I'll give them credit.
I'm not, so much, concerned about the bugs: I know they'll be addressed. I'm more concerned about the utter lack of usability testing, or perhaps lack of usability testing feedback.
Delays continue to plague development of the popular Fedora Linux distribution, with Fedora 18's original November ship date now pushed back to January 2013 at the earliest. Ordinarily, the Fedora Project aims to ship a new version of the OS twice annually, with new releases arriving on the Tuesdays closest to May 1 and …
Given the proposed names of the next Fedora after this one, I shall follow Peter and Erwin and propose a law...
"In any discussion of any facet of a GNU\Linux distribution on The Register's Forums, the probability of a post remarking on the un-usability of the default desktop environment will approach unity from below within log(article wordcount) posts"
It's about bug squashing in the installer and updater chaps.
Fedora is the test bed for a whole series of ports of the RedHat distribution. Even Oracle takes from RedHat, and then undercuts RedHat in attempts to hurt RH. That is, Oracle steals from it's mother.
As far as interfaces is concerned. In the first month of use, I went from disliking Gnome3 to getting more familiar with it, and looking for the good stuff in it. I am now certain that Gnome3 and Unity are going to persist, while KDE will come in a close third. Other gui interfaces such as xfce will also be next to KDE.
I use both Ubuntu and Fedora, and really, can't choose between them for superiority. So, give credit for being able to download free Linux Distributions. How much less cost than that can you have?
Ubuntu is now realizing that you need a revenue to be able to continue to provide a free distribution. In this regard, they are building in interfaces to the Ubuntu Store, to Amazon, etc. This should not detract from the excellent distribution that it is. Fedora serves as the test bed for RedHat.
By the way, both distributions are looking to China, where a billion people will result in many many adopters of either. China also has it's local distributions, and we should look to see some of them made available in English.
I commend you for being easily amazed [somehow the patent on exercising a cat with a laser pointer comes to mind, uncalled-for], but....
Did I say anything about "extended support"?
Did I say something about Cannonball-style-barreling-down-the-upgrade-highway-with-bug-police-in-tow?
Oh dear having trouble expressing yourself clearly are you?
You had a whine about the release cycle being too short for you to "get used to" each version.
You said precisely bugger all about any issues with bugs.
Given nothing actually forces you to upgrade to the new version (despite oooh the version you are on going out of support) it seems the only person who can't see that your problem is to do with the support life cycle is you...
Oh yeah and I must have missed the definition of fanboi extending to encompass anyone who's ever glanced at the release timeline ever.
> You said precisely bugger all about any issues with bugs.
There may be special needs involved here, did you see that this was an article about Fedora. And bugs. And Fedora being delayed because of bugs?
> it seems the only person who can't see that your problem is to do with the support life cycle is you...
Holy! It's good that we have the Internet where perceptive persons can tell me the world is ok and the problem is actually with me.
RHEL 7 is scheduled for the second half of next year, so presumably the reason they keep delaying it instead of kicking these features to Fedora 19 is because they want them in RHEL 7.
If I were deploying RHEL, this wouldn't make me eager to push out RHEL 7.0 anytime soon, as one wonders what other lower profile features are being crammed into Fedora 18 when under normal circumstances they'd be pushed back to Fedora 19, simply because they "must" be part of the next release of RHEL.
Not releasing untested, unsecure and functionally deficient software? Come on, what are beta testers, sorry, end users for?
They should be commended for holding things back rather than just keeping chucking unfinished work out? No that any other software company would do this to anyone would they?
Ever since Fedora tried to switch a 6-monthly schedule (either because Ubuntu does or, more likely, they want to release just after each GNOME release - well, the GNOME 3 UI disaster has put paid to that reason, IMHO), they have utterly failed to release *any* alpha, beta, RC or final Fedora on time.
The main schedule pressure for Fedora shouldn't come from Ubuntu or GNOME, but actually from RHEL of course. Considering they release RHEL every 3 years, that would still leave 4 * 9-monthly Fedora releases, so I think Fedora should move to a 9-month release cycle instead of 6.
BTW, am I the only one worried that F18 RC has actually been cancelled?! Cue loads of panicky post-F18 final fixes being released...
The fact that they don't release it at that point should be a positive point. Unlike the bug-laden pigs that have been some of the more recent Ubuntu releases (by that I mean anything post-10.04), precisely because of the ridiculous, must "release every 6 months and must have features X, Y and Z".
Features, Quality, Time. Choose any 2.
You are right that Fedora should go for 9 monthly schedules. Alternatively,. if Fedora had more developer and Q/A resources, they would be able to stick to six monthly cycles.
When you go to 9 month release cycles, there is a tendancy to overload the next cycle with too many new things. In fact, smaller steps of 3 months, with fewer new things might be better. Then there are the ones who say, continuous improvements are best (Debian style).
I think that Fedora takes on too many new features in the 6 month schedule, thus giving credance to your suggestion.
It may very well be that Fedora was late in starting to look at secure boot. Perhaps they should have started one year ago, and not just after Fedora 17. If Secure boot is a success, other distributions will share the code and follow suite. Yes, we have a delay. By the way, I have installed F18 on a clean disk, and for the most part it works just fine. I really find that the applications work as well as ever. It is delayed because of secure boot , the problems with the logon interface and security checking (sudo, logon after timeout and lock, etc.)
Two releases/year? FFS, I don't want this kind of moving target! Software up to date? Yes. Security updates? Hell, yes! But TWO releases/year?
Too much pressure on the devs, too much hurry to assemble the beast, too little time to testing and adjusting.
OpenSuse tried the 6 month cycle. Just returned to the annual release - thanks God. It seems Fedora is struggling too. Time to back pedal a bit, and release once each year.
Something that could be done (already was, in fact) is to create a second repository, with more up to date packages. OpenSuse does this, with the project "Tumbleweed". I'm sure other Distros do the same, with another name.
On could, even, create various repositories - each one for a different group. One for KDE, another for Gnome, a third for XFCE, and so on.
"In October, the Linux Foundation proposed a "stop-gap measure" that uses a Microsoft-authorized digital key to allow Linux to run, but the Fedora Project board insists on a solution that works without the Microsoft key."
Here's a solution: get a UEFI key from another supplier and sign your bootloader with it. It's not a "Microsoft key" - MS just provide an easy (relatively cheap) way for Linux vendors to get one. It costs $99, which goes to Verisign, not MS (http://mjg59.dreamwidth.org/12368.html). Canonical look as though they might eat the expense of setting up their own key infra, and good for them (http://www.phoronix.com/scan.php?page=news_item&px=MTEyNDY).
It's not a Microsoft key. They just made a nice and easy way for Linux to get signed. It's Secure Boot key, part of the UEFI spec, which was started by Intel as EFI and is owned by the UEFI Forum (http://en.wikipedia.org/wiki/Unified_EFI_Forum). Microsoft are part of that forum, but so are IBM and Apple, and others.
Imagine if MS provided a way to put a Verisign key in a web server, and that enabled HTTPS on that server. And without that Verisign key, unless you just bought one from another supplier, you couldn't serve HTTPS without it giving a warning in a user's browser. And imagine you could use nonsecure HTTPS (what would that be called?) if you didn't want to use HTTPS?
Would people say this: "I mean yeah sure - for now you can use non-secure HTTPS, but because some websites are deemed "secure" users will soon stop using the non-secure sites. This will kill Linux on the desktop, because otherwise it'd have been a sure thing next year."
They probably did. Turned out fine.
Biting the hand that feeds IT © 1998–2020