How is this different from APKs on Android?
In a move that could see Ubuntu veer even further away from the Linux mainstream, Canonical has proposed a new software packaging format designed to make it easier for developers to publish apps for Ubuntu's tablet and phone–friendly future incarnations. "This is not aimed at changing packages that are already part of the …
This could be done with deb packages and a small extension to the dpkg/apt system allowing non-root installs.
Sure, it probably wouldn't be as fast as a whole new packaging format. But why reinvent the wheel for a small improvement in application install time?
My alternative approach would involve:
- A specification for "user-debs", stating no dependency other than the Ubuntu API, no scripts, and all files in one directory. Also probably a flag in the manifest indicating it is a "user-deb"
- A modification to dpkg/apt to allow them to be run by non-root users (for "user-debs" only)
With Canonical's current obsession with reinventing the wheel, I don't doubt they will go ahead with a brand new packaging system.
> a small extension to the dpkg/apt system allowing non-root installs.
Fedora et al. use a policy to control installations; I often set up customer boxes to permit non-root installations of software from trusted erpositories if the user is at the console. That's a config file of half a dozen lines...
What Canonical is punting is nothnig new, just done differently to everyone else in the multiverse. Now who'da thunk it?
"so it should come as no surprise if Ubuntu bases its mobile app packaging system on homegrown code"
Ah, homegrown is one thing, proprietary quite another.
It seems to me as if Canonical is heading in the direction of "We invented it, so we control how it's used and how much it will cost you."
If that turns out to be the case, and I will wait for further developments before judgement, then how is that different from "open core" as practised by the likes of Oracle?
As a worst case scenario it looks as if Shuttleworth may be trying to recoup some of his investments by piggy-backing on Gnu/Linux and charging for pay-for extras and facilities. The next step of course would be to head down the subscription route. Not very nice, but a potential pot of gold at the end may well be too much temptation to resist.
That's not to say that this is what's going on, but from a pessimistic viewpoint it is worth watching out for.
Ubuntu was going to be a success on the desktop. Didn't happen *
Ubuntu was going to be a success on Netbooks. Didn't happen.
Ubuntu is going to be a success on Tablets. Hasn't happened.
Ubuntu is going to be a success on Mobile Phones. Won't (IMO) happen.
This is a solution for a problem that will not arise.
* Unless you count a percentage of (by most accounts) 1% of market share a success.
The poster did say a percentage of 1%.
The Linux zealots here and on other sites keep telling us a large choice of Desktop Environments/Window Managers is "a good thing". So given the number of the latter plus the number of Distributions it's likely that Ubuntus' Market Share is a small percentage of 1%
Hmm... One of the things I find most irritating about Linux, which I generally really like, is the fragmentation. The fact that Ubuntu and RedHat may as well be different OSes, like AIX and HP-UX. I really wish that the major distros would get together and sort out package distribution, startup procedures and locations, etc. It would do a lot for the OS as a whole.
(Oh, yes, I know Linux is the Kernel, but it's a nice easy shorthand for the OS name)
Biting the hand that feeds IT © 1998–2022