Proposing Proposals !

[Patrick McC^very]

Hi Everyone

In the thread,  "minix repo BSD repo, testing pkgs"  roarde had said that he thought we were ready for Minix Distros. 

I agree and I want to post about it.

There are 50 zillion Ubuntu based distro and they are not working together to get a single product out the door. It's a lot of wasted effort.

Could we set up a spot were people could upload their Minix ISO install images. Perhaps we could call them proposal  instead of distros ?

For instance, say I setup a desktop and several libraries/applications. I love GnuCOBOL but I don't expect anyone else here to. If I set up some text editor XYZ, then this could get ripped out of my proposal leaving the rest behind and it could be used in the main distro that the project formally commits to.

ISO images would not have source, I am not sure if it would be best to have the option to upload source or if the main contributors could just ask for code with configure settings set or patched makefiles.

What do you think?

-Patrick

[Jean-Baptiste Boric]

Hi,

I take it you want to want to create something like TrueOS (which is basically FreeBSD for mortals).

I personally think MINIX is nowhere near mature enough to make such a proposal meaningful. There are simply too many missing pieces (no USB on x86, no SMP, no 64 bit support, no kernel threads, no standard sound API, lackluster hardware support...) to make MINIX a viable everyday desktop system. We can't compete with Ubuntu by a long shot.

For example, if I were to install MINIX on my freshly mounted Ryzen-based computer I would have to give up on 28 GiB of RAM out of 32, 7 processor cores out of 8, the M.2 SSD, Bluetooth and Wi-Fi connectivity, my USB keyboard and mouse... I'm not even sure I could SSH into it, I'd probably need to use the serial port to interact with it (and I specifically picked a motherboard with one). I couldn't even install it with the installation CD since the computer doesn't have a CD drive.

That doesn't mean those issues can't be solved with enough hard work, but the fact remains that if I want to install MINIX 3 right now on a real computer, with decent hardware support and without wasting 90%+ of its computing power, I would dust off my Athlon 64 rig from 15 years ago. Running an everyday desktop operating system inside a virtual machine to work around that defeats the point of "everyday desktop operating system".

The situation is improving, for example the modern networking stack just got merged. However, consumer hardware is improving at a faster rate. MINIX for me is fundamentally a hacker's operating system, in a good way: simple to get into, easy to tinker with and improve, but it ain't no Ubuntu.

[Patrick]

Hi Jean-Baptiste !

When I mentioned Ubuntu, I didn't mean to say that we should try to go
toe-to-toe with Ubuntu.

My personal goals with Minix right now is to get one monitor setup with
Minix and the other with a BSD install or I might just stick with
Trisquel Linux. I want to learn how to write device drivers and I want
to interface with custom hardware. I want to move as much of my work
flow over to Minix as possible.

Getting these running are short term goals:

Gvim(done)
GnuCOBOL(done)
Clang Assembly(already there)
Latex(to create formatted documents from within Gvim)
xpdf
xchat or weechat
sox
libvorbis
mplayer
youtube-dl
chimera2(x11 browser)
PulseAudio
tcl8.4 + tk8.4
wx-common
yasm
f2c
kicad
xcircuit
xpaint
xloadimage
samba
libsdl
lighttpd
nginx


and many more. I have the source files ready to go, I am just stuck with
a bad network driver and can't get them transferred over without tarring
them and making ISO CD images. It's getting tiresome...

I have no plans to run LibreOffice on Minix or watch youtube videos
anytime soon, or ever actually.

Even with all of Minix's current limitations, I think it was a lot to
offer and I think it will have business value in the near future. Do
companies need a computer to watch youtube or ways to get circuits
interfaced to software fast....

Arduino and friends are great but what if the application requires a
desktop anyways...why bother with microcontrollers unless it's something
deterministic and these devices can't be avoided.

With the proposal idea, people could hack away and they could share.
Right now I don't see much of a mechanism for this. The wiki is there,
true, I am not saying there isn't anything but if we had ways of
showcasing our work to each other, that might have real value and if new
users could download proposal installs that had lots of software already
up and running I think we could retain people better.

Case in point is ctwm. It's so much better then twm and it can be
customized to look really great. Why not have a .xinitrc and .ctwmrc
files set up to impress users with there first call to startx !

I am willing to do this but it wouldn't be much fun if new users never
found it.

-Patrick