It was a surprise to me to learn that project to create a complete
computer system for schools I've involved in,
Debian Edu / Skolelinux, was
being used in India. But apparently it is, and I managed to get an
interview with one of the friends of the project there, Shirish
Who are you, and how do you spend your days?
My name is Shirish Agarwal. Based out of the educational and
historical city of Pune, from the western state of Maharashtra, India.
My bread comes from giving training, giving policy tips,
installations on free software to mom and pop shops in different
fields from Desktop publishing to retail shops as well as work with
few software start-ups as well.
How did you get in contact with the Skolelinux / Debian Edu
It started innocently enough. I have been using Debian for a few
years and in one local minidebconf / debutsav I was asked if there was
anything for schools or education. I had worked / played with free
educational softwares such as Gcompris and Stellarium for my many
nieces and nephews so researched and found Debian Edu or Skolelinux as
it was known then. Since then I have started using the various
education meta-packages provided by the project.
What do you see as the advantages of Skolelinux / Debian
It's closest I have seen where a package full of educational
software are packed, which are free and open (both literally and
figuratively). Even if I take the simplest software which is
gcompris, the number of activities therein are amazing. Another one of
the softwares that I have liked for a long time is stellarium. Even
pysycache is cool except for couple of issues I encountered
I prefer software installed on the system over web based solutions,
as a web site can disappear any time but the software on disk has the
possibility of a larger life span. Of course with both it's more a
question if it has enough users who make it fun or sustainable or both
for the developer per-se.
What do you see as the disadvantages of Skolelinux / Debian
I do see that the Debian Edu team seems to be short-handed and I
think more efforts should be made to make it popular and ask and take
help from people and the larger community wherever possible.
I don't see any disadvantage to use Skolelinux apart from the fact
that most apps. are generic which is good or bad how you see it.
However, saying that I do acknowledge the fact that the canvas is
pretty big and there are lot of interesting ideas that could be done
but for reasons not known not done or if done I don't know about them.
Let me share some of the ideas (these are more upstream based but
still) I have had for a long time :
1. Classical maths question of two trains in opposing directions
each running @x kmph/mph at y distance, when they will meet and how
far would each travel and similar questions like these.
The computer is a fantastic system where questions like these can
be drawn, animated and the methodology and answers teased out in
interactive manner. While sites such as the
Dr. Math FAQ on The Two Trains problem (as an example or point of
inspiration) can be used there is lot more that can be done. I dunno
if there is a free software which does something like this. The idea
being a blend of objects + animation + interaction which does
this. The whole interaction could be gamified with points or sounds or
colourful celebration whenever the user gets even part of the question
or/and methodology right. That would help reinforce good behaviour.
This understanding could be used to share/showcase everything from how
the first wheel came to be, to evolution to how astronomy started,
psychics and everything in-between.
One specific idea in the train part was having the Linux mascot on
one train and the BSD or GNU mascot on the other train and they
meeting somewhere in-between. Characters from blender movies could
also be used.
2. Loads of crossword-puzzles with reference to subjects: We have
enormous data sets in Wikipedia and Wikitionary. I don't think it
should be a big job to design crossword puzzles. Using categories and
sub-categories it should be doable to have Q&A single word answers
from the existing data-sets. What would make it easy or hard could be
the length of the word + existence of many or few vowels depending on
the user's input.
3. Jigsaw puzzles - We already have a great software called
palapeli with number of slicers making it pretty interesting. What
needs to be done is to download large number of public domain and
copyleft images, tease and use IPTC tags to categorise them into
nature, history etc. and let it loose. This could turn to be really
huge collection of images. One source could be taken from
commons.wikimedia.org, others could be huge collection of royalty-free
stock photos. Potential is immense.
Apart from this, free software suffers in two directions, we lag
both in development (of using new features per-se) and maintenance a
lot. This is more so in educational software as these applications
need to be timely and the opportunity cost of missing deadlines is
immense. If we are able to solve issues of funding for development and
maintenance of such software I don't see any big difficulties. I know
of few start-ups in and around India who would love to develop and
maintain such software if funding issues could be solved.
Which free software do you use daily?
That would be huge list. Some of the softwares are obviously apt,
aptitude, debdelta, leafpad, the shell of course (zsh nowadays),
quassel for IRC. In games I use shisen-sho while card-games are evenly
between kpat and Aiselriot. In desktops it's a tie between
gnome-flashback and mate.
Which strategy do you believe is the right one to use to
get schools to use free software?
I think it should first start with using specific FOSS apps. in
whatever environment they are. If it's MS-Windows or Mac so be it.
Once they are habitual with the apps. and there is buy-in from the
school management then it could be installed anywhere. Most of the
people now understand the concept of a repository because of the
various online stores so it isn't hard to convince on that front.
What is harder is having enough people with technical skills and
passion to service them. If you get buy-in from one or two teachers
then ideas like above could also be asked to be done as a project as
I think where we fall short more than anything is in marketing. For
instance, Debian has this whole range of fonts in its archive but
there isn't even a page where all those different fonts in the La
Ipsum format could be tried out for newcomers.
One of the issues faced constantly in installations is with updates
and upgrades. People have this myth that each update and upgrade
means the user interface will / has to change. I have seen this
innumerable times. That perhaps is one of the reasons which browsers
like Iceweasel / Firefox change user interfaces so much, not because
it might be needed or be functional but because people believe that
changed user interfaces are better. This, can easily be pointed with
the user interfaces changed with almost every MS-Windows and Mac OS
The problems with Debian Edu for deployment are many. The biggest
is the huge gap between what is taught in schools and what Debian Edu
is aimed at.
Me and my friends did teach on week-ends in a government school for
around 2 years, and
some experience there. Some of the things we learnt/discovered
there was :
- Most of the teachers are very territorial about their subjects
and they do not want you to teach anything out of the
- They want any activity on the system in accordance to whatever
is in the syllabus.
- There are huge barriers both with the English language and at
times with objects or whatever. An example, let's say in gcompris
you have objects falling down and you have to name them and let's
say the falling object is a hat or a fedora hat, this would not be
as recognizable as say a
Pagdi so there is need to inject local objects, words wherever
possible. Especially for word-games there are so many hindi words
which have become part of english vocabulary (for instance in
parley), those could be made into a hinglish collection or
something but that is something for upstream to do.