#Modula Keyb.
9 messages in this thread
Who knows what lurks within a windows FIB ???
Actually, it is not important to know this – but I am sure the Amiga opens some
other interesting files besides the windows and devices. It would be nice to be
able to peruse thier vital info – some devilish hacking could be done, for fun
and profit.
Alas, the OS befines the length and breadth of the Amiga universe. I just don't
have the time to take it all apart like I did my C64 and IBM. I just get the
strange feeling that this machine is trying to make life difficult for me –
sort of a 'do it my way or I'll make you suffer' impression. I know that the
people at Commodore & MetaComCo thought they had the perfect user interface,
and the perfect OS – but perfection is relative and thier product seems to be
the software manifestation of the closed Mac idea.
All in all, my greatest complaint is lack of time to work with it until I make
it do things MY way – but unemployment is the only solution to that !
— Jim Mason
Jim,
Not closed at all, and in fact rather a little too open for some tastes. The
thing is, everything is fully documented, from the highest level down to the
pins on the hardware, just waiting for deciphering. Unfortunately, it isn't
easy, and considering the complexity of the OS, it really is surprising that
they could document it as well as they did. Now in only they had done it in
assembler, a lot more people could have understood it.
Regards, Larry.
Are you implying that Assembler is easier than C, or are you referring to BCPL?
C, I don't have any problems with, but why on earth they used BCPL is beyond
me!
– Steve –
Steve,
Either one. To me, assembler is a universally "right" language for any
computer, and as such, is the only "right" way to document the machine's inner
workings. I say leave it up to the compiler writers to write and document the
interface for their own language.
BCPL isn't so bad, except for the pointers, and even they make a lot of sense
when implemented on a true 32 bit machine. The stack goes the wrong way though,
but that only points out the foolishness of trying to cram one architecture
into another.
Regards, Larry.
Assembler might have been easier to understand in some cases – even if one did
not know 68K assembler very well.
The sea of structures within structures, while perfectly logical, are often
difficult to follow. I have found that 'Big Picture' tree diagrams showing the
relationships between structures – within the context of a particular
operationsuch as defining & drawing shapes, opening files etc. – can be of
great use in giving the user a better "feel" for the job at hand.
The RKM has this disturbing habit of giving pages of information about somthing
and then – 20 pages later – adding a " Oh yes, Did we mention ? " sort of
sentence containing the vital information to make the first bit work. I find
this a very difficult style to follow, since it becomes necessary to read
practically every sentence in the book just to make sure you havn't missed
somthing vital.
There ARE ways to make this sort of OS more comprehensible. You just can't try
and show the whole of it at once – it's impossible. Too many things are tied to
too many other things. I feel that each TYPE of task needs its own
documentation – very thorough – as opposed to making the scope of the
discussions too wide. Since Commodore has has plans to expand the Amiga market
upwards and downwards, it would be a good move to prepare a new revision of the
RKM and try a better approach.
— Jim Mason
You mean someone has found a technique to make programmers read the WHOLE
manual????? Incredible!!!!!
James, you haven't really lived until you try to find a "bug" in a system
whichexists ONLY because someone forgot to ask him/herself the question "What
do I do if this code gets executed twice at once?" real-time re-entrant
programs take a different discipline to write, and MSDOS and the C64 stuff is
childs play by comparison. I know we would all prefer that our computer
systems behaved "like God intended" (whatever it is we learned first) but the
sad fact is, sometimes there are better new ways to do things. This is ,for
example, why languages evolve and so forth. Just in case you're wondering, no
I wouldn't have designed AmigaDOS the way it is.
….but it ain't a bad real-time multi-tasking OS, which no other home computer
had when it was announced.
Cost/Performancewise – the Amiga OS is great. I have no problems there. Alas it
doesn't seem to get much better as time goes along (except for crash
prevention). But I can live with that too.
Unfortuneately, there are as many notions as to how God intended computers to
behave as there are notions of Gods themselves. My only special beef with the
Amiga is that there seem to be more of those " can't get there from here "
problems than with other computers I have dealt with.
By the way, I HAVE lived – according to your definition. Writing interrupt
driven code on IBMs, where one is sure to have it executed many times before
the first call is finished is indeed a pain in the posterior. I often see
people complain of code that is not re-entrant, but I have never personally
seen a good article or book that addresses just what sort of code works best
for these situations, and what pitfalls lie waiting.
— Jim Mason
Haven't run into any of those "can't get from here to there" things. Perhaps
since my normal job has been with imbeded real-time systems for the last 15
years I happen to think like the people who did AmigaDOS.
In which case, I guess I owe a bunch of people on this net an apology for
not understanding what is relatively obvious to me. <sigh>