Is this true ?

[Garry Ricketson]

I recently read this comment: "minix3 doesn't support dual core processors or threading."
, on another forum,  can anyone confirm if it is true ?

Thanks

[Jean-Baptiste Boric]

I recently read this comment: "minix3 doesn't support dual core processors or threading."
, on another forum,  can anyone confirm if it is true ?

Once upon a time MINIX 3.1.8 had experimental SMP support, but it bitrotted a long time ago. You can use MINIX 3 on multi-core processors, but then it will only use one core.

[Garry Ricketson]

Ok, thanks,  for clarifying on this,

[Martin Vahi]

I'm a newby myself, trying to learn MINIX, but
in Estonia a good salary for a senior software developer
in 2017 is about 4k€ after paying all taxes. That's about 10k€/month
for salary fund burn rate per developer. And that's a GOOD SALARY.
Most people get paid roughly the same money that
janitors get paid in Finland, about 1.5k€/month after all
taxes have been paid. The Finnish janitor salary is
considered a good "average" salary in Estonia for
secretaries and various state clerks.

So, 1M€ buys 100 man-months at 10k€/month. In the past and
for doctoral students, the Finnish janitor salary would have
been good enough, so 1M€ would have bought about

    1M€ / 3k€ ~333 man-months of development time.

The MINIX3 has had millions of € for development and
the code base and experience of MINIX1 and MINIX2.

Yet, no GCC, no SMP.

I as a freelancer, who is struggling to make my ends meet
am certainly envious. In Estonia a whole house can be bought
for about 300k€. For 3M€ one can buy 10 houses in Estonia.

That's how I see the monetary side of the MINIX3 development.

Oh, and buy the way, Estonia is not even with the cheapest
place with the smallest salaries in Europe. Germany has
bigger salaries and cheaper prices. That makes the picture
even more contrasting, extreme, from a perspective of an
Estonian like me.

I know that my current posting is not very productive, but
that's one issue that I have always wondered about, when
studying the MINIX3 project. It just catches my "eye" very effectively.

Thank You for reading my comment.
 

[David van Moolenbroek]

Hello,

On Tuesday, August 22, 2017 at 2:30:44 PM UTC+2, Martin Vahi wrote:

The MINIX3 has had millions of € for development

That is simply not true, and I really wish people would stop perpetuating this myth. Professor Andrew Tanenbaum's research group received 2.5 million euros of grant money from the European Union for the purpose of conducting *research*. As a result, the money was spent mainly on research, and a relatively small amount was spent on development in support of the research. The research resulted in a large number of publications - several in top venues - and a number of doctorates. That was how the money was intended to be spent. MINIX 3 has never had "millions of euros for development" and with that as a given, it is amazing how much MINIX 3 has been improved during that time, and how incredibly more usable it is now when compared to MINIX 2.

Also, keep in mind that not everyone necessarily agrees with you on what is important for the project. For example, we are purposely moving away from gcc to llvm for a variety of reasons, including gcc's incompatible license. You may not realize how much work it takes to maintain something as hairy as a gcc port, even if it was working before. In fact, a mysterious, complex breakage of gcc after a necessary upgrade is the direct reason why our ARM port is currently non-functional and why we have not been able to release 3.4.0 yet.

Finally, I would suggest that you contact the financial department of the Vrije Universiteit Amsterdam in order to find out how much it actually costs to employ a research group. I suspect you will be surprised.

I know that my current posting is not very productive

Correct.

Regards,

David

[Martin Vahi]

teisipäev, 22. august 2017 14:14.44 UTC kirjutas David van Moolenbroek:
> ...

> On Tuesday, August 22, 2017 at 2:30:44 PM UTC+2, Martin Vahi wrote:
> >
> >  The MINIX3 has had millions of € for development
> >

> Professor Andrew Tanenbaum's research group received
> 2.5 million euros of grant money from the European Union
> for the purpose of conducting *research*. As a result,
> the money was spent mainly on research, and
> a relatively small amount was spent on development
> in support of the research. The research resulted in
> a large number of publications - several in top venues - and
> a number of doctorates. That was how the money was intended to be spent.
> MINIX 3 has never had "millions of euros for development" and
> with that as a given, it is amazing how much MINIX 3
> has been improved during that time, and how incredibly
> more usable it is now when compared to MINIX 2.

> ...
> Regards,
> David

Thank You for the Answer.

I like to point out that I am NOT trying to tell others,
how others should run their project. That includes the
GCC versus LLVM choice. As I said in one of my posts,
I'm a person, who believes that the one, who does the real job,
executes the task, must have the final say on how to execute
the task and obviously if the developers prefer LLVM for whatever
reason, then it's THEIR CHOICE TO MAKE. I'm not protesting their choice,
not to mention, that I'm absolutely NOBODY in the MINIX project.

However, what regard to the argument that the money was
spent on _research_ and that _research_ in operating systems
is somehow something different than DEVELOPMENT of operating systems,
then as long as operating systems are SOFTWARE,
the _research_ is a SUBTASK of DEVELOPMENT. If it weren't
a sub-task of the _development_, then nothing
technically good comes out of the _development_ and
the _development_ might as well be placed into qoutes, like "development".
That is to say, in my view algorithm development and the
STUDYING OF THE PROBLEM is, in my view, ESSENTIAL PART of the _development_.
If people document their studying activity by writing
scientific papers, then that helps everybody and is a
very welcome phenomenon, but millions of € for the studying
activity is still A LOT by Estonian "standards", regardless
of how they write it to the books. I'm not saying that
they wasted it or misused it and I believe that may be
according to Your local price range everything was within
competitive price ranges, but totally independent of what
I think or don't think, what I write here or omit, millions
of € for a software project, even if most of the money is
locked for studying the problem, is A HUGE PILE OF MONEY and
the project deliverables are beautiful and good, but
not that impressive, if weighted to the HUGE PILE OF MONEY.


It's just, how it appears to me and to many people, who
earn roughly a Finnish janitor's salary by doing software
development. You may argue that may be the Estonian
software developers, who earn so little, ~2k€/month after
all taxes have been paid, are just remarkably crappy at their work,
specially given the global market, but, if You use that
argument, then I can't give You any counter-arguments, because
other people, including You, should judge, how crappy or good
I am at what I do.

That is to say, I just reported, how the
Western-Europe style pile of money and the
project deliverables appear from North-Eastern Europe perspective
and the summary of that perspective is: beautiful results, but
the efficiency is really shoddy.


Thank You for reading my post.

[Garry Ricketson]

Thank you again on the reply, it is good to know it (Minix3) still can be installed
on a machine with dual core processors.
    I am going to be installing Minix 3 on a dedicated server, the servers are all "dual core",  my hopes are they won't have any problems installing.
 At the moment I am anxiously waiting  to hear how the install goes. Since I do not have "hands on access", I am depending on some one else to
install it, my access will be remote only. 
  If anyone else is at all interested in this "project",  I will be happy to "share" the server, with others that have a real interest in Minix 3, the server will be paid
for, for 1 year, after that it will depend on how much interest there is and other factors.
----------------- off topic-----------------------

I don't know why this guy:

Martin Vahi

This message has been hidden because you reported it for abuse. 

  did not start a separate topic  , the comments and rumours are of no use to me.

I know that my current posting is not very productive

Correct.

 No it is not productive, and when one know it is not productive , it would be appreciated if nothing is posted. 

 

On Monday, August 21, 2017 at 9:06:21 AM UTC-5, Garry Ricketson wrote:

[David van Moolenbroek]

Hello,

On Tuesday, August 22, 2017 at 10:46:53 PM UTC+2, Martin Vahi wrote:

That is to say, I just reported, how the Western-Europe style pile of money and the project deliverables appear from North-Eastern Europe perspective and the summary of that perspective is: beautiful results, but the efficiency is really shoddy.

Your assessment is rooted in a fundamentally flawed understanding of the goals of scientific research. Moreover: nobody asked!

Please refrain from posting these unconstructive, insulting, ignorant, off-topic diatribes. If misguided rants are your thing, start a blog.

Thanks,
David

[Martin Vahi]

kolmapäev, 23. august 2017 12:11.03 UTC kirjutas Garry Ricketson:
> ...

> ----------------- off topic-----------------------
> > I don't know why this guy:
> > Martin Vahi
> > This message has been hidden because you reported it for abuse.
> >  did not start a separate topic,
> >  the comments and rumours are of no use to me.
> >  > I know that my current posting is not very productive
>
> Correct.
> No it is not productive, and when one know it is not productive,
> it would be appreciated if nothing is posted.

Thank You for Your answer.

The reason, why I self think that there is
still some productivity in my posts about the
monetary inefficiency of the MINIX project, despite
the posts not being _VERY_ productive, is that
social systems are also SYSTEMS, like mechanical SYSTEMS,
electrical SYSTEMS, sofware SYSTEMS.
If there is some elephant sitting in the room,
then that elephant can "clog the gears". For example,
the Debian Linux project is not just a technical
project, but it is also a social project.

    (2016 11 xx Mini DebConf at Cambridge Session by
     Lucy Wayland About Diversity and Inclusion BoF)
    https://www.youtube.com/watch?v=a7yyWpIpeUg

The way I see it, it is precisely the fact that the
social aspects, the SOCIAL SYSTEM, have been taken
care of, what allowed both, the Linux, and the Debian
projects to get huge amounts of developers working on it.
On the other hand, the MINIX project seems to lack
manpower and volunteers and that is not some
"rumour by Martin Vahi", but a statement that has been
claimed by MINIX3 core developers at the MINIX3 Google Groups
forum and at GitHub discussion threads.

The difference between the MINIX and Linux+Debian is
that in the case of Linux and Debian the projects did
not get millions for the core group. That set the
expectations low for the rest of the volunteers and
by the time the IBM and the rest came in with their
truckloads of money, the Linux was pretty much
up and running. In the case of MINIX there exists
the question, why should a volunteer make a great
effort to contribute to the project, if the core
team is "swimming in money" and has not even made
sufficient effort to get those features
that are important for the volunteers, like the SMP,
completed.

Censoring me here is not going to make the problem
go away, but addressing the problem might help
the MINIX project as a SOCIAL SYSTEM, not just as a
technical project, to get some wind under its wings.
I was hoping that I don't have to spell it out
like that, but it seems that people didn't figure
it out on their own.

Basically, the way I see it, the path forward might be
to offer the nonpaid volunteers a FULLY FUNCTIONAL SYSTEM
(read: has SMP,GCC,pthreads, etc.) and then
the nonpaid volunteers can contribute small bits,
port packages/ports, add drivers, etc. like it is done
in Linux and with Wikipedia articles. But currently,
the leap for porting GCC, which is needed for GCC specific
userland software, and the leap for fixing the SMP,
which even the paid team found "too big amount of work", is
unrealistically huge things to ask from unpaid volunteers,
not to mention, if even the paid team, who already knew the project
internals, didn't do it, then why should the volunteers,
who have to start learning the project internals.


All in all it seems to me that the MINIX has been a
research project first and production project for
"real life" applications second. The requirements for a university
research projects tend to differ a lot from the
requirements of an open source software project that
is meant to be in production use. For example, the
prioritization of interesting scientific articles over
basic SMP support is clearly one such difference. Neither can
any production oriented operating system project ignore
the userland software that just happens to be GCC specific.
So, some transition period from one development model
to another, social system wise and priorities wise,
can be expected.

Thank You for reading my post.

P.S. I'll shut up in this thread now, which means that
if I answered any followup questions in this thread,
then I would break my pledge to shut up in this thread :-D