Re: STiNG, CAB, and Multitasking (strikes back)

[Jo Even Skarstein <joska@nuts.edu>]

On Mon, 9 Mar 1998, Peter Rottengatter wrote:

> On Mon, 9 Mar 1998, Yves Pelletier wrote:
> > In my ideal world, (the one where poverty and disease are
> > eradicated :-), MiNT would be the very first thing in the auto
> > folder, as befits a loadable operating system module.  Then,
> > TSRs and other programs would load behind it.
> 
> I agree here, though I must add, when talking about major irritations,
> then one should take into account that the whole business of having
> some programs before others is totally confusing for many users that
> aren't that deep into TOS details, especially as most desktops sort

I have never had any problems with sorting the auto-folder, but I know
that some people have. But in most cases the documentation mention it
if a program needs to be run before or after some specific piece of
software.

I can't see what the big deal is, some programs depends on the
presence of other programs in other situations as well. With proper
documentation this isn't a problem at all.

> the files listing of the AUTO folder. This is one major point in favour
> of trying whereever possible to remove such dependencies. Another one

It's not just MiNT, a lot of programs (like NVDI, Nova VDI, MetaDOS,
MagiC, HS-Modem...) depends on the presence (or no presence) of other
programs or features installed by other programs. Can all these be
"fixed"? Doubtful, the running-order in c:\auto is something we all
have to live with.

> Now in the current situation of bending system vectors, the change that
> removes this dependency for a fair number of programs is simple and 
> small and safe, and it does not even worsen the compatibility with other
> programs. Therefore I see no reason to dismiss the change. Are there any ?

Well, if this particular fix doesn't have any negative effects I can't
see any reasons not to do it. But it's also a matter of time and
resources, there are a lot of other things in MiNT that's more urgent
to fix.

I saw you referred to the MiNT-docs previously Peter ("MiNT should be
the last program in the autofolder..."), and you have a very good
point there. The current docs are not very good (in many cases
outdated), a "MiNT for dummies" should definelately be written. 

BTW. why can't programmers put in checks in their TSR's when possible?
The only program I've seen recently that does this is MetaDOS, which
says "MetaDOS must run before MiNT!" if it detects MiNT. In many cases
it's so easy to put in these checks (not only for MiNT, but for mosts
features it needs or doesn't work with) and notify the user instead of
just continue.


/*
** Jo Even Skarstein    http://www.stud.ntnu.no/~josk/
**
**    beer - maria mckee - atari falcon - babylon 5
*/