LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Distributions (https://www.linuxquestions.org/questions/linux-distributions-5/)
-   -   Ubuntu Based Rolling Release Planned! (https://www.linuxquestions.org/questions/linux-distributions-5/ubuntu-based-rolling-release-planned-4175449554/)

linuxPCplus 02-10-2013 10:33 AM

Ubuntu Based Rolling Release Planned!
 
I am considering the possibility of creating a FULLY Rolling Release distro based on Ubuntu. This project will be sponsored & funded by my company Linux PC Plus. But in order to make this vision a reality, we need volunteers! We need developers, programmers, graphics designers, etc. We especially need folks with experience creating/contributing to Linux distributions. If you have helped develop any Linux distro (especially rolling releases or Ubuntu based distros), maintaining repositories, etc, PLEASE CONTACT ME!!! I understand this is a HUGE task & it will be a difficult challenge to undergo. But with the right people, I believe it CAN BE DONE! If you are interested in contributing & being a part of this exciting project, please email wolf@linuxpcplus right away. Include a brief summary of your experience & why you want to be involved.
MORE INFO WILL BE AVAILABLE ON http://linuxpcplus.com IN THE NEAR FUTURE.

unSpawn 02-11-2013 07:45 AM

Quote:

Originally Posted by linuxPCplus (Post 4888504)
I am considering the possibility of creating a FULLY Rolling Release distro based on Ubuntu.

Just as Ubuntu devs are pondering the very same idea? That's quite a coincidence.


Quote:

Originally Posted by linuxPCplus (Post 4888504)
we need volunteers! We need developers, programmers, graphics designers, etc. We especially need folks with experience creating/contributing to Linux distributions. If you have helped develop any Linux distro (especially rolling releases or Ubuntu based distros), maintaining repositories, etc,

Just being curious, given your long list, what would be your role? (Please don't say "management" ;-p) What are your skills? Do you have something tangible to show people who would be interested investing time and effort?

DavidMcCann 02-11-2013 12:54 PM

Perhaps I being slow here, but does it really make sense to base a rolling-release on something that isn't? If you want a rolling-release distro that's like Ubuntu, shouldn't you do what Ubuntu do and base it on Debian Testing? But isn't that already done by LMDE?

linuxPCplus 02-12-2013 12:56 PM

Let me rply to unSpawn first;
No, not really coincidence at all. It is something I have pondered for a long time. Although, I will admit the recent stories about Ubuntu did put some fuel on my fire. Of course, as you know, Canonical has made it clear they have no such intentions for Ubuntu.
My role? Well, first I wll be honest, I am NOT a Linux Developer/programmer, just an enthusiast with more than 10 years Linux experience (though I am currently preparing for my LPIC exams) and 30 years computer experience. I am web designer & a small business owner with primarily Linux based business.

As such, I would provide the web design for the project, the main servers for the distro itself & it's repos, & a good percentage of financial needs for the project. I would also provide all marketing & promotional needs for the project. I would work directly with the core team of developers on the project to help decide the direction of the project. I will possibly contribute artwork as well.

My leadership role would likely be minimal as I would most likely assign leadership roles to qualified participants. I know my skill set & would not pre-doom this project by taking on responsibilities I am not qualified for.

I will not BS anybody with bloated stories about my skills & experiences. I will never make claims to skills I do not possess. This is exactly why I am trying to gain interest & find a team of folks who DO have the skills that will be needed & whom would like to see such a distro released.

Please note, as I said in my original post, more info will be available in the very near future on the http://linuxpcplus.com website, so please watch there for further details.

Now, to reply to you David.

LMDE is a partially rolling release (and there is some debate about that), not a FULL RR. I am active on dozens of Linux forums, blogs, etc & I have seen so many folks showing a desire for a RR based on Ubuntu. Currenlty, the closest thing to an Ubuntu based RR is NOT LMDE, it is Bodhi. But even it is only a partly rolling release.

This distro, if ever realised, would offer the standard Ubuntu repos as well as it's own distros. This is similar to what other distros such as Bodhi & Manjaro do. Both have their own distros, but allow access to their parent distros repos as well (Manjaro, based on Arch cannot access the Arch repos, but can access the Arch User Repository-AUR).

The distro's own repos would offer the latest stable or semi-stable packages that are not yet available in the Ubuntu repos. For example at the time of this writing neither LibreOffice 4 nor KDE 4.10 are available in the Ubuntu repos. Both are stable & official releases, but you will have to wait until at least 13.04, likely 13.10 for them to be in the Ubuntu repos. Sure, a user can manually add the ppa's & then install them, but an rr like this would make it so much easier, automatic in fact.

There are also many Ubuntu users who hate the fact that every 6 months they have to do a new installation in order to have the latest version of Ubuntu. I hear this complaint ALL THE TIME! This new RR I am planning would allow people to keep everything they love about Ubuntu, but would also mean they never have to reinstall it. Install it once, configure your update options, & you will always have the latest version.

I do not want to give too much away right here, but I am also happy to answer any questions or comments! The whole point of this post (which is repeated on several Linux forums) is to measure interest. So, far, interest seems pretty strong between all of the platforms this post exists on.

As I said, please watch the website for further details. I plan to post more there in the next week or so.

unSpawn 02-12-2013 02:15 PM

Quote:

Originally Posted by linuxPCplus (Post 4890076)
The whole point of this post (which is repeated on several Linux forums) is to measure interest. So, far, interest seems pretty strong between all of the platforms this post exists on.

Good to hear that. Good luck with the project.

k3lt01 02-12-2013 03:16 PM

If you are basing it on Ubuntu the only way you can really do it is to keep the repository pointing to testing but Ubuntu doesn't hava testing so you will probably need to create your own repositories and pull Ubuntu's development release into that on a daily basis. Even then it will only be semi rolling much like LMDE is. Good luck with it though I'll be keeping an eye on the progress.

linuxPCplus 02-13-2013 12:06 AM

As I stated, the distro will have it's own repositories in ADDITION TO the standard Ubuntu repos.
Thank you for watching!

TobiSGD 02-13-2013 06:07 AM

Quote:

Originally Posted by linuxPCplus (Post 4890446)
As I stated, the distro will have it's own repositories in ADDITION TO the standard Ubuntu repos.
Thank you for watching!

You will definitely run into problems with that approach on any distribution with automatic dependency resolving. This is basically the same as mixing Debian Stable with Debian Sid, which will for sure break the system.
The whole point of a (true) rolling release distro is to always have the latest packages in the repositories. This does not work well with the approach to also have older packages in the system.

I also think that you will have to seriously limit yourself, at least at the beginnings of the project to the point you have a reasonable number of package maintainers. Almost any package in Ubuntu is already not the latest version at release time (with very few exceptions), so you will need package maintainers for at least any package that comes by default with your distribution, but that will only be of use if the user does not want to install additional software, in which case you need additional package maintainers for the additional packages, since those packages in the Ubuntu repositories will also not be the latest.

Don't get me wrong, I do not want to discourage you, but I think that you have not thought through all aspects of rolling release systems. There is a reason why all true rolling release distros are independent and not based on distros following a release cycle.

DavidMcCann 02-13-2013 12:20 PM

Mixing repositories can be a mess. As a CentOS users (very small repo) I've got software from a number of repos but I'm using the priority-setting plugin in yum. I think you can get a similar protection with apt and it would need to be set by default.

I get your point about the demand for a rolling-release distro in the Debian family but with the user-friendliness of Ubuntu, but I still think you'd do better basing the repository on Debian Testing.

linuxPCplus 02-13-2013 07:57 PM

Quote:

Originally Posted by TobiSGD (Post 4890658)
You will definitely run into problems with that approach on any distribution with automatic dependency resolving. This is basically the same as mixing Debian Stable with Debian Sid, which will for sure break the system.
The whole point of a (true) rolling release distro is to always have the latest packages in the repositories. This does not work well with the approach to also have older packages in the system.

I also think that you will have to seriously limit yourself, at least at the beginnings of the project to the point you have a reasonable number of package maintainers. Almost any package in Ubuntu is already not the latest version at release time (with very few exceptions), so you will need package maintainers for at least any package that comes by default with your distribution, but that will only be of use if the user does not want to install additional software, in which case you need additional package maintainers for the additional packages, since those packages in the Ubuntu repositories will also not be the latest.

Don't get me wrong, I do not want to discourage you, but I think that you have not thought through all aspects of rolling release systems. There is a reason why all true rolling release distros are independent and not based on distros following a release cycle.

You may very well be correct. This is why this thread exists, to get the discussion going. You are likely right that I have not thought thoroughly enough about about how a rolling release works. This is why I am so glad these threads are garnering this much discussion. It helps identify potential problems & hopefully ways to address them.
Please see my reply to DavidMcanns most recent comments about the project for more thoughts.

linuxPCplus 02-13-2013 08:02 PM

Quote:

Originally Posted by DavidMcCann (Post 4890903)
Mixing repositories can be a mess. As a CentOS users (very small repo) I've got software from a number of repos but I'm using the priority-setting plugin in yum. I think you can get a similar protection with apt and it would need to be set by default.

I get your point about the demand for a rolling-release distro in the Debian family but with the user-friendliness of Ubuntu, but I still think you'd do better basing the repository on Debian Testing.

Not surprisingly, this seems to be a common theme in all of the trheads I have started on the topic on various sites. And to be honest, while I am not giving up on the project, I am considering the advice you & so many others have given: basing the distro upon Debian instead of Ubuntu, but with the user friendliness Ubuntu is so well known for.

If I can give something that resembles the Ubuntu experience users love, but in a rolling release, I will accomplish my truest goal in all of this. If I do base it on Ubuntu, I will still want full comnpatibility (at least as close as possible) with Ubuntu repos & PPA's.

k3lt01 02-14-2013 03:11 AM

As already pointed out you wont really be able to mix repos without alot of work, that very well could be innefective anyway, to create a rolling release based on Ubuntu. I wouldn't even base it on Debian Testing (LMDE from linux Mint is suppsoed to be based on Debian Testing) because the "rolling" idea stops at freeze time which lasts for a fair while (6 months or so). Even Debian Sid cannot be considered a Rolling Release as it also slows down when Testing is in Freeze.

Getting back to the repository for a minute. If you mix and match repositories (official Ubuntu, your own Ubuntu, PPAs) you will create conflicts and I wonder how many people you will have working with you that will be able to fix these conflicts. Also many PPAs come and go so will you allow all the PPAs or just a select few.

If I were you I'd be giving up on either the Ubuntu based rolling release idea of the mixing and matching so many repositories idea. I'd also look for a different distro anyway. If you like the "user friendliness" of Ubuntu then use Debian Sid and modify it to suit because it will take less effort to do that than it would to use Ubuntu as you seem to be suggesting. If you are not worried about various tools like package management maybe you should research Fedora, or Siduction, or others listed on Distrowatch or LQ.

I'm not trying to turn you off the idea but I would hate to see you put in a huge amount of effort only to find your initial choices maybe weren't the best ones.

DavidMcCann 02-14-2013 12:04 PM

Once we get out of Debian territory, there's already a very user-friendly rolling-release distro: Fuduntu.

Other rolling-release distros, like Arch and Debian Sid, tend to be bleeding-edge, which is definitely not user-friendly, so a bit of work would be necessary to slow things down: you couldn't just use their repositories as they stand.

k3lt01 02-14-2013 12:27 PM

Quote:

Originally Posted by DavidMcCann (Post 4891668)
Other rolling-release distros, like Arch and Debian Sid, tend to be bleeding-edge, which is definitely not user-friendly, so a bit of work would be necessary to slow things down: you couldn't just use their repositories as they stand.

Sid isn't bleeding edge, Gnome is 3.4 in Wheezy and 3.4 in Sid, hardly a great leap forward in anyones books. Sid slows down alot when Testing is in freeze.

273 02-14-2013 01:03 PM

Well, running Sid for a while now I've not noticed any major breakages other than the change over to multilib, and that was solved by pinning a testing repository, so I'd guess that means a modified Sid with some testing packages to prevent breakages could work?
If you've enough people/time I dare say things like jockey could be pulled from Ubuntu to your private repository and kept at the version which matches Sid (either from LTS Ubuntu or current depending)?
I'll admit these are only suppositions based upon experience of running Debian based distros so are probably a little off the mark.

linuxPCplus 02-15-2013 09:43 PM

While I do agree it will be alot of work, I have to say that this would not be the first distro to successfully mix repos. Look at Manjaro,Bodhi, & even the early versions of Chakra. All mix(ed) repos belonging to their parent distro with their own repos. These are not the only ones doing it either.

All that being said, I do take your thoughts seriously & yes, I am looking at ther "parent distro" options. Let me try to be more clear about what I am trying to accomplish here, maybe this will allow folks to give even more advice & thoughts.

Reality is that I do not truly care what the parent distro is. Hell, for all I care it could be it's very own independent distro. But the things that absolutely MUST be true are these: First, the distro MUST be as user friendly (or more so) as Ubuntu. Second, the distro must be at least as stable (by default) as Ubuntu (but users should be able to easily enable less stable "testing" repos if they wish). Next, package management MUST be as easy or easier than Ubuntu. The distro MUST be rolling. One of the primary goals is to create a distro users can install, and keep up to date without EVER needing to manual upgrade as they must with standar releases such as Ubuntu.

There are other features I would LIKE to have, but these are the absolute musts (at least that I have determined so far)! A nice software center that works & looks as well as say Ubuntu Software center or Deepin Software center would be the ideal. Highly configurable auto-upgrade options would be important. For example, lets say the repos are updated daily (hypothetically). A user should be able to decide if they want to manually upgrade when they are ready or auto-upgrade daily, weekly, monthly, bi-monthly, so on.

Free choice is another essential. Users should be able to choose ANY DE they want. What would be cool is to have sort of a build it yourself setup for the initial distro download that lets the user create an ISO (in an easy GUI way) with the packages they want pre-installed. Of course, they could always choose from a variety of pre-built iso's too.

As I have pointed out in my earlier posts, this is the earliest stages here. It is literally just an idea in my head that I wanted feedback on. Do not misunderstand me, I am committed to this project! I have given alot of thought to this (& still do) &I fully intend to make this a reality. The dialog I am getting in these forums is helping me to determine more specifcally what I hope to create here. So thanks all for the feedback! Keep it coming!

Fuduntu IS user-friendly, but not Ubuntu friendly. I feel it is missing some key features that I want for this distro. Have any of you ever used Zorin? It is an Ubuntu based distro designed to make the transition from Windows to Linux "seamless". THAT is what I want for this distro, but in a rolling relase format. I want a Windows or Mac user to be able to pick this distro up & be able to use it almost immediately. It needs to be THAT friendly.

Randicus Draco Albus 02-15-2013 10:28 PM

Quote:

Originally Posted by linuxPCplus (Post 4892724)
I want a Windows or Mac user to be able to pick this distro up & be able to use it almost immediately. It needs to be THAT friendly.

The default installation of most distros can be used by by Windows and Mac users without difficulty. It is when people want to modify something it becomes necessary to become familiar with the system and learn a little. Essentially, what you are suggesting is a rolling release where the developers do all the work, so the users do not need to do any. That is fine, as long as you have enough people who can do all that work.


I also cannot help getting the impression that you want to create a one size fits all distro.
Quote:

What would be cool is to have sort of a build it yourself setup for the initial distro download that lets the user create an ISO (in an easy GUI way) with the packages they want pre-installed. Of course, they could always choose from a variety of pre-built iso's too.
User-friendly enough to use without learning anything -point and click in pretty GUIs; and with advanced options for customisation. Of vital importance to remember is that Windows and Mac achieve "friendliness" by taking away choice. As I stated above, you would try to create a system where the developers would have a huge amount of work.

Personally, I think such a project is over-ambitious, but if you have enough determination and people to help you, go for it and good luck.:)

k3lt01 02-15-2013 11:01 PM

Quote:

Originally Posted by Randicus Draco Albus (Post 4892743)
as long as you have enough people who can do all that work.

This, to me, is the clincher. It is a huge task and much more than 1 person can achieve.
Quote:

Originally Posted by Randicus Draco Albus (Post 4892743)
Personally, I think such a project is over-ambitious, but if you have enough determination and people to help you, go for it and good luck.:)

I agree to an extent. I had similar thoughts with Cobber but I realise it can't happen when it's just me doing Cobber so I changed my thinking to creating a stable release along with various levels of testing/unstable all using Debians native repositories with a small repository that is native to Cobber. To me anything more is to much work for one person.

Randicus Draco Albus 02-16-2013 12:40 AM

Quote:

Originally Posted by Randicus Draco Albus (Post 4892743)
Personally, I think such a project is over-ambitious,

I probably should clarify what I mean by over-ambitious. I do not mean a project that is too big, but rather, a project that encompasses too much. That is why I made reference to MS and Mac in connection with user friendliness. So-called "user friendliness" in terms of the ease of using Mac and Windows is achieved by limiting freedom. Creating a system that has the perceived ease of use of MS and Mac without locking the system down and allowing user freedom would be a daunting task. Others might disagree, but in my opinion, it would be trying to merge conflicting goals. If it is possible, it would require a lot of work.

TobiSGD 02-16-2013 04:57 AM

Quote:

Originally Posted by linuxPCplus (Post 4892724)
While I do agree it will be alot of work, I have to say that this would not be the first distro to successfully mix repos. Look at Manjaro,Bodhi, & even the early versions of Chakra. All mix(ed) repos belonging to their parent distro with their own repos. These are not the only ones doing it either.

Mixing repos is not a bad thing per se. The problem starts when you try to mix distros that are bleeding edge with stable repos. This will inevitable lead to problems with different versions.

Quote:

Highly configurable auto-upgrade options would be important. For example, lets say the repos are updated daily (hypothetically). A user should be able to decide if they want to manually upgrade when they are ready or auto-upgrade daily, weekly, monthly, bi-monthly, so on.
In a rolling release distribution auto-upgrades are something that should be avoided at all costs. Have a look at the most successful rolling release distro, Arch: It is highly recommended to check their website everytime before doing an upgrade so that you can work around caveats that may be existent. The software world moves fast and you won't be able to intercept all possible problems, especially if one of your goals is to have a highly configurable system.

Quote:

I want a Windows or Mac user to be able to pick this distro up & be able to use it almost immediately. It needs to be THAT friendly.
My personal opinion: For those users, usually with low knowledge level and the absence of motivation to learn or doing anything on the command line, a true rolling release is very difficult to bring to a level that will be accepted.
That is why distros like LMDE are not using a true rolling release scheme.

snowday 02-16-2013 10:44 AM

Windows is not rolling-release (look how many are still using XP!) and Mac OS in not rolling-release. Windows and Mac users are much more comfortable with a nice stable system that doesn't constantly change under their feet; this is why Ubuntu, Mint, Debian, Fedora, Slackware, Red Hat, etc. are the most popular distros.

DavidMcCann 02-16-2013 10:50 AM

The question of work also arises with the goal of offering all the desktop environments. I've used 7 desktops and 12 window managers, and there are still ones I haven't encountered yet!

Even if you keep to the most popular, the extras are often inferior to the flagship one. For OpenSUSE, KDE is better than Gnome which is better than Xfce. Why? Because they have more KDE users, since it's the only one on the live disk, and few Xfce testers because it's not in the beta-test releases. Similarly, Mint is not so good when it comes to KDE or Xfce. Supporting everything needs a lot of work, and it's also difficult to get it tested.

Over the years I've come to the conclusion that you need to have a very clear vision of what a distro is supposed to do and whom you expect to use it. That's one reason why Slackers are happy: Slackware is very focused on their needs and preferences, even if it's caviare to the general.

linuxPCplus 02-17-2013 05:04 PM

So I have actually done some more research & thinking. It seems to me that making a Ubuntu based rolling release should (in theory) not be as hard as most seem to think.
I mean Ubuntu has testing repos. You can download the next version any time you want. I have already been playing with 13.04 for example.

I may be totally wrong in the assumptions/theories I am about to propose, I will admit that up front. But lets think about this. Keep in mind, its all theoretical.

So, what if you created a distro that can access the exact same repos that are used for the Ubuntu version in development? Of course, you would want to make it easy for a user to decide how often the system updates. In other words, daily, weekly, monthly, etc.

You would also need to find a way to make the updates automatic. or at least partly automatic. One of the points of a rolling release is to eliminate the need to reinstall every time a new version comes out. So you would have to find a way to get the system to update from the Ubuntu testing repos without having to reinstall.

I do not intend to imply that this would be an easy task. Alot of challenges would face us in developing this distro. For one thing, we may need to have direct help from Canonical (to get the needed level of access to the repos). Extra steps would need to be implemented to increase stability of the distro since the Ubuntu testing repos are obviously not totally stable by nature.

Perhaps the distro could have it's own repos that pull from both the testing & stable Ubuntu repos, but instead of including EVERYTHING in the testing repo, we (well, the development team) include only those items that are tested & known to be relatively stable. In addition, the maintainers could add additional packages that are stable.

For those more brave users, they could enable another set of repos that include the full Ubuntu testing repos packages.
(BTW, sorry if I am confusing anyone here, I am thinking out loud. I am hoping that sharing my thoughts will help create a dialog that may some day lead to this idea becoming a reality).

So essentially, this distro would have it's own repos which are built from the Ubuntu repos, rather than actually using the Ubuntu repos directly. It would have a "Stable" repo built from both the main Ubuntu repos & APPROVED packages within Ubuntu's testing repos. The second repos would be the "Testing" repo, built from the main Ubuntu repos and a wider range of "testing" packages. Finally, there would be a third repo modeled around the Arch User Repos concept. This repo would be made up of packages that may or may not be approved by this distro's team, but are contributed by the distro's users.

So I have alot more going through my crazy head about thiss, but I think I have given some good food for thought & discussion. Like I said, this these ideas are primarily theoretical,and as such, may not be workable. But I wanted to share it & see where it leads!

TobiSGD 02-17-2013 05:19 PM

Quote:

Originally Posted by linuxPCplus (Post 4893962)
It would have a "Stable" repo built from both the main Ubuntu repos & APPROVED packages within Ubuntu's testing repos. The second repos would be the "Testing" repo, built from the main Ubuntu repos and a wider range of "testing" packages.

And at this point you would have a semi rolling release, like LMDE, not a true rolling release. I still think you shouldn't limit your base to an existing distro with a stable release cycle, and I still think that you don't get the implications of what you propose. True rolling release distros are not compatible to distros with a stable release cycle, alone due to software versions.

Make your distro an independent distro, you can still port Ubuntu's (or whichever distro's) tools to it, but you will get rid of the problems that inevitably will occur when you mix older packages into a rolling release distro.

linuxPCplus 02-17-2013 05:39 PM

Quote:

Originally Posted by TobiSGD (Post 4893980)
And at this point you would have a semi rolling release, like LMDE, not a true rolling release. I still think you shouldn't limit your base to an existing distro with a stable release cycle, and I still think that you don't get the implications of what you propose. True rolling release distros are not compatible to distros with a stable release cycle, alone due to software versions.

Make your distro an independent distro, you can still port Ubuntu's (or whichever distro's) tools to it, but you will get rid of the problems that inevitably will occur when you mix older packages into a rolling release distro.

Agreed, it would be a semi-rolling release. And I did in fact state that I do not know all of the implications, a fact I admitted from the start. That is the point of this discussion.

I am one of those who is on the fence about LMDE's status as a rolling release. I mean the current LMDE iso comes with GIMP 2.6 and 2.8 has been out for more than a year!

I admit I am leaning towards an independent distro, with Ubuntu's best tools/features ported over. The end goal is a rolling release that is as user friendly as Ubuntu, but more up to date. And perhaps, this is the best way to accomplish this.
(Do you relaize that kdenlive is on 0.9.4, but Ubuntu repos still have 0.8.2? That release is over 2 years old & 0.9.2 is the most recent "official" release!)

Randicus Draco Albus 02-17-2013 06:01 PM

Quote:

One of the points of a rolling release is to eliminate the need to reinstall every time a new version comes out. So you would have to find a way to get the system to update from the Ubuntu testing repos without having to reinstall.
Why is it necessary to re-install? Back in the days of yore when I used Ubuntu, I upgraded with the update manager. Granted, and this may be applicable to this theoretical system, I did not have any packages from outside the main repositories.;)


Quote:

we may need to have direct help from Canonical
the Ubuntu testing repos are obviously not totally stable by nature.
have it's own repos that pull from both the testing & stable Ubuntu repos
Why must Ubuntu's repos be used? The stated goal is a system that is as "easy" to use as Ubuntu. How many "user friendly" or "beginner" distros are tweaked versions of Debian? There is also a beginner distro that is a tweaked version of Slackware. Remember that almost all of Ubuntu's packages are slightly modified Debian packages. Which is probably why others who have posted in this thread recommend using Debian as the base. Personally, Ubuntu is that last system I would use as the base of another. I am reminded of the proverb; garbage in, garbage out.

It seems to me that the real objective is a distro that looks and feels like Buntu, but is better. Is that not what Mint is supposed to be?

Quote:

It would have a "Stable" repo built from both the main Ubuntu repos & APPROVED packages within Ubuntu's testing repos. The second repos would be the "Testing" repo, built from the main Ubuntu repos and a wider range of "testing" packages. Finally, there would be a third repo modeled around the Arch User Repos concept. This repo would be made up of packages that may or may not be approved by this distro's team, but are contributed by the distro's users.
And how would that make the system as easy to use as Mac and Windows? That would require users to read the system's documentation and learn how to administrate their systems, which is contrary to to the Microsoft/Apple/Ubuntu philosophy of keeping the users ignorant. As I eluded to in an earlier post, you want to make the system both complex and easy. If you want to create a system as easy to use as you described earlier, it would offer packages from one repository. Offering stable, unstable and experimental packages, and leaving it to the users to decide how to mix and match them, requires a level of knowledge well beyond point and click. It is like wanting an engine that has 500 horse power and gets 50 miles per gallon. The best advice I can give is to focus on what is most important: ease of use or a remarkably complex system. You need to decide which demographic of users the system would cater to.

k3lt01 02-17-2013 08:51 PM

Quote:

Originally Posted by linuxPCplus (Post 4893962)
So I have actually done some more research & thinking. It seems to me that making a Ubuntu based rolling release should (in theory) not be as hard as most seem to think.
I mean Ubuntu has testing repos. You can download the next version any time you want. I have already been playing with 13.04 for example.

At the end of every development cycle everything stops totally for a period of a couple of weeks. I was an Ubuntu Development tester for a couple of years, actually there are a few of us in LQ that were Ubuntu Dev testers, and can tell you with certainty you do not want to base your rolling release on Ubuntu's development release. Debian Sid (which most Ubuntu development cycles are based on) is infinately more stable than, even stable, Ubuntu releases. As an example, there is a reason why Squeeze and Ubuntu 10.04 have the same specs and that is because 10.04 is based on Squeeze yet it was released 12 months ahead of Squeeze.

linuxPCplus 02-18-2013 01:25 AM

Quote:

Originally Posted by Randicus Draco Albus (Post 4893994)
Why is it necessary to re-install? Back in the days of yore when I used Ubuntu, I upgraded with the update manager. Granted, and this may be applicable to this theoretical system, I did not have any packages from outside the main repositories.;)


Why must Ubuntu's repos be used? The stated goal is a system that is as "easy" to use as Ubuntu. How many "user friendly" or "beginner" distros are tweaked versions of Debian? There is also a beginner distro that is a tweaked version of Slackware. Remember that almost all of Ubuntu's packages are slightly modified Debian packages. Which is probably why others who have posted in this thread recommend using Debian as the base. Personally, Ubuntu is that last system I would use as the base of another. I am reminded of the proverb; garbage in, garbage out.

It seems to me that the real objective is a distro that looks and feels like Buntu, but is better. Is that not what Mint is supposed to be?


And how would that make the system as easy to use as Mac and Windows? That would require users to read the system's documentation and learn how to administrate their systems, which is contrary to to the Microsoft/Apple/Ubuntu philosophy of keeping the users ignorant. As I eluded to in an earlier post, you want to make the system both complex and easy. If you want to create a system as easy to use as you described earlier, it would offer packages from one repository. Offering stable, unstable and experimental packages, and leaving it to the users to decide how to mix and match them, requires a level of knowledge well beyond point and click. It is like wanting an engine that has 500 horse power and gets 50 miles per gallon. The best advice I can give is to focus on what is most important: ease of use or a remarkably complex system. You need to decide which demographic of users the system would cater to.

Quite simply, because reinstalling is MUCH faster. Running a ditribution upgrade from update manager takes (in my experiences) 2 hours or more, and often causes alot of breakages or other problems. Reinstalling takes about 30-45 minutes.

The next comments have been answered in previous posts where I mentioned the distro having it's own repos.

The repos would not make it as easy as Mac or Windows to use. That would be accomplished through other GUI tools. For example, I already have a look changer simialr to the one found in Zorin (actually based on it) to make the system look like Windows or Mac. There would be several other tools that would make the distro look & feel enough like Windows or Mac to make it comfortable or familiar to Windows/Mac users.

"Offering stable, unstable and experimental packages, and leaving it to the users to decide how to mix and match them, requires a level of knowledge well beyond point and click."
Other distros have managed to do exactly this quite successfully: Bodhi, Manjaro, & Vector for example. The big difference here would be that where as in these distro's you have to manually change the source lists, in this new distro it would be done via a nice little gui.

Try to think of it as a scalable system. When a new user picks it up with little or no Linux experience, they get a fast, stable, & up to date system. As they become more familiar with Linux, they can "tweak" it to get more advanced features, package updates, & bleeding edge software.

I will repeat once again since some don't seem to be getting it: this is an idea, a concept in my overactive imagination. The more this discussion progresses, the more I am able to "tweak" or "sharpen" the image of what I hope to create. The open source community is one of the most creative communities in the world! Opensource developers are well know for doing things that "can't be done" or that are "too complicated". I hope I am carrying on that tradition!

Randicus Draco Albus 02-18-2013 02:26 AM

Quote:

The big difference here would be that where as in these distro's you have to manually change the source lists, in this new distro it would be done via a nice little gui.
Like Synaptic?

Quote:

Try to think of it as a scalable system. When a new user picks it up with little or no Linux experience, they get a fast, stable, & up to date system. As they become more familiar with Linux, they can "tweak" it to get more advanced features, package updates, & bleeding edge software.
Instead of a singe, scalable system, perhaps Debian's division of Stable, Testing, Unstable is a model to consider?

Quote:

Running a ditribution upgrade from update manager takes (in my experiences) 2 hours or more
You are blessed with a quick connection. My upgrades took several hours.:( Upgrade time is the price paid for stability.


Quote:

There would be several other tools that would make the distro look & feel enough like Windows or Mac to make it comfortable or familiar to Windows/Mac users.
There are two schools of thought concerning that issue. Some believe a Linux system that looks just like Windows will be appealing, because it offers a familiar, and comfortable, environment. Others believe more people will be attracted by something different. Why switch to something that is the same as I already have? No right or wrong answer there.


Quote:

this is an idea, a concept in my overactive imagination. The more this discussion progresses, the more I am able to "tweak" or "sharpen" the image of what I hope to create.
Then it is necessary to consider all ideas and criticisms offered. Both good and bad. You can be almost guaranteed that as ideas and information unfold, your vision will metamorphose. Possibly into something you did not expect. Such is the journey of creativity.


Edit
Another idea to consider that would require more work, but would have a better chance of bearing fruit, would be to have two separate releases. A stable beginner's system that only has stable older packages, upgraded at regular intervals; and a rolling release with cutting edge packages for advanced users. Upgraded packages would move from the rolling release to the stable release after the bugs have been worked out.

A stable rolling release is improbable, but a dual-release distro would offer both. The distro's web site should have copious documentation and/or links to it, encouraging users of the stable release to read documentation. The immediate goal being to help new users move up to the rolling release, but also with the long-term objective of helping them become knowledgeable Linux users.

TobiSGD 02-18-2013 05:31 AM

Quote:

Originally Posted by linuxPCplus (Post 4894175)
"Offering stable, unstable and experimental packages, and leaving it to the users to decide how to mix and match them, requires a level of knowledge well beyond point and click."
Other distros have managed to do exactly this quite successfully: Bodhi, Manjaro, & Vector for example. The big difference here would be that where as in these distro's you have to manually change the source lists, in this new distro it would be done via a nice little gui.

No, other distros do it not that way. All the distros you mentioned build their packages on the same stable base. If you have a stable and a rolling release branch of your distro (or, as you mentioned earlier, for example Ubuntu's current release and the future version) you have two different base systems, with different versions of software. Mixing those two repos will break the system and using a GUI instead of editing a text file will not prevent that.

Quote:

I will repeat once again since some don't seem to be getting it: this is an idea, a concept in my overactive imagination. The more this discussion progresses, the more I am able to "tweak" or "sharpen" the image of what I hope to create. The open source community is one of the most creative communities in the world! Opensource developers are well know for doing things that "can't be done" or that are "too complicated". I hope I am carrying on that tradition!
I don't want to discourage you and this may be sound hard, but I think that your project in the current state is doomed. While you can imagine how your distro should be, you seem to lack the knowledge to filter out which parts of your imagination are viable and which not, especially with limited manpower. Open source developers may be able to do great things, but the problem here is that you are not a developer, you are a manager. You can imagine things, but if they are viable is beyond your current knowledge.
Before going further I would recommend to get more knowledge about this topic. Try to go through LFS one time, so that you can see what actually happens under the hood of a Linux system. Go ahead and mix some packages of the Ubuntu testing branch into the current stable branch and see what happens and where the difficulties with that really are. Then go through your vision again and test it against your new knowledge. Become at least a little bit of a developer, planning a distro is a job for developers, not managers.

linuxPCplus 02-18-2013 08:19 AM

Quote:

Originally Posted by TobiSGD (Post 4894281)
No, other distros do it not that way. All the distros you mentioned build their packages on the same stable base. If you have a stable and a rolling release branch of your distro (or, as you mentioned earlier, for example Ubuntu's current release and the future version) you have two different base systems, with different versions of software. Mixing those two repos will break the system and using a GUI instead of editing a text file will not prevent that.

I don't want to discourage you and this may be sound hard, but I think that your project in the current state is doomed. While you can imagine how your distro should be, you seem to lack the knowledge to filter out which parts of your imagination are viable and which not, especially with limited manpower. Open source developers may be able to do great things, but the problem here is that you are not a developer, you are a manager. You can imagine things, but if they are viable is beyond your current knowledge.
Before going further I would recommend to get more knowledge about this topic. Try to go through LFS one time, so that you can see what actually happens under the hood of a Linux system. Go ahead and mix some packages of the Ubuntu testing branch into the current stable branch and see what happens and where the difficulties with that really are. Then go through your vision again and test it against your new knowledge. Become at least a little bit of a developer, planning a distro is a job for developers, not managers.

The truth is, I agree. I can admit that you are right. I am learning as this discussion unfolds that while this is a nice dream, making it reality will be a major challenge. Do I still think that some version of this dream is possible? Hell yes! But unless I learn more about the system, this project cannot succeed.

Just to let ya know, I am not "all talk". I am actually returning to school on March 18 to finish my Netwrok System Administration degree and I still studying for the LPIC level 1 exams. So yes, I AM educating myself & making the effort to "expand my mind" so I can be less a manager & more a contributor.

Some of you may have looked at my website & noticed that I do offer Linux tech support services. I just want to point out that I do not personally provide these services, I have employees who do all of the support & I require my support agents to be either CompTia or LPIC certified. I only point this out because I realize this discussion may have created some concerns about my tech services being qualified.

I am shelving this idea for te time being. Once I have gained the appropriate knowledge & skills, I will revisit it. Thank you to EVERYBODY who gave me constructive criticism & advice. While I did not like all that I heard, I recognize it was never a personal attack, but just frank honesty which I truly appreciate!

unSpawn 02-18-2013 08:30 AM

IMHO there are more fundamental flaws which go for a lot of "distributions" out there. One is the idea that taking an original distribution and "making it better" is somehow a Good Idea. Related to that is the idea that it is OK for this new "distribution" to defy Laws of Nature and to do so without consequence, somehow showing progress without gradual, natural evolution. IMHO that is a move away from core values associated with traditional craftsmanship and a dismissal of the fact that maturation takes time and effort. Over-cultivation, prostitution of any resources and lack of original creativity will reveal themselves in time always. Phrased differently: if one asserts Linux is a healthy ecology then it should show signs of true biodiversity. For a reason. You shelving the idea shows that. Well done.

linuxPCplus 02-18-2013 03:15 PM

Quote:

Originally Posted by unSpawn (Post 4894377)
IMHO there are more fundamental flaws which go for a lot of "distributions" out there. One is the idea that taking an original distribution and "making it better" is somehow a Good Idea. Related to that is the idea that it is OK for this new "distribution" to defy Laws of Nature and to do so without consequence, somehow showing progress without gradual, natural evolution. IMHO that is a move away from core values associated with traditional craftsmanship and a dismissal of the fact that maturation takes time and effort. Over-cultivation, prostitution of any resources and lack of original creativity will reveal themselves in time always. Phrased differently: if one asserts Linux is a healthy ecology then it should show signs of true biodiversity. For a reason. You shelving the idea shows that. Well done.

LMAO! Thanks (I think)! IF I ever create a Linux distro (or at least help one be created) I want it to be successful. I do not want folks to feel it is just aother distro to add to the bottom of the already long & growing list. It will need to be a distro that appeals, that makes a difference, & give tru measurable benefits to its users.

I think completeing my education (both formal & informal) will be the key to making this vision a reality!


All times are GMT -5. The time now is 09:51 PM.