[Hamara-devel] Sugam on Debian and the about page.

Wed Mar 25 11:45:07 GMT 2015

On 2015-03-25 02:38, Gurvinder Dadyala wrote:
> On Monday 23 March 2015 09:30 PM, shirish शिरीष wrote:
>> Hi all,
>> 
>> There are and were lots of thoughts after my discussion with Vikas on
>> Friday evening, but as week-end happened and had to meet with friends
>> and socialize a bit. As the about page would and should be a
>> collaborative effort, some numbers, some discussion before making
>> writing them are in order but even before building sugam using debian
>> note which Vikas shared and my thoughts why I said it's better to have
>> or use Debian as the base even now than Ubuntu/Canonical.
>> 
>> a. It doesn't make any sense other than to say that we have done sugam
>> using Ubuntu (kind of like bragging right) but as far as support or
>> anything is concerned we wouldn't be able to give/share any of that.
>> Even how much testing we would be able to do ourselves is not known at
>> this point of time. If it's on Ubuntu I would probably not be so
>> inclined to volunteer my time on that, on Debian it would directly or
>> indirectly serve Debian as well because if any bugs we uncover (if and
>> when) it would help to make Debian better as well as upstream. On
>> Canonical/Ubuntu I haven't seen them collaborate on the same level, so
>> in a sense the commons are served. This serves both ethical and
>> commercial interests as resources are being used where there is no
>> danger of being withered away.
> Ubuntu community is not that open as it should be there in Open Source
> software contribution. Most of the time you will find issues which are
> solved on Debian and same issue is still pending in Ubuntu. Our choice
> of using Ubuntu for sugam was not for specific reasons as it was also
> for the first time we were trying to do something like this.
> May be because as a user we have been using Ubuntu for a long time, so
> we decided to just choose Ubuntu as base for Hamara. However because
> we are still at initial stages we can definitely go for Debian as base
> system because Debian ecosystem is much more closer to Open Source
> Software concept as compared to Ubuntu.
>> 
>> b. If you look at the differences between Ubuntu and Debian, that
>> would be around 20% and Unity is one of the major differences in
>> there, apart from that there is no much difference (except if you take
>> Canonical partner offerings but that is a different story altogether.)
> That is absolutely correct, the only noticeable difference one would
> find is Unity. Though I have not used debian much, however have spent
> some time when i was working for ARM build. Ubuntu as far as I have
> seen from last few years have started focusing on UI rather than bugs.
> You can yourself find same bugs in the next build which might never be
> fixed or are automatically closed because of no activity. At the end
> of day it also depends on end user, who always goes for word of mouth.
> For example you will find people knowing about Ubuntu but not about
> Debian.
>> 
>> c. The other reason which I think I shared or didn't share in my
>> earlier post is flexibility. In Debian I could use the BSD kernel or
>> GNU Hurd if the situation so requires (although when and if GNU Hurd
>> would become a reality is another story in itself) but this is
>> possible because of de-coupling or not having hard paths to a specific
>> kernel type. This is and would not be true of Canonical.
> I have not tried this but I guess that is best part that you can use
> some kernel rather than sticking to mainline repo as in case of
> Ubuntu. Because sometimes as a user you might want to try another
> kernel for some issue which you think is not yet fixed in Ubuntu
> mainline kernel. De-coupling and not having hard paths are actually
> proofs that as a user, you are not being forced to use only specific
> type of kernel and are free to make your choice as a user.
> I have not heard of GNU hurd, but will definitely check
>> 
>> For Debian see https://www.debian.org/ports/kfreebsd-gnu/ and
>> https://www.debian.org/ports/hurd/ . You can find quite a good lists
>> of the various ports and their progress at
>> https://www.debian.org/ports/ - look at both the official and
>> unofficial ports to have an understanding of the depth of both the
>> range of the efforts and people trying to do them.
> You are right Debian has more ports and support far more architectures
> than Ubuntu. Another reason for this may be because of not so open
> system provided by Ubuntu from contributor perspective.
>> 
>> So, the hard path would probably be tackling d-i (debian-installer)
>> but this is precisely where people could help. What would be needed is
>> the latest snapshot of debian-installer, the installation-report
>> package (already part of d-i) and putting whatever modifications made
>> to the UI. See https://www.debian.org/devel/debian-installer/ and see
>> the mailing list archives of debian-boot
>> https://lists.debian.org/debian-boot/2015/03/threads.html . There is
>> also #debian-boot on oftc for help and support.
>> 
>> Another tool people could look at would be boxer as have heard good
>> things about it but as haven't used it can't say one way or the other
>> (haven't needed to) https://wiki.debian.org/FreedomBox/Tools
>> 
>> d. Another interesting sub-project which was taken and is being taken
>> seriously by folks is using clang to uncover compiler errors alongwith
>> gcc. See http://clang.debian.net/ .
>> 
>> All of the above shows that people have real reasons for choosing and
>> using Debian as a base (for me it's all bugs in the open except for
>> CVE's) and getting timely support.
>> 
>> Apart from all above, if you use Ubuntu now and then want to switch to
>> Debian that will involve pain both in terms of tools, time and
>> knowledge which will be completely unnecessary with lower costs now
>> then in the future.
>> 
>> This concludes my comments about building sugam on Debian.
>> 
>> Now for the about page two re-curring themes were there with Vikas
>> while we were chatting/discussing on the phone.
>> 
>> a. Localization - This is where we have been bad, how bad we are at it
>> can be guaged at https://www.debian.org/international/l10n/po/hi (if
>> we just take hindi at the language to be handled)  I have shared some
>> brief on that at
>> https://flossexperiences.wordpress.com/2014/11/05/debutsav-2014-experiences-day-2/
>> but as shared have no real expertise on it. Having more people having
>> hands-on and adding to the .po files would help all and any free
>> software immeasurably while helping development of Sugam itself. This
>> is a huge problem area but if we are able to do move some tiny
>> percentages on it, the benefits are immense. As an e.g. you could take
>> Nokia's making money off cheap mobiles which used local languages as a
>> starting point, the point is, if the language of all different parts
>> of the desktop and softwares are in local language, the take-up would
>> be larger, English as a language is a barrier for many Indians and
>> hence technology. If technologies such as Raspberry Pi2 or beaglebone
>> black or any of the newer ARM-based devices could be delivered to
>> people in local language, there is and would be a quantum leap in
>> understanding, doing business, entertainment etc. which we can't think
>> of. Just think of a kid sitting somewhere and using blender in his/her
>> local language and making an animated movie on his Pi2, just blows the
>> mind. There are unlimited use-cases for localization and the
>> 'job'/work  is well-defined, just need people who know it, are
>> passionate about it and willing to devote time on it.
> Localization always helps in increasing user base and more people
> would love to provide support in terms on localization rather than
> just bug fixing and testing. Specially for ARM based devices
> localization can be the main driver in convincing people for using
> Hamara.
>> 
>> b. EDA - The global EDA industry is somewhere between 9 to 19 billion
>> dollars depending on whose numbers you believe in. India would
>> probably be somewhere in 1 or 2 percent as there is some EDA happening
>> but Modi's 'Make in India' could probably change that. Add to that
>> technologies which would influence the EDA industry i.e. IOT, Wearable
>> Computing, Mobile and Social, each of which are estimated to each
>> generate around 250 billion dollars world-wide by 2020, the
>> opportunities are unlimited. It's how much risk we are willing to
>> take. There is quite a bit of EDA tools in FOSS which go hand-in-hand
>> in the whole open-source hardware philosophy as well. There was an
>> initiative to package FOSS EDA tools number of years back (a friend of
>> mine Archis Gore and me and several others) , we did a release but
>> were unable to do much due to number of reasons (one of the top being
>> the students our main target group didn't have FOSS in syllabus) so
>> all the effort went into drain. This is/was about 10 years back. Tools
>> have grown a lot since then but would need somebody who is passionate
>> about the EDA industry and can share his/her wisdom with the rest of
>> us.
>> 
>> c. Sell systems with Hamara Linux - Don't need to say more. Actually
>> have shared quite of my own quandary at
>> https://flossexperiences.wordpress.com/2014/11/05/debutsav-2014-experiences-day-2/
>> 
>> Both the ideas as shared by Vikas are rich and have untold dividends.
>> I am excited to be part of hamaralinux as we get the show on the road
>> :)
> _______________________________________________
> Hamara-devel mailing list
> Hamara-devel at lists.hamaralinux.org
> http://lists.hamaralinux.org/listinfo/hamara-devel

It's good to see everyone detailing out benefits of basing hamara/sugam 
on Debina. The discussion in the thread gives enough reason for us to 
switch to debian as an underlying operating system. I am not going to 
say anything(technical) extra to advocate this move but detailing out 
few things which will be good from an user points of view and things we 
need to keep in mind while going forward this route.

Wins over Ubuntu

1.)Hardware Compatibility : While our focus is around providing hamara 
on all(high end/low end) hardware, this is an important thing which we 
will be getting if we move to debian. Debian currently develops for 13 
hardware architectures (a lot :) ), ranging from standard 32 and 64 bit 
Intel to Arm and PowerPC, with support for two more in progress. By 
contrast, Ubuntu supports 32 and 64 bit versions as computer desktop 
environments, and is developing Ubuntu ARM for mobile devices. Using 
debian is a clear win here.

2.) Stability : For an end user stability is something very important 
and even more important if they are using linux ;) (People are not 
worried about stability of MS but if you ask them to use linxu the first 
question will be about it's stability.) As far as stability is 
concerned, a stable release of Debian is more robust than the rock of 
Gibraltar. You can say goodbye to blue screens and random system 
crashes; Debian will almost never go down – which is why it is so 
popular for enterprise-grade applications and web hosting!

What we need to take care of-

1.) Installer : Debian is said to be a developer-first Linux 
distribution. Although it is robust, secure and powerful, it isn’t 
exactly designed for those new to Linux. The developer community has 
worked tirelessly in the last few years to make the installation process 
and basic setup easier, but it is still leaps and bounds away from 
Ubuntu’s usability as far as installer is concerned. We need to work on 
this part and see how we can improve the installer.

2.) Desktop/UI : This is really an important part and the general 
conception is that ubuntu has got an edge here because of their "Unity" 
thing. However using gnome for a long time my personal preference is 
gnome over unity. I am not sure about a new user though if they would 
like gnome/unity/kde/xfce/lxde/whatever. May be we can provide multiple 
desktop environment with simple switch option in the OS ??

3.) Software Availability : I have heard Ubunt has got thousand of 
softwares in their repository. Does debian lacks here ? Need to check 
this.

4.) Package Manager : We can't/shouldn't expect everyone to run "sudo 
apt-get bla bla bla". We need a good package manager for hamara. Not 
sure how good package manage debian has got ? Anyone wants to confirm ?

The another topic being discussed in the thread is about hamara website 
about-page I GUESS? Sorry guys but I couldn't really understand the 
points there :(

Thanks
Shantanu