20 Jul
2010
20 Jul
'10
7:33 p.m.
I am very happy to say that I have not used a serial
port for anything
in a good 2-3y now, and not for anything more than a very occasional
Whereas I use that interface just about every day for something.
sync of my Psion 7book (and one old PC - Freecycled
away a couple of
years ago - with a serial mouse) since the turn of the century. There
were only a small number of times in the late 1990s.
I really hate RS232. It is the most troublesome interface of any kind
on any computer I've ever used. I celebrate its disappearance with joy
I am the exact opposite. I hate USB wit a vengance...
The problems start with the name, 'Univeral Serial Bus'. It's not
'universal', in that of the 200 or so computers I own, not one of them
has a USB interface or can have a USB interface. It is serial. It's not a
bus. HPIB is a bus, but USB isn't. A bus does not need 'hubs'. One out of
three right is, alas, what I would expect from modern PC companies though
And then there are the technical aspects that I dislike. It's
overcomplicated for a lot of tasks. You can implement an asynchronous
serial interface like an RS232 port in a handful of simple logic chips.
Heck, you can do it electromechanically if you have to (ASR33 anyone).
USB takes at least a microcontroller at each end. Which is often more
than I want to deal with.
But the main problem is that the USB interface is totally non-symmetric.
OK, RS232 is no-symmetric in that there are DTEs and DCEs, but a
null-mdoem cable will often sort that out., There is nothing simialr for
USB. To give an example, I can link my HP95 palmtop (RS232 interface) to
an RS232 printer or plotter. But I couldn't link a modern USB palmtop to
a USB printer without a PC in between. Or I can link my HP48 calculator
to my HP95LX (and since both run kermit, can trivially transfer data
between them),. but I couldn't connect an HP50G with a USB interface to a
mdern PDA.
I strikes me that USB works fine if you want to do what hte manufacturer
intends you yto do and don't want to think about it. The problem is that
almost all the time I want to do things that hte manufacturers of the
devices have never thought of. And I do want to think about it. I'll
stick to RS232 ports.
and I hope never to have to use such a port again. All
the crap with
baud rates, stop bits, parity bits, flow control and all that hateful
1960s-ish nonsense is just a fading memory now and I hope I never have
to refresh it.
It's odd that I rarely have problems with this.
Apple did serial ports right on the Mac. You plugged things in, they
Actually the Mac serial ports are a right pain. Not only because of the
stupid conenctor (which is a pain in the rear to wire), but also because
Apple didn't adhere to any standard. It's not-quite RS422 on the data
lines with RS423 flow control lines. ARGH!!!!!!!
/De gustibus non est disputandam./ However, I would
point out that,
quite aside from my personal life, in my career as a field engineer
It woreis me -- a lot -- that a 'field engineer' would have problems with
the simple concepts of an RS232 interface.
and general IT bod, my Psions were great helps to me.
They held my
appointment books, my client contact details, and numerous databases
of reference info, from RSR232 (F/X: *spit*) pinouts and PC I/O port
and IRQ assignments, standard definitions, command references and all
sorts. That, surely, would be of use to you?
Possibly. But a paper notrbook does just as well. And doesn't need
batteries. This sort of information doesn't chnage, so having it on paper
is not great problem.
Hmmm. I found (and this is not atypical from
discussions with others)
that it takes a couple of hours at most to learn to use an RPN
calcualtor, but after that you never want to go back. It's simply so much
more convnient and easy to use. You don;'t ahve to worry aobut the order
of operations -- the operatios are performed in the order you type them.
So remembering whether -2^4 is (-2)^4 or -(2^4) is not a problem any more=
.
I very rarely use a calculator at all any more - usually just for totting up bills and things. I am perfectly happy with
conventional
arithmetic, never find it a limitation or hazard, and generally
dislike having to adapt my habits to the patterns of logic of
machines. The purpose of computers is as an aid to my mind; they, I
feel, should come to me, not me to them. Thus I never learned, nor
Sure. Thing is, I think in RPN (as do a lot of people without realising
it). RPN is natural if you don't want to have to plan out the whole
calculation before you start -- 'So we've gpt a 10k and a 15k resisotr in
paralell (10000 1/X 15000 1/X + 1/X) and then we have a 22k in parallel
with a 4k7 (22000 1/X 4700 1/X + 1/X), those networks are in series (+)'
and so on... I would never want to go back.
even tried to, any assembly language or machine code;
I bought faster
computers and continued to use BASIC. I dislike C, Perl and so on. I
seldom used to use hex, preferring decimal and binary.
I routinely work in hex and octal (and binary), have no problems handling
any of them in my head. Deciaml is a pain in the rear. And while I use
BASIC, it's certainly not my favourite language...
There is of course an art to finding the right balance, but I tend to
favour ease over efficiency, simplicity over complexity and so on.
I feel it's a big mistake to confuse 'easy to kearn' with 'easy to
use'.
Something may be easy to learn, but if it doesn't help you solve complex
peoblems (that you need to solve), it's no real use. It's better to spend
an afternoon learning to use a more powerful system whciuh actually will
help you with your problems.
RPN ius like that. Anyone can press the keys on a calculator in the same
order as the equation is written in the textbook. They might get the
right answer, they might not (if the precendence rules are differnet).
But RPN is certain.
Same applies to physical tools too. I feel it's better to learn to solder
and do you electronic constuction and modifications in a reliable way
than stick with those plugblock breadboards.
-tony