The end of March is near, and in about a month, we will be seeing Lucid Lynx, or Ubuntu version 10.04.
After finally cutting my last machine over to Karmic a few weeks ago (my X60 tablet), I’m wondering if Ubuntu is developing new versions too fast. I know some may think how can development be bad or too fast. The answer is simple: stability (or lack thereof).
The main reason why I waited to cut over my last machine was because I was worried about having to tweak my machine so that I could get my tablet functions working. It took me about an hour to work through, and I got it pretty close to perfect. (I can’t get one feature, auto-rotation when in tablet mode, which was working in Jaunty to work in Karmic.) The other reason why I waited was because, aside from a couple of issues, my machine working fine, and I didn’t feel the need to upgrade. After all, regression sucks.
But I did finally bite the bullet because I noticed my personal T61P at home was running much cooler and more efficiently after I cut it over to Karmic. I also had a problem with that bone-jarring alert beep occurring on shutdown or when I over-deleted/backspaced (entering one too many of either) in menu dialogues with Jackalope on my X60. The sound was so annoying I became like a Pavlovian dog; after a few days jumping out of my seat (because of the alert) after beginning the shutdown cycle, I learned to turn off my external speakers before doing so.
So what’s my assessment of Karmic? Actually, on my personal machines, I have had no major issues with it. However, after my last cut-over, my feelings are mixed. I have one issue with Karmic and dual-head mode. Karmic seems to have better monitor detection capabilities and allows me to extend my desktop over two monitors. The only problem is that if Compiz is running, X routinely freezes up. After many months of routine use my X60 (I plug into an X6 multi-base in my office), I’ve had to learn to do things differently. I have to disable Compiz before plugging into my base. Not a big deal; however, after many bouts of forgetfulness, I’ve needed to restart at lest two dozen times.
So, for the first time in my computing life, I have decided to run alpha software to see what Lucid Lynx has to offer. Since I was planning on fresh installs anyway, I decided to upgrade to the 64-bit version instead. I started off slow. I put Lucid Alpha 1 on my other Thinkpad (a T60 with integrated graphics). Amazingly, things worked quite well with little tweaking. It is alpha software so there were a few glitches here and there, but very minor. Since it worked so well, I decided to install Lucid on my T61p at home (this one uses NVIDIA graphics); the Alpha 3 was already available, so I installed it instead. Aside from some issues with the current NVIDIA drivers, the experience has been very positive as well. I began thinking, Wouldn’t it have been better to hold off on Karmic and wait for Lucid instead?
I also began thinking about Ubuntu’s quick release cycle. Do they need to release new versions every six months, or should they give themselves more time in between versions to try and release something more refined say once every year? I’m not a programmer and have the faintest idea what six calendar months equates to in computer software development time, but after using every version of Ubuntu since Feisty Fawn (not to mention several other distros in between), I find myself wondering if it is too much. Couldn’t the upgrade cycle be cut back to allow for partial upgrades of the current version instead. For example, instead of going to a new version of the entire OS, couldn’t new versions of apps like Open Office or Firefox be moved into the repos instead? (I know, things we can, as Linux end users, do anyway.)
Why not give the Ubuntu developers time to work out the kinks and get things more stable? Don’t get me wrong; this isn’t some off the cuff rant because I am angry or disappointed with Ubuntu. In fact, this line of questioning is quite the opposite. Like every good citizen of every country, contrary to what many conservatives in this country believes (they call it unpatriotic), we have a responsibility to ask questions and critique the status quo to see if there is room for improvement. The only way things stay at the top is if it is willing to adapt and change.
I know. You must be saying isn’t that what the six month cycle insures. Well you’re right. It does insure a new version every six months, and seemingly, a change in the status quo. The problem is of course time and quality. Look at Microsoft as a prime example. Although their release cycle is very slow and not a good model; how many times have they released a sub-standard product because of profit margins? Vista was a joke because it was poorly designed, too ambitious (in terms of driver development and third-party interest), and was on a timeline that it could not keep up with. So what did MS users get? A product that no one wanted to use or move to (like Windows Me). People were clinging on to XP or jumping ship to Mac and Linux. I had never seen so many new Macs in my life until Vista was released. (Which may not be a good thing either…but that’s another post for another time….)
Well, Lucid is still in alpha and we won’t know how good it is until it is finally released in April. So far, aside from moving the minimize, maximize, and close icons from the right side to the left, a la Apple (something that can be easily moved back), it has been very positive and promising. The load times seem to be faster–although nowhere near ten seconds–so far I’ve been averaging about 30-35 seconds to working desktop, and I do like the new themes and splash screen they’ve installed. I do like the fact that the apps seem to have finally caught up; all of my favorites, Firefox, Thunderbird, Kompozer, etc. are there in their newest versions (sans the Gimp…it is in the repos and can be easily installed).
I look forward to when Lucid is finally released in a few weeks, and regardless, I will be a happy Linux user. But while using whatever version I finally decided upon come that time, I will still be pondering whether Ubuntu could benefit from a slightly less ambitious release cycle. Cutting edge is good…bleeding edge…not so much….
[...] Here is the original post: Should the Ubuntu Release Cycle but Cut Back? « The Chronic Agnostic [...]
A fast development/deployment cycle like this is confusing for new people. I’ve been using Ubuntu for a few years and finally sat down to read a little about the release cycle. I didn’t understand the LTS at all. Just thought new stable versions were being released. I’m sticking with 8.04 until a few months from now.
I guess the LTS version is Ubuntu’s way to keep something stable and well-developed in place. Overall, I think what Ubuntu releases is stable; the LTS cycle may be a little too long (but a good idea to have for enterprise deployment) to wait for the latest and greatest. I just six months may be a little too quick…maybe a yearly release? You could always have special versions if you wanted to release something really cool and divergent before the traditional release cycle.
Thanks,
J.
If you want stable stick with LTS releases. Just don’t drag the rest of us down with your whining. I look forward to each release and run it from Alpha 1.
You can still download and install 8.04 and it still has 1 year of support. You have no reason to complain. If you want newer apps running on a stable base then use the backports. That’s why they are there.
Yes, as I stated in my original post, I know newer apps can be installed on older releases…I don’t think I’m whining…I’m critiquing…I’ve been able to roll with each new release with little issue; however, know of some who have struggled with updgrades because of regressions. And quite honestly, I look forward to each new release too because it’s exciting to see what the developers have done. Just wondering if there’s a better choice between 6 months and the LTS cycles.
Sorry to drag you down. Not my intent…
J.
I’ve been installing each Ubuntu version as they’ve been coming out since I started using Ubuntu nearly a year ago. Now that I’ve committed to using only Ubuntu on my home system, I’m feeling paranoid about throwing my system into disarray every time a new release comes out.
I think I agree with you that such a fast release cycle is not optimal. I want to upgrade to 10.04, but I’m a bit scared of doing it too quickly. Have they solved the bugs with monitor switching when compiz is on with intel graphics cards? Do hibernate and suspend work better in this coming version? Are any of my packages going to break on upgrading? That stuff intimidates me!
Hey Inkhorn,
Yeah, those are good questions. I’m sure people have others. I know the LTS releases are there for stability, but I always feel compelled to move when the new versions come out. I have talked to others who feel that Karmic wasn’t a strong release. Just wondering if more time in between versions would bring more stability and fewer regressions. And again, I’m not questioning because I’m unhappy, but because I want to see a stronger end product. I’m not an Ubuntu guru, but I’ve been able to tweak when I’ve had to. Aside from Karmic, I’ve been generally happy when upgrading to new versions.
BTW, how’s the new job?
Thanks,
J.
Do you have your home directory in the same partition as your OS directories or in a separate one? Following that, when you install a new version of Ubuntu, do you do a distribution upgrade or do you do a fresh install? I like a little bit of change once in a while but I’m afraid of big trouble coming my way.
Because of my new job (data analysis for a non-profit charity organization in Canada) I’ve not been using my own computer too heavily for the last month. It’s been weird, going from being the master of my own computer domain to being relegated to the permissions of a network user who can’t install anything without the explicit permission of the IT directors. It will probably take some getting used to!
Otherwise the new job is suiting me quite nicely! Academia was getting too stifling for me (the time commitment was soul-sucking) and now my new job gives me a decent paycheck, much opportunity to play with/analyze data like I enjoy, and a much better life-work balance!
Inkhorn
Hi Inkhorn,
For the most part, I do have my home directory in the same partition with my OS. I have one box where I have them separate. Because I have access to a network drive, I really don’t have too much in my home directories anyway.
I have done distribution upgrades and fresh installs in the past. I prefer the upgrades because I don’t have to redo all of my configuration files; however, at times, I have made some radical changes that required a clean install (e.g. moving to 64 bit).
I’m glad you’re liking your new job. I agree with academia being a huge time commitment without much in a paycheck!
Aloha…
J.
Too fast? I think it’s too slow … How needs stability, the LTS versions are just wonderful and the time between LTS versions can’t be called “too fast”, I think
Ubuntu is still lacks about “bleeding edge” some distros can give, so I can’t say it is too fast, and for sure, in production (both server and desktop) I install only LTS versions which – surely – needs to be “very” stable. Maybe the time before the LTS release can be greater, yes. But not between “ordinary” releases.
Hi LGB,
I guess the LTS releases really is the answer for those who yearn for stability. I just can’t convince myself to stick with an ‘old’ release knowing there’s something new available.
Thanks,
J.
[...] After finally cutting my last machine over to Karmic a few weeks ago (my X60 tablet), I’m wondering if Ubuntu is developing new versions too fast. I know some may think how can development be bad or too fast. The answer is simple: stability More here [...]
I have to agree with your post, and feel rather annoyed with the commenters attitude toward your subject and the write-off that the LTS is a viable solution for the desktop edition.
LTS Desktop is a joke. It is solid and appropriate for enterprise users and for server users, but for the desktop end user, it’s not really an acceptable option.
When you step back and look at the larger picture, you see Canonical dealing with two sides. End users that want stability and sanity, and software developers that want to push the capabilities limit at all times. So what you find is a group of end users that want stability, but then find themselves not capable of doing something that they can do easily on a Mac or Windows. so when they try to stick with the LTS version, they find that if they want some software to do X they have to have something new which means upgrading…which means breaking things, etc.
I’ve used Ubuntu off-and-on since Dapper. Dapper was nice, it wasn’t cutting edge but I couldn’t stay with it (I had to dual boot) because of software limitations. The first viable “use all the time” product was Feisty, which pulled me off of Dapper after testing Edgy as nothing special. Then when Hardy came along I was psyched to get onto another LTS edition. But the choices made for Intrepid (X.org, etc) basically made Hardy a worthless product for the Desktop edition. Why? Because software developers said, Intrepid has the new stuff so we can build our software with dependencies on it. If you look at the current backport products for Hardy, they are extremely limited.
I guess what I’m trying to say is that just like your comment, there isn’t a viable option between stability and “cutting edge” that they are trying to gather on the 6 month cycle. And what this means is that they are alienating (or bare minimum annoying) a majority of the users that can push the product more main-stream. If Intrepid wasn’t released and was in beta another 6 months, software developers would keep ensuring their product worked with Hardy and the end user would have been happy. Based on the past, there has always been at least 1 worthless release every year…and it always seems to be the October release. Edgy couldn’t pull me away from Dapper, Gutsy was a buggy beast compared to Feisty, Intrepid was horrible but the changes forced me into use, Karmic was not half bad but nothing special. Now it seams Lucid is a great upgrade. To me, they should drop the October cycle or just make it a “maintenance install” with some of the new features, but still based on the previous cores.
I think it’s a great idea to reevaluate the release cycle. Like you said, have 6 more months to work out the kinks in an “alpha/beta” mode and you’ll have much happier end users. Then the people that want the cutting edge can run the betas, that’s kind of the benefit of running a beta right? You get the new stuff if you can deal with the bugs?
Hi Trae,
Thanks for the comprehensive and well argued response. I agree that the LTS versions seem geared towards enterprise applications. Your typical end user probably will want to upgrade to the newer release to have the benefits of the upgrade. But, as you so aptly point out, there is a chance something will break–even stuff in the repos (e.g. Kompozer in Karmic). I guess, what I’m really suggesting is what you’re suggesting above–partial upgrades that eliminate the need to wait for the new version to fix bugs (and the need to backport) and having a longer beta phase to polish the final release. After all, it’s about quality…not quantity, right?
Thanks again…
J.
I’d like to see “brand new features” during one release
and “maintenance updates” during the second release.
“Security updates” and “production limiting updates” [see note] could happen at any time.
One “mistake” I see is marking a routine release for Long Term Support. An LTS release must include a set of features that all work with minimal opportunities for instability. “The latest release” might not meet these stability and operability goals.
__________
Note: “Production Limiting” situations are those where software defects prevent an end-user from using an application for its stated purposes. Clearly application crashes and wrong answers fall into this category, but also situations where the application mostly works but certain options or config details fail to behave correctly.
Hi Dan,
Yeah that sounds like a great way to use the current release cycle. And you’re right about the LTS releases; they do have to conform to certain specifications which may require cutting edge applications/technology get relegated to the non-LTS versions–because of some forms of instability.
Thanks,
J.