Make Debian as the base.

Provide us your suggetions and New Ideas to improve Hama GNU/Linux
shirish
Posts: 1
Joined: Tue Feb 24, 2015 6:49 pm
Contact:

Make Debian as the base.

Postby shirish » Mon Mar 16, 2015 7:25 pm

Hi all,
Can anybody confirm who or what is the parent distribution of Hamara. From the repo. given it seems it comes from the Ubuntu stable.

http://archive.hamaralinux.org/hamara/dists/namaste/

I made it out because of the universe and multiverse repos which is how Ubuntu names it repos.

I want to know and ask why Ubuntu was taken as a parent. Also there doesn't seem to be any clarity as to who the targetted users are. Is it a general distribution or are there any aspects/groups which you would like to stress on more?

I went through the skimpy hamaralinux.org website but couldn't see any reasons why Ubuntu was chosen.

I come from the Debian user-land and I see many benefits that hamara could have if they chose to have their packages and development strategies vis-a-vis Ubuntu. Let me enumerate them as under :-

a. Instead of a 6 monthly cycle we have around a 2/2.5 year cycle. So in terms of either user-support or support in general, it would be much more easier to handle support for your users than constantly running around trying to get the new release baked.

b. Debian uses Experimental > Unstable > Testing > Stable repos :- A new package (just debianized app.) hits usually unstable/sid where it sits/may sit for a long time before getting into testing. Testing is what usually becomes the next release/version so there is predictability in what is coming down the chute. So people who like certain softwares can use/re-use it in their day-to-day usage, report and triage bugs and make the transition to testing smoother.

c. Baking time is longer - In Debian a new release usually hits unstable (and sometimes experimental, especially during freeze which is where we are now ) and then depending on issues raised or not would make it to testing/new version after 10 days. Most of the developers are fast so you could have packages with quicker turn-around too while at the same time if there are any serious bugs it would not move and give users ample warning before installation of a package with severe bug/s. The user can then go to those bugs and choose whether to install or not the said package.

d. Mixed repos. :- This is the biggest grouse or issue I had with Ubuntu. If a package had a defect, they will not fix the package for the stable as once it's in stable they want to fix it for the new version. For a stable package Update (SPU) to happen lot of things need to happen :-

1. The package is a high-priority package (meaning it has paying customers) which means it is much more likely that if you find a bug in Apache or Nginx it would be fixed, but if you find a bug in say VLC it probably would be fixed for the new release, tough luck for the current one.

2. The package bug is not a corner-case bug but more of a mainstream one which is easily reproducible.

So for a user if he wants to check a new release of a package, he has to abandon the full stack and come into a new uncertain stack (alpha releases) where packages might/would be broken.

In Debian you can get away with mixed repos. (which I have been doing for years) where you have a new package (for e.g. I am using Iceweasel 36-.0-1 which was released just three days but using the tried and tested glibc, kernel and other parts of the stack which I know and have been using for about couple of years.

e. More number of packages - The upcoming release would have around 40k of packages in Debian main and the number will always increase. There are 1k of active Debian Developers (DD's), bigger community, much bigger infrastructure support then probably what Canonical has.

f. Transparency :- While we do have non-free repo. the idea of even the non-free repo. is to nudge people to open up the software and get it into main. The motto is also to make all bugs public. Also most CVE's and zero day bugs are tackled within the first 24-48 hours. You can look at most of the recent issues (for e.g. Heartbleed SSL or any of the others) and check the response time that Debian had in pushing new packages as well as helping in diagnozing and making test-cases for the problem as well as the solution.

There are many more arguments but hopefully will have convinced you on the technical merits of the distribution. There are other worthy contendors such as Arch, Gentoo, Slackware etc.

I do hope if somebody has an answer for this and whoever answers this puts it in the FAQ as well.
flossexperiences.wordpress.com

vik
Posts: 32
Joined: Thu Aug 14, 2014 3:55 pm
Contact:

Re: Make Debian as the base.

Postby vik » Thu Mar 19, 2015 10:54 am

Can anybody confirm who or what is the parent distribution of Hamara. From the repo. given it seems it comes from the Ubuntu stable.

http://archive.hamaralinux.org/hamara/dists/namaste/
Yes, we build hamara on top of Ubuntu 14.04.1
I want to know and ask why Ubuntu was taken as a parent.
We followed the distro building process initially established by gnewsense and furthered by Trisquel. It was more straightforward to build on top of Ubuntu, but we do have plans to move to a Debian base in our next release.

Also there doesn't seem to be any clarity as to who the targetted users are. Is it a general distribution or are there any aspects/groups which you would like to stress on more?
Indian users are our focus and in particular we want to target:

- Young users via helping Indian Schools implement linux
- Indian Developers by ensuring good support for developer toolsets
- Non English speakers by ensuring good support for the main 5 Indian languages first
- Indian users who do not yet have access - by supporting low cost arm devices and encouraging a DIY approach.
Instead of a 6 monthly cycle we have around a 2/2.5 year cycle. So in terms of either user-support or support in general, it would be much more easier to handle support for your users than constantly running around trying to get the new release baked.
I agree - this would be a better approach for hamara - we are not keen on following the Ubuntu way of 6 monthly releases.

d. Mixed repos. :- This is the biggest grouse or issue I had with Ubuntu. If a package had a defect, they will not fix the package for the stable as once it's in stable they want to fix it for the new version.
We have our main repo and the devel repo. I think mixing packages from these is fine for experienced users, but for the main repo - we are happy to bug fix packages in devel and move them to main once tested.

We may well need a third repo now for experimental :)
e. More number of packages - The upcoming release would have around 40k of packages in Debian main and the number will always increase. There are 1k of active Debian Developers (DD's), bigger community, much bigger infrastructure support then probably what Canonical has.
Well, moving towards Debian was always the intention!
I do hope if somebody has an answer for this and whoever answers this puts it in the FAQ as well.
Will do and thanks for the post, very helpful!
Vikas Tara
Founder - Hamara Linux


Return to “Suggestions & New Ideas”

Who is online

Users browsing this forum: No registered users and 1 guest

cron