New DOS shell for Windows coming

[Franc Zabkar]

On Sat, 18 Jun 2005 08:28:02 +1000, Franc Zabkar


Uhh.. did Win98se *EVER* cut it? :>

Give me a break. It took me ages to move on from Win95B.

I've used a number of command lines over the years and never found any
of them to be intuitive. Usable and sometimes very useful, yes, but
definitely not very intuitive.

Some that come to mind include DOS, CPM, PRIMOS, KRONOS, RSX-11M,
CGOS. I found all of them relatively easy to learn. In fact Primos is
so intuitive that I was able to connect a Prime mini (an unfamiliar
machine) to a DG mini and perform file transfers merely by referring
to Prime's online help. And I did this within a couple of hours, using
only edit and copy commands, via an RS232 interface.


- Franc Zabkar

 

[Nate Edel]

As far as the current situation goes, I think it's pretty
clear that current Linux distributions are by far the most
inefficient and resource hungry systems ever created. The system
I'm on now (Linux Fedora Core 2) just doesn't run well with
less than 512 MB of RAM, and it needs at least 768 if I
want to avoid excessive swapping when I have a lot of Firefox
tabs and other stuff going at once.

You could try a more efficient distribution - or, if you use KDE on Fedora,
even try switching to Gnome... or a lightweight Window manager, for that
matter.

KDE on Gentoo seems to run OK on 384 and well on 512 (w/ a P4-1.7ghz machine).
Gnome on Gentoo seems to run OK on 256 and well on 384 (ditto).
blackbox on Gentoo seems to run OK on 128 and well on 256 (ditto).

blackbox on Gentoo (using kdrive as a X server rather than XOrg) with a 2.4
kernel runs OK (on a 96mb Pentium MMX 233 system, no less!)

 

[Alexander Grigoriev]

Well said, bro!

A scripting language is a kind of programming language.
Bash is reminiscent of BASIC. I think of it sort of as
BASIC's crazy uncle. It's got a weird and limited set
of data structures, weird and limited looping constructs,
a weird and limited notion of function, etc. BASIC itself
is pretty mediocre, but writing for its crazy uncle Bash
is pretty unpleasant.

The sad thing about Unix is that the technical incompetence
of its developers not only makes for lousy technology,
but also blinds its users to the need for improvement. They think
weird things, like a scripting language isn't a real programming
language, so you don't need to know anything about programming
languages to create it. They think they can write things like
that by the seat of their pants, knowing nothing about computer
science, or even being that good at programming. And they end up
congratulating themselves for the superb job they've done, because they
don't know enough to realize that they should be working within the
framework of 4 decades of programming language research. In that
tradition, the crazy uncle of one of the lousiest languages to
survive from the 1960's wouldn't count for much, but in the
Unix world, it's treated with reverence, like it's some
powerful and ingenious invention. So you've got these technically
incompetent folks, who know nothing about computer science,
who are even more clueless about usability and marketing,
calling each other gurus and congratulating themselves for
knowing more than the folks in Redmond.

And what you end up with, is the Unix people let BASIC's crazy uncle
out of the attic, and made him run a big part of the Unix
show. And the other parts of the show are run by the other
loony systems, like X Window and Emacs, which are designed
with equal technical incompetence.

 

[Alexander Grigoriev]

One of design objectives of Windows 95 Gold was running in 4 MB. Of course,
it's not quire happy with it. But I've run it with 8 MB and it was not bad
at all. 8 MB Compaq-branded DIMM then costed 300.
When (around 1990?) I was told that I need 8 MB to run NT 3, I thought:
that's a big system!

I've seen Windows 2000 booting with 32 MB. Like watching paint dry.

 

[Alexander Grigoriev]

Man, you should not torture yourself with Gnome. This crap is as bad as
OpenWin.

 

[Tony Hill]

One of design objectives of Windows 95 Gold was running in 4 MB. Of course,
it's not quire happy with it. But I've run it with 8 MB and it was not bad
at all. 8 MB Compaq-branded DIMM then costed 300.
When (around 1990?) I was told that I need 8 MB to run NT 3, I thought:
that's a big system!

I've seen Windows 2000 booting with 32 MB. Like watching paint dry.

Ugg, I couldn't imagine trying to use Win2K with only 32MB of memory.
A while back at a summer job I had a computer that ran Win2K with 64MB
of memory and that was painful enough! I got a fair bit of swapping
by just booting the OS, let alone running any applications.

Of course, I was still WAY happier with this solution vs. the other
option of using Win98. I'd much rather due things slowly once than
having to do them quicker but re-doing my work all the time due to
Win98 crashes. I know some people claim that they can run Win98 for
weeks on end without it crashing, but for me it would crash nearly
every day (often multiple times in a day) on *EVERY* PC I've ever used
it on.


As for Linux, I've run it in a variety of configurations with a wide
range of memory sizes. One of the real nice points of Linux is how
flexible it is, on one system I could run a really trimmed down
version and it'll work just fine on 16MB (might even be able to
squeeze into 8MB). On my main system I've got 512MB so I load it up
with all the features, full KDE GUI, etc.


Honestly though,I don't think there's any point in talking about
"bloatware" these days. 1GB of memory costs only $100 and that's
plenty of memory for 99% of users out there today. Hardware has
pushed ahead much faster than software. If you ask me, I'll take the
bloat any day if it results in better features. New hardware is
dirt-cheap, if the problem requires another 512MB of memory, I'll
gladly spend the extra $50 to get that 512MB of memory when the
alternative is less functionality from the software. Now obviously
bloat without adding functionality isn't going to help anything, but
so long as the software is improving, I'm all for it.

 

[Tony Hill]

Some that come to mind include DOS, CPM, PRIMOS, KRONOS, RSX-11M,
CGOS. I found all of them relatively easy to learn. In fact Primos is
so intuitive that I was able to connect a Prime mini (an unfamiliar
machine) to a DG mini and perform file transfers merely by referring
to Prime's online help. And I did this within a couple of hours, using
only edit and copy commands, via an RS232 interface.

The fact that it took you a couple of hours and required using the
on-line help means that this interface was definitely not intuitive.
Easy to learn, maybe, but definitely not intuitive.

As someone more famous than myself once said:

"The only "intuitive" interface is the nipple. After that it's all
learned."

 

[Franc Zabkar]

The fact that it took you a couple of hours and required using the
on-line help means that this interface was definitely not intuitive.
Easy to learn, maybe, but definitely not intuitive.

I'm including the time taken to familiarise myself with totally
foreign hardware, and the time taken to make up a cable. The actual
commands I used were PRINT on the source machine and EDIT on the
target machine. I needed to consult the online help to find out how to
place the target machine's text editor into insert mode. IIRC, the
subcommand was "I" which *is* intuitive.


- Franc Zabkar

 

[Grumble]

Actually I intend to migrate to Linux when Win98se no longer cuts it.

Have you ever tried the *unoffical* Win98SE Service Pack?
http://exuberant.ms11.net/98sesp.html

 

[Nate Edel]

When (around 1990?) I was told that I need 8 MB to run NT 3, I thought:
that's a big system!

It came out in 1993, sometime around the end of the summer or the fall.

NT 3.1 (which was the first released) raised the requirement shortly before
the release date from 8mb to 16mb. I got 16mb of non-brand ram for about
$500

Then it turned out that it didn't support Plus hardcards, and I had to get a
new hard drive and controller to run it.

I've seen Windows 2000 booting with 32 MB. Like watching paint dry.

It's not much better on 64. The real running footprint of just the OS is
about 80MB, so 96MB is the practical minimum to do *anything*.

 

[chrisv]

IIRC TI calculators used a 9-level stack, albeit a hidden one. In fact
one could access the stack on a TI-59 using undocumented instructions.
So it's clear that TI machines *evaluated* the expressions in the same
way as HP calculators, it's just that TI calculators *accepted*
expressions as they were written.

It's irrelvant how TI handled the calculator evaluated the
expressions. It's the UI which is under discussion, here, and Robert
(and many others, including myself), prefer the RPN UI.

No mental gymnastics were required.

Are you implying the more "mental gymnastics" are required to use an
RPN calculator? Because that would be false.

Furthermore, the often touted claim that HP calculators were much more
economical with keystrokes was a myth which I disproved many a time.

"Much more economical" sounds to me like a straw-man argument. "At
least as, and often more, economical" is the truth, not a "myth", in
every case I've compared the two methods.

In any case, whether or not RPN is better than algebraic is
irrelevant. My contention is that a person should not have to adapt to
technology, but that technology should adapt to him. For example, I
should be able to pick up any unfamiliar calculator and key in "1+2=",
not the counterintuitive "1 Enter 2 +".

A superior (or even only preferred) tool is quite often worth the
investment in how to use it. It's ludicrous to discount RPN
calculators because they may take a couple days getting used-to.

The inner workings should be
transparent to the user interface.

True, but many of us feel the RPN UI is vastly superior, especially
when used in conjunction with multi-line displays.

 

[keith]

It's irrelvant how TI handled the calculator evaluated the
expressions. It's the UI which is under discussion, here, and Robert
(and many others, including myself), prefer the RPN UI.

So do I. It took me about four hours to master RPN when I got my HP45,
and I never looked back. Since, I have had a difficlut time with
arethmetic calculators.

Are you implying the more "mental gymnastics" are required to use an RPN
calculator? Because that would be false.

Of course there are more "mental gymnastics". Learning something new does
take a certain abount of mental capacity. ;-)

"Much more economical" sounds to me like a straw-man argument. "At
least as, and often more, economical" is the truth, not a "myth", in
every case I've compared the two methods.

When the TIites saw how I did vector sums they shut up about the
superiority of TI's UI.

A superior (or even only preferred) tool is quite often worth the
investment in how to use it. It's ludicrous to discount RPN calculators
because they may take a couple days getting used-to.


True, but many of us feel the RPN UI is vastly superior, especially when
used in conjunction with multi-line displays.

Even with a single line. I've never used a calculator with multiple lines
and I much prefer RPN.

 

[Franc Zabkar]

Have you ever tried the *unoffical* Win98SE Service Pack?
http://exuberant.ms11.net/98sesp.html

Looks interesting. I'm loathe to try it on my system, though, because
I've already applied a lot of the security patches. In any case the SP
doesn't provide a proper "undo" feature, so I'd only apply it to a
newly built machine. I did apply several of the tweaks, though. One
that had a dramatic effect was the addition of
"ConservativeSwapfileUsage=1" to the system.ini file. That alone was
worth the look.

Thanks.


- Franc Zabkar

 

[chrisv]

It took me about four hours to master RPN when I got my HP45,
and I never looked back. Since, I have had a difficlut time with
arethmetic calculators.

I know. It's like "Now how do I work one of these things again? Oh
yeah. 6737 minus 1498 equals. Yuck."

 

[keith]

I know. It's like "Now how do I work one of these things again? Oh
yeah. 6737 minus 1498 equals. Yuck."

....and then you find that you really meant 'minus 148', or wanted to do
other things with '6737' that you hadn't anticipated. Arethmetic
calculators are for the feeble-minded.

....got a good RPN calculator for Linux?

 

[Franc Zabkar]

It's irrelvant how TI handled the calculator evaluated the
expressions. It's the UI which is under discussion, here, and Robert
(and many others, including myself), prefer the RPN UI.

One's *preference* for certain UIs is not at issue here. I'm merely
claiming the obvious, namely that algebraic calculators have an
intuitive UI, whereas RPN does not. Take a kid who has never heard of
RPN, give him a HP calculator without a manual, and then see how long
he takes to work out what to do.

FWIW, I'm not averse to using non-intuitive interfaces, especially
when they provide powerful features. An example is the teco character
editor which I have used extensively in the past.

As for my preference for certain UIs, these days I prefer the Windows
GUI for many tasks, and the DOS CLI for others. My current calculator
is a scientific, statistical, hex/bin/oct, algebraic one. It cost me
$5 about 15 years ago.

Are you implying the more "mental gymnastics" are required to use an
RPN calculator? Because that would be false.

Algebraic calculators allow me to enter an expression exactly as it is
written. The machine decides the order of processing. OTOH, RPN
calculators require that I pre-process the expression and submit it in
a form which the machine can digest.

"Much more economical" sounds to me like a straw-man argument. "At
least as, and often more, economical" is the truth, not a "myth", in
every case I've compared the two methods.

In my day HP's marketing was making such claims. It seemed like their
mantra. Personally I've never experienced a difference of more than
about 10%. In any case, TI calculators had twice the memory, twice the
features, and cost half as much, so the UI did not affect my
purchasing decision.

A superior (or even only preferred) tool is quite often worth the
investment in how to use it. It's ludicrous to discount RPN
calculators because they may take a couple days getting used-to.

I would never discount them for that reason.

True, but many of us feel the RPN UI is vastly superior, especially
when used in conjunction with multi-line displays.

- Franc Zabkar

 

[Franc Zabkar]

So do I. It took me about four hours to master RPN when I got my HP45,
and I never looked back.

It took me less time to learn the Hangul alphabet. RPN was even
easier, but then I already understood a little Polish.

Since, I have had a difficlut time with
arethmetic calculators.

.... and spelling and grammar.

When the TIites saw how I did vector sums they shut up about the
superiority of TI's UI.

I just showed the HPites my TI59's magnetic card reader, its ROM
modules (eg stats, engineering), its printer interface, and its vastly
superior program memory. The HP97 didn't come close, not even at twice
the price. HP calculators had the glam (for reasons that escape me),
but TI had the grunt.


- Franc Zabkar

 

[Franc Zabkar]

Arethmetic calculators are for the feeble-minded.

I just knew it wouldn't take long for an RPN snob to show his true
colours.

BTW, the correct spelling is "arithmetic", and the correct terminology
is "algebraic" notation. *All* calculators are "arethmetic" (sic).


- Franc Zabkar

 

[Robert Redelmeier]

...got a good RPN calculator for Linux?

Google? http://www.linuxfocus.org/English/January2004/article319.shtml

-- Robert

 

[Felger Carbon]

...and then you find that you really meant 'minus 148', or wanted to do
other things with '6737' that you hadn't anticipated. Arethmetic
calculators are for the feeble-minded.

Feeble-minded? Did I hear my name called? Once upon a time I was
enormously infatuated by the computer language Forth, because I could
understand how it worked, because it could work "close to the metal",
and because it ran very well on my then-favorite CPU.

Alas, my feeble mind was never able to think in a Forth (RPN) fashion,
despite extensive efforts to re-educate myself (' a couple days of
re-education may be required'). And I was never able to make sense of
HP calculators, even though I recognized their obvious value. To this
day I don't use Forth (never have) and I use TI calculators (34s
scattered around the house).

Felger Carbon
feeble-minded non-RPN person