On Mon, Jul 10, 2023 at 4:04 PM Jonas Smedegaard <
jonas@jones.dk> wrote:
Hi Yogesh,
Quoting Yogesh Powar (2023-07-10 09:50:36)
> >
> >
> > This is kind of an obvious comment, but before doing something like
> > this you should figure out what kind of outgoing software support
> > there would be. If there isn't any, then the students are unlikely to
> > use these computers, and will use something they find easier. Unless
> > they are already very adept with Unix-like systems in general, and
> > Linux-based systems in particular, which seems unlikely.
> >
> Yes. Makes sense.
>
>
>
> > And from a more bottom line perspective, if there isn't anyone to
> > maintain the machines, they will eventually stop working. Even Debian
> > can't completely run by itself, and does require some maintenance.
> >
> True.
> They have a computer teacher. May be we can get him started to to mainten
> the system. Or teach him how to seek help.
Beware that the teacher will likely have limited time and attention.
Ack. I missed this point.
Even assuming the the computer teacher is enthusiastic to learn about
linux, he wil probably have spare time to invest in this new task, on
top of his existing duties as a teacher.
I've made that mistake many times of over-estimating the time of others,
when their enthusiasm was equal to my own.
Ack.
> Another option is we begin with only 2 machines instead of 20. And once a
> month or two revive them if broken.
If you start small then the systems become a premium rather than a
commodity for the students: You want most possible students to explore
and make mistakes in their own pace, not standing in line for few
moments with them or only the top students getting access to them.
I meant, we start from 10% and then to 100%. Assuming the efforts for installation
depends on the number of machines. (I know you have suggested Debian Installer
in later part of the email)
When the school is already ok with allocating the hardware for more
systems, I suggest to make use of that - e.g. by a) investing some (of
your own!) time into constructing a dead simple (for the teacher!)
routine to reset a system to pristine state, and b) propose to have
plenty machines available but also some spares (for the teacher to not
expect *all* of them to be in working state *always*), i.e. setup all of
them (after having created some automated tool that only requires time
for them all to complete) but suggest the teacher to only expect 12 or
15 of them in class, to have spare ones to quickly replace to not waste
time during class.
It's a good idea to reserve some machines to save time at the time of
machine breakdowns.
Even with a fully automated reset procedure, you should still expect to
pay the school a visit now and then, as such procedure can only
meaningfully cover *software* flaws - hardware slowly falling apart is a
different art which I suggest you don't try train the teacher in
handling but instead yourself or a team you form act as "mechanics" to
make "routine checkups" on the system park.
Ack.
The simplest way to offer automated reset of software is to compose a
[preseeding] file for Debian Installer for a fully automated install,
and then burn e.g. 10 identical install CDs with that preseeding file
added. It might take several hours for an install to complete, but
when fully automized you can setup many machines in parallel and the
only real cost is time. The result is certain to be a pristine generic
Debian installation that teacher and students alike can consult the
internet for help using (as opposed to a custom setup that you might
think is "better" for that school but to some extend binds them to you
as the expert of what exact system they have).
Sure. Will explore the Debian installer's pre-seeding file.
If you don't see the benefit of keeping it generic (which I strongly
recommend you to consider first), there are numerous semi-standardized
custom systems, the most famous probably being [DebianEdu] and [FAI].
Will explore DebianEdu & FAI.
I have some experience with running Debian systems at elementary schools
and highscools for the past 25 years, using plain Debian, DebianEdu,
FAI, and more custom-tailored ones. Some years ago me and my partner
made a 3 month [visit to India] to meet fellow hackers interested in
custom-tailoring Debian. I am quite happy to share opinions and
practical details, if you are interested.
Ok. Wow.
Regarding localization, perhaps Debian nowadays perfectly covers the
needs or Marathi - I honestly don't know. But if you consider making
improvements, you might take inspiration from the [Sweecha Project],
whom I have found to be nice folks working on that topic for another
language - Telugu. They might have some insight to share e.g. about
which subtasks are easy or hard.
Whatever you do, enjoy! I know it is an exciting task to engage in :-)
Thats on the spot.
- Jonas
[preseeding]: https://www.debian.org/releases/bookworm/amd64/apb.da.html
[DebianEdu]: https://wiki.debian.org/DebianEdu
[FAI]: https://wiki.debian.org/FAI
[visit to India]: https://couchdesign.dk/india/
[Sweecha Project]: https://swecha.org/content/swecha-project
Ack