librelist archives

Re: Roadmap for LuaOS

From:

Stefan Reich

Date:

2011-12-09 @ 01:33

Hey Yu, thanks for writing in & for your compliments.

A little side note: I currently spell the OS's name "Lua OS", but if 
people like "LuaOS" more, we can also switch to that. It does looks kinda 
catchy too.

(FWIW, "JavaOS" was also written without a space back then. But then 
again, we don't necessarily have to follow Sun's example in everything :)

Regarding your question - it seems like there are two directions in which 
Lua OS is currently expanding:

Direction 1. Improving the Linux part. There is always something to 
improve around there. My current favorites are these:

1.a Make a new start menu. I have written one in Java, but it would be 
nice to make a new one in e.g. Lua+IUP. (This would also start faster :)) 
This task should not be hard for any decent coder, and it would be great 
to have a nice searchable start menu that can later be enhanced with some 
AI or some such. Anyone up for this? :)

1.b Improve the Linux sound system. This is currently the one thing that 
annoys me a bit in LuaOS. (Huh. Am I switching to that spelling?)

It seems that sound can currently only be played by one app at a time. If 
Exaile is set to 'paused', no sound will play in Iceweasel. That is kinda 
ridiculous... :)

Also, setting volumes and choosing a default device (e.g. speakers vs. USB
headphones) is complicated. Especially when one of them is not plugged in 
all the time.

I guess there is probably some solution for all of this that does not 
involve all too much work. Maybe actually write a sound server in Lua/C - 
with a nice little UI? I haven't gotten into the details of that yet 
anyway. Maybe a Linux expert has some comment on this?

Direction 2 is of course: Building parts of the next-gen architecture.

"Lua OS Next-Gen" (the envisioned purely Lua-based OS) is a combination of
different innovations on the language and/or runtime. So we can tackle 
these one by one and build some prototypes; each focusing on a specific 
feature. The first such prototype was "Safe Lua", next was "Mobile Lua" - 
and probably some more are to follow.

Hope this gives you a little idea of where things are going :)

Incidentally, a few people have suggested switching from a Linux basis to 
a BSD basis. I see that they like BSD, but I don't really buy into the 
argument that BSD is better than Linux for our purposes - or even 
necessarily "the right thing" to use as one person suggested.

BSD, to me, is far away, and I have no intention of really giving it a 
spin. Also Linux seems a lot more popular. BSD is pretty exotic as a 
desktop platform. I doubt that all the things I'd want to use would run 
easily on BSD.

Everything that exists as free software is available for Debian. Is the 
same true for BSD? Again, I find it just too exotic.

Finally, they say BSD is developed by a more "hierarchical" organisation. 
But this organisation seems kind of shadowy. I have some idea of who Linus
Torvalds is. But I have no idea who the makers of BSD are. Maybe they are 
not media people? Or maybe there is some sort of secret?

Either way, Linux just seems like a much more natural basis for a popular 
desktop OS than BSD.

Cheers -
Stefan

I almost wrote luaos :)

Direction 1 seems like what Google has been doing so far with ChromeOS. I 
think this would be a good start, since writing an OS from scratch seems 
such a huge task you might lose energy at some point. Better use linux 
framework for now then.

Direction 2 is 'writing the OS from scratch' right? Using LuaJIT for speed
perhaps? That'd be very interesting as the OS would be a real novelty, 
because I've never heard of an OS successfully written in a dynamic 
language. You'd have a lot of room to innovate, but what a huge task!

Question: Have you ever thought about starting a "WikiOS" where people 
would edit the OS as they will like in Wikipedia? Or maybe start this only
for the writing of LuaOS libraries and drivers while you stay with the 
kernel and the design of it? The big picture of LuaOS, so to speak!

Anyway good luck with the project! It is very refreshing! I am not a 
programmer per se, and yet not a complete noob in it if you consider bash 
scripting as programming :) I'm a network admin by trade.

Regarding linux/bsd.. I install OpenBSD as a firewall in many companies. 
Most of the medium/big companies I've been involved in uses OpenBSD as its
sole front defense against threats from the Internet and from inside too 
(along with a well designed network of course, with DMZ, IDS/IPS, etc). 
It's almost the default firewall around I'd dare to say. Take a look at PF
as an example, it's really powerful! But I never used it as a desktop, 
though I hear they have a package base with all these goodies, and again I
hear it is well designed (easy to install/remove and it just works). No 
personal experience with Desktop there to confirm though.

Regarding the *BSDs, I'm not aware of the others since for my work OpenBSD
is the only choice, but AFAIK FreeBSD is the most popular, and OpenBSD the
most secure and free from patents.

Also to install software on OpenBSD they have what they call 'ports tree'.
It is so easy they should advertise it more :) Basically its just 
'pkg_add' this, 'pkg_delete' that, and they'll take care of dependencies 
for you.

Take a look at the software they offer here: 
http://artfiles.org/openbsd/5.0/packages/i386/
For how to use it take a look here: http://www.openbsd.org/faq/ports/index.html

Anyway, I've been a Windows guy as network admin most of the time, using 
linux and bsd when it fits, but OpenBSD has made my life so much easier 
that I'm starting to consider trying it a Desktop on my laptop!


Yu Fang

莊周夢蝶

On 11 December 2011 00:28, Yu Fang <yufang56@yahoo.com> wrote:

> Direction 2 is 'writing the OS from scratch' right? Using LuaJIT for speed
> perhaps? That'd be very interesting as the OS would be a real novelty,
> because I've never heard of an OS successfully written in a dynamic
> language. You'd have a lot of room to innovate, but what a huge task!
>

There are a myriad of Forth-based operating systems out there: basically
one per embedded Forth system (plus there's ForthOS if you want to count
that).

-- 
"Perhaps people don't believe this, but throughout all of the discussions
of entering China our focus has really been what's best for the Chinese
people. It's not been about our revenue or profit or whatnot."
--Sergey Brin, demonstrating the emptiness of the "don't be evil" mantra.