Re: Wholesome choice?

In <gkarhc$9lj$1@xxxxxxxxxxxxxxxx> "Keith F. Lynch" <kfl@xxxxxxxxxxxxxx> writes:

Ben Yalow <ybmcu@xxxxxxxxx> wrote:
"Keith F. Lynch" <kfl@xxxxxxxxxxxxxx> writes:
What would it mean for me to be on the Internet? If I can send and
receive email, post to and read Usenet, and produce and view web
pages, I'd say I'm there. Panix provides all of that, and deals with
the technical, security, and file-backup issues so I don't have to.

You can send and receive a limited fraction of email (due to your
hardware restrictions).

I can send and receive anything that will fit in my Panix quota.
However, I've chosen to block all emails longer than a medium-sized
novel, since in my experience all such emails that are directed to
me are spams or viruses. Nobody writes hundred-page letters to me,
online or off.

The rest of us get more than just text. We get word processed documents,
and spreadsheets, and PDFs, and lots of other formats that you can't
handle.

People don't send you that stuff because they know you've cut yourself off
from anything more complicated than your hardware/software can handle.

You can view some pieces of some web pages, although not see much
of their content.

I can see all of their content, including source code, except for the
non-text portions, which are usually unimportant.

For you, they're unimportant. For the rest of the world -- we want to see
maps, or pictures in news stories, or listen to a piece of music, or look
at the video of the most recent SMOFcon presentations by the Worldcon
bidders, or read the comics, or .... For everybody but you, those kinds
of things are important.

You can't use any of the P2P services, send files to/from your local
machine to another machine on the network, etc.

I can and do upload and download files from my local machine
using Kermit. I see little reason to do so except keeping local
archival copies.

My impression of what you call Internet access is that it's the
ability to look at colorful pictures of the net rather than to
actually *be* there. If I'm at a library terminal and I want to
check my email, read Usenet, update a web page, or do anything else
whatsoever except look at a web page, chances are it doesn't have ssh
and it blocks telnet:// URLs. (Yes, I'm aware that Panix eventually
installed https://shell access, to work around this.)

Of course, that's because you cut yourself off from things. The rest of
us want that stuff, even if you don't.

I've had a machine on the Internet for decades, as have most of the
people in this newsgroup.

Decades? Twenty years? Since 1989? Non-shell accounts were rare and
expensive then, if they existed at all. (Note that this was years
before the Web.) I suspect most people here didn't have *any* net
access then, except perhaps at work or at school. And that those who
did had either shell accounts or UUCP-only accounts. In fact, I've
found no mention anywhere online of SLIP or PPP accounts before 1992.

Google ybmcu@xxxxxxxxxxxxxxxx You'll see things from before 1989.

Other people have computers, not just a terminal.

So do I. Several of them. And I prefer not to pay good money to
place them where any sufficiently clever and malicious person can
read and alter all my files. Especially since it wouldn't gain me
much of anything.

If your opertaing system had reasonable security, you wou;dn't worry about
that.

Using Emacs on a terminal, please decide if the continent outline
colors on the current Australia ad on the back cover of the current
Worldcon progress report translate well to B&W ...

That's not a text-editing task. Its inability to do that task does
not make it "a very simple text editor." Anyhow, "Turing complete" is
obviously limited by one's input and output devices. For instance no
software of yours, no matter how sophisticated, can vacuum your floor
unless you have vacuum-cleaning peripherals.

Right. And you've limited, by your hardware choices, Emacs to be even
less than it would be in an X environment. It still won't do layout, but
it does a lot more than your version does.

It's the sort of things that people working on cons need to do.

How did people do it before computers?

Expensive typesetters. And expensive color separations, or black/white
layout. But typewritten publications just aren't good enough any more.
People who get those tend to decide that the convention that's sending
them out is hopelessly obsolete, and won't go. So the convention dies.

(And, BTW, pdftotext won't work on those PRs -- it scrambles the
orders of the words.)

I suggest you remake those PRs so that it doesn't. I'm sure I'm not
the only fan who relies on pdftotext to read PDFs. Having the words
in the right order is far more important than having the continents
display correctly in black and white. (Is there anyone still using a
black and white graphical momitor screen?)

As before, I'm willing to concede that you might not be the only person to
do something. There might well be a second person.

But it's not the PDF that's broken -- it's pdftotext. The PDF is valid,
and displays properly. It's only when a broken piece of software that
doesn't really understand how to do the conversion tries to read it that
it breaks.

Any of dozens of readers can handle it properly. The file will display
just as intended.

That would be a better argument if that underlying philosophy were
not profoundly flawed on numerous levels.

How so? Both are, of course, based on Unix-like platforms (MacOS X
actually exposes the Unix to end-users).

I could spend all day answering that, but I have over 100 other rasff
messages marked as needing replies. (I spent last night at a PRSFS
meeting instead of trying frantically to cut down the backlog here.)
So I'll mention just a few highlights:

* The idea that a GUI interface is somehow more intuitive than a CLI
interface. To do any but the most trivial tasks on a computer, you
need an interface with the power of a language. You have to *learn*
a language -- there's no intuitive way to edit a file, read a
newsgroup, block spam, update a spread***, etc., any more than
there are intuitive names for common objects rather than English (or
French, or Japanese) words for them. Words have to be learned, and
grammar has to be learned to be able to say anything meaningful
with them. Typing at a command line is more "linguistic" than
pointing and clicking with a mouse, just as speaking is more
"linguistic" than pointing with your finger and grunting.

Except that people, in fact, pick up those interfaces when they have no
computer literacy whatsoever, and can use them easily. They can't always
use them as well as power users, but they get lots of work done, and they
can do so with far less training than it takes to get somebody to use a
Unix shell.


* The idea that a GUI interface is faster than a CLI interface. This
is true only is for people who can't touch-type, and touch-typing is
so easily learned that it's senseless not to do so unless you plan
to use computers very little over your lifetime.

I don't see what touch typing has to do with anything (although, since
I've only been using computers for 40+ years without touch typing, I guess
I wouldn't).

* The anti-concept of "opening" a file. It means one thing to view
a text file on a screen, another thing to view a picture, another
thing to run executable code, another thing to listen to a sound
file, etc. By conflating these very different concepts under the
word "open," Windows has opened the door to malware. People who
want to view the contents of an email message end up inadvertently
running malicious program code that's embedded in it instead, as
they don't have any concept of doing one but not the other -- at
least not a concept they can convey to the computer.

That particular door began to open when the concept of strict separation
of program and data went away. Which happened at about the time the
stored program computer became the standard -- probably late 50s or so. I
don't think Microsoft is responsible for architecture changes in the 50s.

* The idea that bugs and crashes are inevitable and unavoidable,
and that it's reasonable to ask people to pay for upgrades to
Microsoft's grossly defective products which simply correct some
of these inexcusable serious flaws. I'll note that these aren't
even new and original kinds of bugs, but simple things like buffer
overflows or like forgetting that leap years exist.

Bugs and crashes *are* inevitable. I haven't seen a system without one in
those 40+ years. People who think they have perfect software -- with no
bugs, and which never crashes -- are deluded.

* The idea that there is a perpetual arms race between "hackers" and
victims, and there's no sure and simple way of keeping bad guys
from viewing and altering all your files, or of keeping them from
remotely reprogramming your computer to send spams and viruses.
And that anyone with anything but the latest version of Windows
with the latest patches is especially at risk.

Having been on both sides of that race, going back to well before
Microsoft ever existed, I can say that competition between bug finders and
bug fixers isn't a Microsoft issue. Just talk to the systems staff at
lots of places about what April 1 was like.

* The idea that you don't really own or control your own computer, but
that Redmond, or even content providers, are free to monitor your
usage, block some kinds of usage (such as skipping FBI warnings on
DVDs, saving the text of websites you view, or returning to the
previous web page you viewed), and updating and installing software
on your machine without your explicit permission (e.g. Sony
rootkits, Microsoft's "mandatory" upgrades), or remotely disabling
your machine if they mistakenly think you may have an unauthorized
copy of something. I contrast this with the well-tested idea that
software on a production machine should *never* be updated or
altered without a very good reason and thorough testing.

Feel free to sign any other license you choose. You do control your own
computer -- and you pick what you want it to do, and trade off your
desires with other people's. If you want something they provide, figure
out how to get them to sell/lease it to you.

* GUI operations are less suitable for building macros out of than
are keyboard operations. It's easy to construct and learn a macro
language to automate repeatedly typing very similar things that
vary in some easy-to-describe way, perhaps even with conditional
branching and looping. It's much harder to construct or learn a
macro language -- if possible at all -- for doing the same thing
with mouse moves. So I often see GUI users doing near-identical
mouse moves, over and over again, once every second or two, for
minutes or even hours. Like knitting, which it resembles, users
may find it enjoyable and relaxing (at least until they destroy
their wrists), but it is not an efficient use of time.

Macros are a minor special case of the scripting language. And you have
much more powerful, and intuitive scripting languages for non-Unix
products.

* Many concepts that have been built into Unix or Emacs for decades
are still missing from Windows. It's nice that cut-and-paste are
always the same control characters, but what if you want to paste
something from two or three cuts ago? Windows seems to be totally
lacking anything like the Emacs kill ring.

* The idea that when computers get enormously faster and have
enormously more memory and disk space, that they should take about
the same amount of time to boot up or to so other tasks, and that
the memory and the disks should be mostly full, even though the
software doesn't actually do much more than years ago other than
look and sound a little ritzier.

Computers do *much* more than they ever could before. You just don't want
yours to do so.

And that's OK -- but the rest of the world disagrees with those
preferences.

--
Keith F. Lynch - http://keithlynch.net/
Please see http://keithlynch.net/email.html before emailing me.

Ben
--
Ben Yalow ybmcu@xxxxxxxxx
Not speaking for anybody
.