing capabilities, preconfigured for a reasonable setup? Or a person wants a high-performance cluster ready to go?

(Reply to this) (Parent)
Re: Vertical releases
(Anonymous)
2006-09-14 08:53 pm UTC (link)
What about something like a package class for these predefined configurations?

(Reply to this) (Parent)(Thread)
Re: Vertical releases
[info]spyderous
2006-09-15 05:18 am UTC (link)
Could you expand on that?

(Reply to this) (Parent)
Re: Vertical releases
[info]mswilson
2006-09-11 05:01 pm UTC (link)
Going after vertial markets is one thing that rPath (http://www.rpath.org/) has been doing. rPath's tools make building releases that are ready-to-run and pre-configured for a particular purpose. Even the Gallery guys have made a version of Gallery running on it (see http://gallery.menalto.com/).

(Reply to this) (Parent)
Gentoo's existing focus: people who don't like express installers
(Anonymous)
2006-09-11 10:05 pm UTC (link)
I am a gentoo user who is quite happy with the system as it is. I built my gentoo box from a stage 0 mini-cd installer with a printed copy of the Guide next to me. I reveled in the ability to "get it right" from the ground up - without having to undo the work of a installer plugging default options in.

Gentoo does have a user-base, and one that has survived so far with a well polished LiveCD and good documentation. How many of gentoo's users are like me, a classical computer hobbyist ?

IMHO building a distro for motivated self-learning linux hobbyists' would be far more fun than building a corporate terminal: office,e-mail,web,photo.

I think the "desktop" vertical is a euphemism for the application bundle above. An express installer, or even kickstart like installer is possible because it targets a user base that does not stray far from defaults, or a typical PC hardware setup.

As for LAMP or other verticals I don't think those are very practical. A proffessional user needs flexibility, integration with existing infrastructure, and scaling. I would guess (WAG) that most of the proffessional users would skip a installer to refine the installation to meet their requirements with precision.

If user's need databases, or webserving, I would think that they would turn to web services focused on an application rather than building something from the bottom up.

In summary I think gentoo does have a focus: the hobbyist/proffessional. Do the gentoo developers really want to re-target the distro ? if so why ?

(Reply to this) (Thread)
Re: Gentoo's existing focus: people who don't like express installers
[info]spyderous
2006-09-11 10:09 pm UTC (link)
Sitting around maintaining ebuilds is not interesting. Helping Gentoo _go_ somewhere is interesting. The question remains, where can we (or at least some of us) go without getting in the way of people doing other things with Gentoo? We developers need goals, we need to know that Gentoo is moving forward (forward to where is another question). Bumping foo 1.0.0 to 1.0.1 is not a goal and is not fulfulling.

(Reply to this) (Parent)(Thread)
Re: Gentoo's existing focus: people who don't like express installers
(Anonymous)
2007-02-07 10:55 pm UTC (link)
Two cents against this...although the vertical direction is great, this does nothing to resolve the problem around the moving version 1.0.0 to 1.0.1 ...that still needs developers/volunteers and more so after the verticals ...having said that...the vertical profile idea is great...it is using the power of gentoo as a metadistribution to its fullest...it will help popularize it and maybe get commercial support so that ppl could be paid for the version advancement...just like Ubuntu does :)

(Reply to this) (Parent)

(Anonymous)
2006-09-12 12:28 am UTC (link)
Sounds like an excellent idea. I think profiles in conjunction with installer configs would do the trick, using the pre-existing Gentoo System architecture and sane pre-configurations to make everything "just work".

In fact the people on The Linux Action show inquired about "Gentoo server profiles", so there's definite interest in this type of solution.

http://www.linuxactionshow.com/wp-content/plugins/podpress/web/aHR0cDovL3d3dy5saW51eGFjdGlvbnNob3cuY29tL2VwaXNvZGVzLy9MaW51eEFjdGlvblNob3dFUDAxMi5vZ2c/LinuxActionShowEP012.ogg

(Reply to this)

Create an Account
Forgot your login or password?
Login w/ OpenID
English • Español • Deutsch • Русский…