Mail Archives: djgpp/2001/06/29/03:44:41
On Thu, 28 Jun 2001 23:28:24 -0400, DJ Delorie <dj AT delorie DOT com> sat on
a tribble, which squeaked:
>True. I guess we're used to the "common" user, who knows where Intel
>keeps their online docs.
Actually, the "common" (or at least "lowest common denominator")
hasn't a clue that they exist.
>It's hard to know how much you need to tell people; assume too little
>knowledge on their part and you could inadvertently insult them.
If they are inadvertently insulted by such a thing, they need a
thicker skin.
>You can always ask for more specific information, which we'd be happy
>to provide, if you need it.
And bandwidth grows on trees? :-)
(Begins seriously considering futures in agriculture...)
>True, but the topic was well discussed, and finding a specific topic
>in those manuals shouldn't be too hard. Intel is pretty good about
>documenting their chips.
Then why are they so bad at marketing them? :-)
(The very box I'm writing this on has an AMD chip in it.)
Your description of technical manuals suggested to me lists of opcodes
and timing and alignment dependencies from which one would probably
have to piece together the global behavior for various alignment
related questions from looking at a dozen different spots. If they
have proper textual discussions of coding for these chips in theory
and practise, and these are rationally organized, OTOH...
(I have an old nostalgia copy of the C-64 Programmer's Reference
Guide. IIRC, it has source for an assembler, some stuff on C-64 BASIC
internals, kernel and sound and video internals, and the CPU opcode
tables, as well as a motherboard circuit layout. I don't recall
anything about instruction timings or alignment, but then again that
bitty 8-bit chip didn't *have* any timing or alignment dependencies,
IIRC ... :-) Seems very claustrophobic to program now, yet there was
something about having a computer whose inner workings could be
documented in a not very huge book and lay within the complexity
horizon of one person's mind...)
>I meant in the general way that we deal with questions and answers.
>If you don't agree with the technical aspects of the software, you
>probably *do* belong here :-)
Why, so you can be flamed to a nice golden brown? Or for reasoned
debate? *grin* Of course, nothing I've seen leaves me to question the
software's technical merits. The working Mandelbrot rendering engine I
put together blows rings around Fractint, although this is partly
because my boundary tracing algoirithm is smarter than theirs at image
borders and on the x-axis, i.e. when forced to detour from the
fractal's contours -- it uses distance estimator calculus tricks to
skip large areas along these detours without missing part of a
dendrite or anything. Of course, Fractint is also still being compiled
in the medium memory model of a 16 bit compiler, for some reason...
>True, but that's the price you pay for easy access. And DJGPP's
>regulars are pretty mild tempered. Most of our flame wars start with
>cross-posts, so you get combinations of people used to different
>"rules of conduct" and that's when the worst personality conflicts
>happen.
Or is that "rules of engagement"?
The insidious case is the "slow-boil" flamewar, 3 of which seem to
have been narrowly averted here of late. Someone posts something that
seems vaguely insulting or demanding (I may actually be guilty of
this, and Eli certainly is) or at the very least is ambiguous in that
regard, and someone gets ruffled feathers and makes no attempt to hide
the fact. There's a gradual escalation, and then someone goes
nonlinear... I've seen it happen before...
>Surprisingly, for DJGPP that's not really true. Many of the regulars
>are from countries whose native language isn't English, and we've got
>enough coverage for most languages. Perhaps English gets a larger
>*volume* of responses, but the other languages are just as likely to
>get useful answers.
It's a small group sampled with some bias from the general population,
so there's bound to be the odd, somewhat large statistical deviation.
--
Bill Gates: "No computer will ever need more than 640K of RAM." -- 1980
"There's nobody getting rich writing software that I know of." -- 1980
"This antitrust thing will blow over." -- 1998
Combine neo, an underscore, and one thousand sixty-one to make my hotmail addy.
- Raw text -