Idea
#10225: Promote Shuttleworth OSS pulse, on project at a time
|
| |
23
|
|
|
Written by mangar the 24 Jun 08 at 06:34.
Category: Others.
Related to:
Nothing/Others.
Status: New
|
|
|
Description
Mark Shuttlework ( http://www.markshuttleworth.com/), offered that all of the OSS projects will have a pulse:
Libraries, foundations (gtk+, qt, x.org, kernel): A stable release every 6 month to a year.
Applications: A stable release X months after the libraries released.
Distros: A stable release X time after Applications are released.
This way, creating a distribution, and developing for Linux and the OSS software stack will have a predictable "pulse", where everyone can collaborate, share patches, and use the same base libraries, instead of the current situation, where each piece of software uses it's own version of the base stack, and the maintainers and packagers runs around like headless chickens, trying to tie everything together, even the Kernel guys approved, since it allows them to know when to concentrate on bug-fixes releases, and relieve them of the need to support an unknown number of versions of the kernel.
There has been some controversy - Aaron Siego doesn't want to tie development to hard time-lines, the Redhat people see it as a way to leech QA resources from them, etc.
Since I believe it is a great idea, that will bring some order and predictability to the OSS stack (that's a good think - it allows developers to have stable building blocks, and users to have the latest versions of software in each release), I suggest that the Ubuntu project will contact each project, one at a time, and bring them into the Pulse, until the majority of project will be synchronized.
Attachments
No attachments.
Duplicates
Comments
|
Ssdg wrote on the 24 Jun 08 at 08:09
|
And loose the schedule freedom of devs? no way.
Let me clarify my point of view.
Do you imagine that firefox would be this awesome success if mozilla released it on a scheduled time instead of now that they think it's ready?
Mark's point of view is the one of a packager some people think there is another way to do it especially with project version managers (like CVS, SVN and bazaar) the distro create a branch the moment they want to freeze the code and they work with this branch. so the project is free to continue to it's goals of quality without being submitted to schedule imperatives.
So, I'll vote -1 because I think mark's idea is intrusive into projects life and because I'd rather get a stable version 5 month late than a pseudo-stable at ubuntu release.
|
|
mangar wrote on the 24 Jun 08 at 08:37
|
@Ssdg
when the alternative is to release beta software as a main component of the desktop (firefox in Ubuntu's case, X.org 1.4.99 in Fedora's case, tons of libraries in all of the distributions case), then yes, a regulated Pulse is required.
If a piece of software is 5 months late, than it can be postponed to the next time slot.
|
|
Ssdg wrote on the 24 Jun 08 at 09:00
|
I meant 5 month after the software release, like FF3 which was released on june while ubuntu was released on april, we as ubunteros should have get this on october, ie 4 month late) (10-6=>4 month).
And I doubt mozilla will submit to this pulse because of their moto: release when ready. (I read standblog.org and I only know the french version of this, so sorry it might be poorly tranlated).
So as far as I'm concerned and as a devellopper myself, I vote this down because I know my work is better when I work on it without deadlines. (but I work faster with deadline, to be honest)
|
|
stoffel wrote on the 24 Jun 08 at 09:01
| |
I like the idea to synchronize release cycles, but I voted -1 to this entry because I think Ubuntu actively promoting the idea will have the reverse effect.
|
|
Vadim P. wrote on the 24 Jun 08 at 12:21
| |
All linuxhaters's fans must vote +1!
|
|
danbhfive wrote on the 24 Jun 08 at 14:08
|
I like this idea. There are projects that are already time based, and there are projects that arent. I think that comments that are pointing out non-time based projects are missing the point.
If a project is time based in its release schedule, then it should be in sync with any other project that is doing the same. If it cant, it cant, whatever.
For all those projects where they can, this is a great idea.
|
|
nevermind85 wrote on the 24 Jun 08 at 14:22
|
I'm voting +1. I like the idea, while being a pseudo-developer myself I know what it's like to have a deadline and you no one wants to release something to the public unless it's ready or you work at Microsoft.
Projects don't need to be tied up to this deadline, currently most projects have at least two branches: stable and last-build. If they can just try (emphasis on try) to release new stable versions of their software within a given time-period; distros will be able to include the new version on their new releases. Anyone can get bleeding-edge packages if they wanted to.
|
|
Auzy wrote on the 24 Jun 08 at 14:36
|
And will Canonical donate any resources to development teams to help make this happen? Or will we just promote it, without actually helping? Its all good and well for Shuttleworth to donate 10 million dollars, or fly space shuttles, but that doesn't actually help this idea.
Project speed is mostly dependant on the number of developers. And Mark wants to speed it up, but isn't contributing full time developers to the projects. Instead my (possibly incorrect) assumption is that they are all working on Canonical projects such as the netboot. That wont help speed up development.
Sorry, but I cant vote for this. I like it, but I have seen little indication that Shuttleworth is serious. If there was an indication that he has hired as programmers, I will happily vote +1.
But as I don't believe people have been hired, then it seems that even mark isn't 100% behind it himself.
|
|
Vadim P. wrote on the 24 Jun 08 at 16:05
|
I think it'll be fair to claim that when you'll need to pay for Ubuntu.
'till then, we can't claim anything.
|
|
Auzy wrote on the 25 Jun 08 at 01:04
|
Yeah, but my point is,
1) Am I right in assuming that Shuttleworth donated money to Canonical.
2) Shuttleworth endorses the idea.
3) So does that mean Canonical is essentially snobbing him by not implementing it?
|
|
Vadim P. wrote on the 25 Jun 08 at 12:46
|
Shuttleworth didn't donate money to Canonical. He's single-handedly funding the whole company (who is not self-sustainable yet, so it's all going into his pocket)
2) yeah
3) um no, it's everybody else who's snubbing him (ie, the projects)
|
|
deadowl wrote on the 25 Jun 08 at 12:50
| |
Since it's Shuttleworth's idea, I'll just go ahead and take it for granted by voting it down. Not that I disagree with it, just that I'm annoyed with the ignorance in this thread.
|
|
Auzy wrote on the 25 Jun 08 at 13:05
|
He wants to ensure consistant releases. This requires a lot more permanent resources.
You can't tell part time developers who don't get paid that they need to get something done on time. They do have lives outside of coding. Its why projects like xserver exceeded the expected time (Comp sci 101, 80% of software projects end up failing).
Is canonical willing to throw more permanent resources in, to ensure that projects stay on time when they are falling behind? Otherwise, its just a statement, its no better then the ETA's we have now on projects without enough developers (Grub2,Xserver). Thats what I'm saying. It will make no difference in keeping most projects on track
|
Post your comment
|
|
|
