Re: [MiNT] What's in, what's out?

[Guido Flohr <gufl0000@stud.uni-sb.de>]

Hi,

On Sun, Nov 28, 1999 at 12:20:59PM +0100, Julian Reschke wrote:
> > The /kern fs provides the system interface.
> 
> So it has a sort of dual interface?

No, a triple interface: the hypergalactic, the eclectic and the sophistic
interface. You don't have to bother about the differences, tho'.  They are
all the same.
;-)

> > Why give it to Frank?  You can upload it to a hundred of ftp servers.
> 
> It has been available for year now. Frank complained about it not being
> open-source, that's why I offered to give him the sources (and take over
> development) if he want's them.

He obviously doesn't.

> > Being human-readable does not necessarily imply not-machine-readable.  You
> > can easily parse the information in your programs.
> 
> Yes. But do you think it's such a good idea to add a text-based operating
> system interface just to make life easier for a few shell scripts, rather
> than defining a proper (and more efficient) binary interface?

Yes.  A machine can read both machine-readable files and human-readable
files.  Most humans tend to feel better with human-readable files.

> > > 2) I cited this as an example for a poor API that found it's
> > way into the
> > > kernel because it wasn't discussed before.
> >
> > If somebody would have stood up and shouted "it's buggy!" it would have
> > been changed.  But you cannot change it now, years later.
> 
> Right? So why did that happen? Because the interface was just added, not
> openly discussed.

Konrad implemented it and there was time enough to discuss it then.
It has generally been the case with the MiNT kernel that discussions were
based on a sample implementation.  Take for example VM: It was
implemented, it proved not to work, and it was removed then.

> > SIGPWR is neither a weird idea nor a misunderstanding.  It is a System V
> > feature and its existance aids in porting power management software to
> > MiNT.
> 
> Yes, how can you say it's important enough to be added as the last available
> signal, if you really can't tell us where and how it is used?

If there will really be another signal that would be desirable we can
still redefine SIGPWR to SIGWHATEVER, unless SIGPWR is already widely
used.  But if SIGPWR is used by then, there was obviously a need for it.

I never said that I wanted to port the power daemon to MiNT.  But if
somebody ever wants to do that (and power management is important for some
users) then init has to be aware of it.  And I do init.

> > Yes, it is.  Without a C compiler you cannot compile your own kernel.
> 
> Compiling the kernel is a nice thing for a few people. Once you start
> telling people that they *have* to be able to compile their kernels, you
> probably won't be able to convince people to migrate from that other
> operating system, right?

My little brother compiled his own Linux kernel the day after he had
installed Linux (he never got in touch with Unix before).  They have
xconfig and he can read.  With the MiNT kernel it would even be easier (no
modules) to build a tailored kernel.  We only need a decent user interface
to the build process.

Ciao

Guido
-- 
http://stud.uni-sb.de/~gufl0000/
mailto:guido@atari.org