[net.micro.amiga] I finally made it!

ewhac@well.UUCP (Leo L. Schwab) (02/08/86)

[ To be eaten by the Ravenous Line Eating Bug of Traal ]

	Phew!

	I finally got caught up with this newsgroup.  If I may, I'd
like to interject a few observations.

Amiga vs. ST vs. MAC vs. C-128 vs. Bambleweeny 57 Sub Meson Brain:

	Interesting.  I enjoy reading emotional stuff, no matter how
mindless it is.  However, it occurrs to me that that's what net.flames
was created for.

FLAME ON

	I have been porting a typing tutor program from the ST to the
Amiga.  Although I haven't actually written any code for the ST, I have
the full source listing to the program I have to port.  Several little
"hidden features" become evident when looking at this code.

	First, the ST is not a controllable environment.  That is, your
code has to determine what kind of system it's running on, and make
adjustments.  For example, one person may have a monochrome system whose
screen resolution is permanently fixed at one value, while another person
may have a color screen, which supports two different resolutions, both
of which are incompatible with the monochrome one.  So your program has
to diddle around, trying to figure out what to do.  In the source I have,
there are endless incantations like this:

	if (hires) {
		[ set character height to something, etc. ]
	} else {
		[ set character height to something _else_, etc. ]
	}

	This is a pain.  Also, from my initial exposure, I don't like GEM.
Apparently you have to reserve memory for what is referred to in my source
listing as "idiotic bindings."  The format of the GEM calls is also
rather confusing.  For example, under GEM, you hafta say:

	vst_effects (handle, 2);

	Whereas on the Amiga, you say:

	SetSoftStyle (RastPort, FSF_ITALIC, 255);

	Everything seems rather non-obvious on the ST.  One thing about
GEM, though.  It lets you turn off the mouse pointer completely.

FLAME OFF

AmigaDOS, TRIPOS, UNIX, OS9, MS-DOS, CP/M, SOLOS :-), .....

	I think AmigaDOS is a pig.  It seems to spend all of its time
thrashing on the disks.  I'm sure TRIPOS was great as a doctoral thesis,
but this is what happens with Gedanken Experiments when you let them go
too far.

	I like UNIX, but I think it is unrealistic to expect a full-blast
UNIX (particularly Berkeley implementation) on a system with only 256K
user RAM.  Not to mention that fact that the UNIX filesystem becomes a
complete loss when you try to make it work with floppies (the system has
to umount the floppy before you can remove it (shades of Macintrash)).

	I've never heard of OS9; where can I learn more about it?

MANX, Lettuce, etc....

	I do not like Lattice.  Whether or not it has to do with the way
the compiler was designed, or the guy who designed it, etc: it makes no
difference to me.  In my view, it is a crock to have to invoke two phases
seperately.  Lattice is also famous for producing bloated code.

	A friend of mine who works extensively with CompuPro boat anchors
has always had the highest praise for MANX.  I have had an opportunity
to play with their beta compiler, and I can see why he thinks so.  MANX
is quick.  MANX is neat.  MANX is (relatively) painless.  MANX generates
an intermediate assembly file (which you can hack on).  MANX gives you
neat utilities (most notably, 'Z', a 'vi' clone (my favorite editor)).
MANX code is *TOTALLY PORTABLE* between all their compilers, no matter
what machine is runs on (that is to say, it will compile successfully
and consistently across all their compilers).

	It has been suggested on the WELL that someone ought to cook up
a FORTRAN compiler so we can make use of all that public domain graphics/
engineering software that apparently exists for FORTRAN.

The 1.1 ROM Kernel Manuals:

	I just got mine today.  A wonderful document; a quantum-leap of
improvement.  It comes in two volumes.  Volume 1 is mostly tutorials and
example code (that works!).  Volume 2 is UNIX-style reference pages, a
dump of the #include files, the IFF-85 docs, and example library and
device driver code.  As well as how to make your own boot disk (not
Kickstart flavor, sorry).

	It's well worth the wait.

	It suddenly occurrs to me that this dissertation shouldn't be
here.  If this is true, forgive me, I'm new to the net.

					Leo L. Schwab
					well!ewhac
					dual!unicom!schwab