Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld) [LWN.net]

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 2, 2008 7:45 UTC (Sat) by kripkenstein (guest, #43281)
In reply to: Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld) by Beineri
Parent article: Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

I'm sorry, it's not simple at all. Canonical was very open about the reasons for not LTSing
Kubuntu, and the one relevant to your suggestion to ship KDE 3.5.x is that shipping 3.5.x
means they need to support it for 3 years. In those 3 years, KDE developers will *not* be
focused on supporting KDE 3.5.x, they will be bugfixing and tweaking KDE 4.x. This leaves
Canonical with a single in-house KDE developer in charge of maintaining KDE 3.5.x, which is a
considerable burden and risk.

to post comments

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 2, 2008 10:09 UTC (Sat) by Beineri (guest, #39002) [Link] (6 responses)

> Canonical was very open about the reasons for not LTSing Kubuntu

Wrong, they pushed forward faulty arguments like the state of KDE4 and the 
one you mention.

> that shipping 3.5.x means they need to support it for 3 years. [..] 
> This leaves Canonical with a single in-house KDE developer in charge of 
> maintaining KDE 3.5.x, which is a considerable burden and risk.

If Canonical/Shuttleworth isn't willing to invest the necessary money to 
be able to support Kubuntu LTS for 3 years (in the worst case alone) but 
on the other side wants to earn money by selling support contracts for a 
version labelled LTS then this is a problem of Canonical/Shuttleworth's 
understanding of how "selling support as added value" works in the Open 
Source business and not of upstream/the KDE project!

About the worst case, Canonical employee claim that upstream will not 
continue to support KDE 3.5 codebase long enough but they never asked the 
KDE project and that claim is wrong. Also they would not be the alone one 
doing the work, as eg Novell supports KDE 3.5 code base as part of SLED 10 
until 2011 (extended even 2013).

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 2, 2008 10:44 UTC (Sat) by kripkenstein (guest, #43281) [Link]

> About the worst case, Canonical employee claim that upstream will not continue to support
KDE 3.5 codebase long enough but they never asked the KDE project and that claim is wrong.
Also they would not be the alone one doing the work, as eg Novell supports KDE 3.5 code base
as part of SLED 10 until 2011 (extended even 2013).

I disagree with the first part, and agree with the second one.

1. The KDE project can say that it intends to support KDE 3.5.x for 3 years, but Canonical is
being realistic here: KDE is a volunteer project. Supporting KDE 3.5.x is not a glorious job,
and hence most people will prefer to hack on 4.x. Furthermore, the KDE project cannot *commit*
in any binding sense to supporting KDE 3.5.x for three years, such a commitment has no
standing. The volunteer developers can change their minds tomorrow - one of the perks of being
a volunteer. Canonical feels that this is too risky, and I tend to agree.

2. You're right about Novell, Canonical wouldn't be alone with the problem. While it doesn't
solve the problem, it might make it much more reasonable. So I do see your point - Canonical
could have made an effort to support Kubuntu as LTS for 8.04.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 0:32 UTC (Sun) by Ed_L. (guest, #24287) [Link] (2 responses)

"If Canonical/Shuttleworth isn't willing to invest the necessary money to be able to support Kubuntu LTS for 3 years (in the worst case alone) but on the other side wants to earn money by selling support contracts for a version labelled LTS then this is a problem of Canonical/Shuttleworth's understanding of how "selling support as added value" works in the Open Source business and not of upstream/the KDE project!"

Er, are you willing to pay Canonical extra for KDE in your 8.04 LTS support contract? Are you willing to pay them at all? Do you know anyone who is?

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 8:12 UTC (Sun) by Beineri (guest, #39002) [Link] (1 responses)

I'm for sure not within the target audience of Canonical. :-) If Canonical 
has no (potential) Kubuntu LTS customers then why can't they openly and 
honestly say that instead of spreading something like "we would like to 
sell a Kubuntu LTS product but upstream/the community doesn't support us"?

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 15:38 UTC (Sun) by man_ls (guest, #15091) [Link]

Both situations are not mutually exclusive. The middle ground could be something like "we would like to sell a Kubuntu LTS product but upstream/the community doesn't support us -- to the extent that the product would be profitable". Maybe they would need to hire x additional KDE developers, but the number of expected Kubuntu customers would only pay for x-2. I think it is coherent with what they are saying.

Talking about your projections of the expected number of customers in public is not always a smart thing to do.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 4:14 UTC (Sun) by tbrownaw (guest, #45457) [Link] (1 responses)

>> Canonical was very open about the reasons for not LTSing Kubuntu

>Wrong, they pushed forward faulty arguments like the state of KDE4 and the one you mention.

Why does their presented reasoning being supposedly faulty automatically mean that they
weren't honest about that reasoning?

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 7, 2008 20:46 UTC (Thu) by jospoortvliet (guest, #33164) [Link]

If you assume they are stupid, you could argue their arguments were 
honest. I assume they're not, so they knew it was bull when they said it. 
Pretty dishonest, if you ask me.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 2:06 UTC (Sun) by dlang (guest, #313) [Link] (3 responses)

how many other projects are not going to support the current version for three years?

is samba still going to be supporting the currently shipping version three years from now?

what about Gnome, GCC, X.org, etc.

the kernel team definantly won't be supporting the current kernel three years from now (they
won't be supporting it one year from now)

if they honestly dropped KDE entirely becouse the current version won't be supported three
years from now by the upstream developers they should have dropped just about every package
from the LTS release.

holding KDE to a different standard then other software packages is the real problem here.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 3, 2008 23:04 UTC (Sun) by Ed_L. (guest, #24287) [Link] (2 responses)

"Holding KDE to a different standard then other software packages is the real problem here."

Where is the different standard, and where is the problem? There is one X.org. There is one Samba. There is one kernel. But there are many desktops. Just what totally unique "must have" feature-set does KDE bring that the people who actually pay for LTS are willing to pay for?

I'm not saying KDE doesn't have some unique features. And I've no quarrel with folks who prefer KDE over Gnome, as a personal preference. The difficulty I have is with people who push their personal prejudice to where they've convinced themselves that because Canonical won't pay for their favorite toys, Canonical's market research must be wrong -- a proposition for which no one here has yet presented any evidence.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 4, 2008 0:25 UTC (Mon) by dlang (guest, #313) [Link] (1 responses)

no distro (including Ubuntu) is required to ship any software.

if they don't want to ship KDE then they should say so, and that will be the end of it.

but claiming that they can't ship it because the upstream developers won't commit to
supporting it for three years is not being honest because very few (if any) of the other
upstream developers will commit to supporting their software for the same time.

so if 3 years of upstream support is a requirement for a package to be installed most other
packages should be ripped out. if it isn't then KDE should be included (or the real reason for
it not being included should be stated)

personally I don't use either KDE or Gnome, they both eat up to much of the system to suit me,
so it's not as if I am personally impacted by the decision in either direction. However, I am
calling BS on the stated reason for not including KDE.

Aaron Seigo talks life, free software and reinventing the Desktop (ComputerWorld)

Posted Feb 4, 2008 15:58 UTC (Mon) by liljencrantz (guest, #28458) [Link]

The other packages, like GCC, Gnome and the kernel are used both by Ubuntu and Kubuntu. I
think the point Canonical is trying to make is that there are many people who are paying for
long term support for Ubuntu, and that the cost for supporting e.g. Gnome and GCC for 3 years
is split up among all these people, butthe  number of people willing to pay for Kubuntu
support for 3 years is somewhere in the vicinity of none, making it economically unsound.