Re: Great SWT Program

On Nov 7, 8:05 am, b...@xxxxxxxxxxx (Bent C Dalager) wrote:
Well, then, instead of messing up your text files, they are instead
completely useless, since who wants to go to all the work of setting
them up again every single editing session?

If it's something you intend to use a lot, you'll persist it in your
.emacs file or somewhere else of your choosing.

Special bindings for editing a single file? Besides, what would they
actually be associated with? The file name I suppose?

M-x local-s<space><space> e insert-k<space><space><enter>

Ooo ... kay? That is definitely cryptic and unobvious and I don't
think any amount of man page reading over any amount of time is likely
to make it obvious that this is supposed to do, or easy to
remember. :P

There's obviously a whole damn language in there with very odd syntax
rules, semantically significant whitespace, and semantically
significant line noise. And it isn't even perl.

No, incremental search doesn't lose out to regular search; I never
claimed it did. Rather, search in general loses out to that newfangled
thing called a "scrollbar".

Even if that were the case, then you'd just use the scrollbar. I don't
see the conflict here.

Text-mode tools were the things under discussion, remember?

Indeed. I still haven't found that kitchen sink though . . .

It's hidden inside the embedded copy of nethack, I think. :P

I don't know why you'd want to make a graphical port of a graphical
program. It seems like a strangely masochistic sort of activity.

I'm talking about a graphical port of emacs, a text-mode application.

I don't know why, but you seem to be persistently using terms
differently from the rest of us.
buffer -> document
emacs -> graphical port of emacs
??? -> emacs
C-x -> Ctrl+x
M-x -> I'm not sure; alt x? Not that it matters.

[snip a bunch of what seems to be nonsense]

Really -- if you're familiar with the software it cuts what is
highlighted, but if you're not it cuts semi-randomly instead?

If you're familiar with the software, you won't expect it to cut
something just because it's highlighted.

You're not making sense. Highlighted means "this is the selection",
which means "this is what cut, copy, and the like should operate upon
if used right now". That's a fundamental part of the language. Using
it incorrectly is just as damaging and trouble-causing as using
English words incorrectly; if a user interface used "Save" on a menu
item to actually mean "Quit without saving" I'd be reaching for the
thing's uninstaller, and then for Thunderbird to send a nasty e-mail
to the developers! Nevermind that you could, in theory, "get used to"
the mislabeled menu items and learn what each one *really* did; it's
still a nasty prank to pull on the user! There is no excuse and no
justification for a UI that actively misleads the user by misusing the
language.

There was altogether too little fucking in the Matrix movies. A bit
more of that sort of thing might have gone some way towards rescuing
#s 2 and 3.

What??

You are denigrating some of the Matrix movies? *gasp*

Infidel! *points* Burn him! He blasphemes most foully!

You can configure emacs to highlight the selection if you really must
have this sort of visual feedback. This has no impact on what gets cut
or not - it's just decoration.

It's NOT "just decoration", it is a key piece of communication between
the software and the user. That you can configure it to tell the truth
doesn't absolve it of being, apparently BY DEFAULT, configured to lie
to the user, or even of being capable of being configured to mislead
the user at all.

I never did such a thing. I believe it is you who have become overly
indoctrinated with the "if it's inversed, then it's a selection"
paradigm. This doesn't hold in all software, and it certainly doesn't
hold in emacs.

Misusing visual highlighting to mean the wrong thing is just as bad as
using "Save" as the label of a menu item that does not actually update
the document file with the changes since the last save. (Using it for
a "Save As" may be acceptable. Using it for "Print", or "Exit", or
"Paste", or whatever certainly would not be.)

So C-w is not "cut" anymore? Now what is it instead?? :P

[snip long and complicated stuff]

See? This is completely unusable software. Nothing works in a stable
way; everything has weird special cases; there's no way to get
anything done without spending more of your time RTFM than actually
doing your fucking job. :P

And yet, despite the mindboggling complexity put in its way by the
very forces of the universe, it does.

"ex" does not match to the end of "extend" unless you completely
redefine the matching rules so that "e" in the query matches "n" in
the target and "x" in the query matches "d". :P

The second search matches the "ex" in "extend". Assuming
it selects to the end of the match, the selection stops with the "x"
in "extend". If it selects to the start of the match instead, then it
stops with the space before the word "extend". Neither of those cases
has the selection include the "tend" at the end. :P

C-w made it complete the word that was started with the "ex".

Complete WHAT word? There are a ton of dictionary words that start
with "ex", including, for example, "examine" as well as "extend". It
can't know which one of those you want to search for, nevermind if you
don't want any word that it has in its dictionary at all...

I also noticed that when I followed the exact instructions using my
emacs, it did exactly what I described it would do.

So you claim.

Savants of the average Joe stripe, most likely.

Bull***.

Nothing in your postings has convinced me of anything except that that
software is even harder to use than I'd originally suspected, and
requires some sort of oddball talent with memorization to use
effectively.

And as I pointed out at a much earlier point, anyone who does not have
the time or resources necessary to train in efficient tool use has a
considerable problem.

Training is unnecessary. People can use my preferred tools quite
efficiently with a minimum of training. That people cannot use your
preferred tools AT ALL without a considerable amount of training is a
problem with your tools, not a problem with people.

Sufficiently distinct to have no false positives and unique are the
same, the last time I checked,

False positives are not a problem. That are easily skipped in search
of the true positive.

Except that skipping them a) takes extra time (which becomes a problem
if there are very many) and b) screws with the behavior of backspace,
as I recall you claiming.

Incremental search doesn't help.

But neither does standard search. In fact, it is considerably /more/
helpless in this situation.

That's why normal people depend primarily on scrolling to navigate,
and search only when they know fairly exactly WHAT they're looking for
and don't know WHERE to find it. Usually this means with an unfamiliar
document that they know contains something. One of the most common
cases in my experience being web surfing: you google something and get
a hit. Click the link and get a huge page full of little chunks mostly
about other topics (e.g. a blog). Finding the thing you were searching
for has proved tricky: you've merely narrowed the location of the
needle down from a whole field of haystacks to one particular
haystack. No problem: Google has told you that a word you're looking
for *is* down there somewhere, and the browser's search will show you
the occurrences. (Also handy at times, for mid-size pages dense with
text, is Google's cache -- it will mark the occurrences of your search
terms, and you can just scroll looking for colored markers; with very
long pages the browser's in-page search is more efficient. Google's
cache may still be useful if the real page won't load, or doesn't have
anything to do with your query; the version Google saw presumably
does, and Google's servers are up if you got this far.)

It's not as fast as flicking the scroll wheel or otherwise using the
fucking mouse.

Then you do that.

OK. I claim victory. End of thread.

But by hypothesis "mon" was not enough. What now, wiseass?

C-s

Why not page up and save yourself some trouble? (Under the hypothesis
that occurrences of "mon" are too numerous to sift through the false
positives in a reasonable amount of time.)

Alt-Tab will take you to a different application altogether.

Not necessarily; it can take you to other windows that belong to the
same application, too.

Ah, yes, more confusion. And this is helpful . . . how?

There's no confusion here -- alt-tab switches among top-level windows;
it's quite simple. Alt-tab by itself switches between the two most
recently used; alt held down and multiple tabs hit will switch to
others. Alt-tab-tab to the third-most-recently-used, for instance. And
you can somewhat visually navigate the choices, as if you hold alt
down and hit at least one tab you'll see a popup with icons and
labels. So a well designed GUI app will let you make various task-
related windows visible from e.g. its Window menu, and alt-tab can
then be used to switch among them at need. There's also the Window
menu shortcuts (or just the Window menu) to switch to a specific one.
If it puts them inside a bigger window instead of at top level, the
Window menu should behave the same way, and you use ctrl-tab instead
of alt-tab to switch among them.

It is if that section takes the form of another open window on your
desktop.

And /only/ then. What a niche solution.

That is how a window system is normally used: functionality you want
to reach quickly is opened in some window or another and parked until
you need to switch to it. When you won't be using it, it gets out of
your way by being closed; when you're not currently using it, it gets
out of your way by being beneath the window you're working in. It's
still available quickly, and it's easier for novices to find (or
discover exists, even) via menus or browsing the open windows.

It is certainly possible for a GUI to be poorly designed. But a well-
designed GUI beats the socks off any text-mode interface it's ever
been my displeasure to try to use.

Assuming there is a "the" other window. Usually there isn't because
most application developers like to put their application in a single
main window.

If there are sub-windows instead you use ctrl-tab in place of alt-tab.
There's usually also a Window menu in either case; see above.

And the example was of someone who did /not/ follow the expected
pattern of usage.

That's not my fault, and it's not the GUI's fault. At least someone
unfamiliar enough with the GUI not to follow the normal pattern of
usage still gets their work done, albeit more slowly; someone
unfamiliar with the extremely complicated pattern of usage of emacs
(your latest examples read like stereo instructions!) will get nothing
at all accomplished.

It's called "you obviously never bothered to actually read any of it";
go back and reread the entire thread, all 1000+ articles of it, and do
not post to it again until you have.

Wow. I must be a fast reader.

No. I think you cheated and didn't reread everything. :P

well, save maybe the hypothesis that you're just stark, raving mad.

I don't know about stark.

What does it actually mean anyway?

It means you're overdue for today's dose of Thorazine, that's what it
means. :P

I read the descriptions of what you do with it, imagine attempting to
operate it myself and accomplish common user goals in it, and quickly
come to the conclusion that I'd have to juggle all sorts of mental
state about both the document and the current state of the software
that I would not have to do to do the same work with Notepad, Word, or
anything similar.

This juggling you refer to is either non-existent or else so trivial
that it's not noticable.

Trivial to a memorization whiz, maybe. Not to a normal human being.

I never suggested doing anything of the sort; only editing *tabular
data* using spread*** software.

And you keep insisting that properly formatted source code is tabular
data, so the connection is obvious.

No, I do not; you're lying and putting words in my mouth here, which
is very rude. Don't do it again.

For proper formatting of source code you use two things: tab and your
IDE's auto-indent.

I have already agreed that training to use emacs is non-trivial.

That's putting it mildly. It obviously takes years to master, to judge
by the arcane stereo-instruction stuff you've been posting lately.

It's time to stop.

Then by all means please do shut up.

/Damn/, I'm fast.

*** you, liar.

NO IT DOES NOT! How many fucking times must I repeat myself before it
sinks in?! Hell, YOU try finding anything in a document you don't
already know backwards and forwards that way.

I do so all the time.

Bull***.

If I get too many false positives, I will tend to choose other means
of navigation.

OK, then. I claim victory. End of thread.

But what about where you enter the search query? It's not in a
separate window, so the cursor in the same window as the document has
to move to wherever this is. Unless of course you type the query in
blind.

The search query appears in the mini-buffer, but the cursor remains
with the search hits.

Then how the hell are you supposed to edit the search query?! I
thought that was the big advantage -- editing the search query on the
fly. Now you're telling me that the insertion point will be in the
document, which just means you can edit the document before hitting
next match. That's the Windows meaning for "incremental search"
discussed in contrast to the emacs one earlier. Make up your freaking
mind!

None are, in my experience, "designed from the bottom up with a mouse
fetish"; just designed assuming that modern video hardware and a mouse
will be available along with the keyboard. And I find them quite
efficient TYVM.

Probably only for lack of having trained in something better.

Bull***! Keyboard only can't possibly be superior to multiple input
options, and a thin drool of visual feedback can't possibly be
superior to rich and detailed visual feedback. Face it; the software
you're describing is a dinosaur. You just don't want to admit it.

HOW??? You claim that everything just extends the existing selection.

There are a number of ways to define the start point of a
selection. The most direct is C-<space>.

Ugh ... ugly.

Whenever you start an incremental search, the start-point (a.k.a the
mark) is reset to the start of that search.

So much for selecting using search to find the second endpoint, then,
if doing so will disturb the position of the first endpoint.

Hah. I call your puny mathematics and raise you fiteen years of
experience in the field.

Mathematics is unassailable. If you claim to have "experience in the
field" that contradicts it then I claim you're lying. Same as if you
claimed that your experience was that if you had five apples in a
basket, and threw in three more, then counted them, there tended to be
seven. Either you're lying or you're bad at counting. :P

[snip assorted BS]

I certainly haven't been bothered to write you an incremental search
training manual. Do your own legwork if you want to actually learn how
to do it.

The more you write, the more convinced I am that I don't want to go
near it ever again. I won't even touch it with a ten-foot SCSI cable
now. Not emacs, anyway.

there, word-for-word, but NOT knowing even approximately where it is.

Except you don't because if you did, then goto-line would have been
just as useful.

Note the word "approximately". Particularly if what you know is that
it's about 1/3 of the way through the document, not that it's within a
page or so of line number xxyyzz.

Seeing as I use incremental search all the time but have yet to
memorize one single file of source code, I can only dismiss your claim
as uninformed FUD.

Seeing as what you claim simply isn't possible I can only dismiss your
claim as an out-and-out lie.

And, of course, we all know that all JBuilder is ever used for is
reading source code - never to write it.

I wouldn't know, not that any of this is relevant; JBuilder was never
under discussion. Changing the subject; how cheap.

It would be when you're using inefficient tools for the task.

But I'm not.

Of course, if I had never had the fortune to come across
emacs I never would have known what I was missing and might have been
in your camp on this issue.

More evidence that it's dangerously addictive. Like cocaine it
apparently makes people feel like they're Superman -- even while
actually slowly weakening them.

That may be it of course - I may just be blessed by lady luck, that's
why I find incremental search so excellent.

More likely you're just plain nuts.

Indeed. Most of the time, I don't actually know /where/, exactly,
something is in the document (but you apparantly do? Do you memorize
all your documents?)

I normally have a fairly good idea of where, proportionally, each
thing is within the document, yes.

and so incremental search lets me go there from
just know approximately what is being said there.

I've just explained for the umpteenth time why that doesn't work when
it's really only "approximately" known.

I would have thought this realisation rather puts the final nail in
the coffin for one of your pet theories.

No, it only indicates that you're lying. Your stories are inconsistent
from one day to the next.

It's very close to it; a fraction of a second versus your several
seconds of fiddling with the keyboard to get to the same destination.

I touch type. I don't take "several seconds" to type a simple search
term.

We're talking a complex query, plus a bunch of weird extra chords, not
simply typing "ant" into a box or something.

But, again, this helps to highlight why you don't think standard GUIs
are inefficient: you work pretty slow to begin with and so the extra
lost efficiency isn't really noticable to you.

Bull***; I work plenty fast enough to begin with. I'm slowed down by
interfaces that make me reach for the help every five minutes, or do a
ton of typing to get anything done when it takes less time to left-
click once than it does to type four or five words and some control
sequences.

Unfortunately, you've suggested that this should be one's primary
means of long-range navigation.

Which, indeed, it is.

For you and other memorization savants perhaps.

[insult deleted]

None of the nasty things that you have said or implied about me are at
all true.

There is such a thing as information overload. Some things, you simply
do not /need/ the GUI to be shouting out at you, because you /already
know them/.

You find GUIs to be "shouting"? How strange. Have you discussed this
with your therapist?

Of course, this conveys only as much
information as a 100-pixel-high scrollbar, and in a manner that still
requires a bit more mental heavy lifting to interpret besides.

Yes, I understand some people find percentages to be difficult.

They don't have to! It's still far faster and less mentally-burdensome
to visually assess a scrollbar than it is to compute mental arithmetic
-- even when that arithmetic is fifth-grade easy.

The scrollbar can certainly be useful in a few niche cases, but
incremental search fast becomes the heavy lifter wrt navigation once
you've learned how to use it.

Bull*** once again, except perhaps for memorization savants.

[insults deleted]

None of the nasty things that you have said or implied about me are at
all true.

Pressing <enter> terminated the search. When the search is terminated,
of course, the hits are no longer highlighted.

So much for no easy deselection.

What you are referring to was not a "selection".

Whatever. Call it what you will.

Then I can select that line and type M-x eval-region to have it run. I
can do this with arbitralily long sequences of commands, and entire
scripts.

Wacky. Though that didn't look like any variant of shell I've ever
seen.

That's not an "extra" step, it's a necessary one.

For you perhaps. Not for me.

Never wanted to paste, modify, *then* submit?

Are you saying that
if you paste into an emacs search prompt, it immediately does the
search and won't let you paste something in and edit it and *then*
search?

No.

Then an enter keypress is necessary to submit the search after all.

[insult deleted]

None of the nasty things that you have said or implied about me are at
all true.

That's where the tab key, the space key, and the enter key come in.

/If/ you want it to take forever, yes.

Bull***. *** off.

But they will. Or are you such a language chauvinist that you deny
other nations the right to localize their fuel gauges?

WE'RE NOT TALKING LOCALIZATION, FUCKWAD, WE'RE TALKING TWO SUPPOSEDLY
ENGLISH-LANGUAGE LOCALIZATIONS ONE OF WHICH HAPPENS TO LIE TO THE
USER.

Or maybe with the steering wheel on the right side of the car, eh?

NOT A VALID COMPARISON.

No. I am sorry, but the visible highlight and the thing that actually
gets cut being completely unrelated is completely unnatural and
abnormal.

/If/ your indoctrination comes from Windows, yes.

NO. ALWAYS. INHERENTLY ABNORMAL.

We're basically talking about the user interface *lying*
here, for Chrissakes.

No, we're talking about learning the interface.

One that lies? No thanks.

I have repeatedly agreed that GUI /could/, in a perfect universe, be
as convenient and efficient as I find emacs to be. But, again, in
practice they turn out not to be.

In practice they turn out to be quite good a lot of the time, and
moreover, actually usable by people that are not memorization savants
too.

What evidence would you require in order to prove that I do, indeed,
believe that?

Nothing will convince me of much of the BS you've been posting lately,
simply because it appears to violate the laws of physics more often
than not.

Then you do not touch type. Touch typing is all about holding the
hands still while only moving the fingers.

How awkward. You'll propably strain something, too.

I do not find that using grep is difficult in the least, and there are
often hours between each time I use it.

Memorization savant. Your mileage will vary.

How would one even discover this key binding though? I don't see how
discovering it would pop up naturally in the course of ordinary
workflow or even using the help to find out what nutty thing they used
for paste instead of ctrl+V. Short of reading the entire manual from
cover to cover, or finding out from a tutor or some other such live
person, I just don't see things like this being stumbled onto.

Actually reading the manual does seem like a decent sort of idea of
course.

Yes; for a phone-book memorizer type, it's probably a breeze. :P

Of course, because of this prominence, it's not very difficult to
learn to use C-g for the same thing.

No, it's just contrary to every other user interface in the world,
contrary to every shred of common sense, and even contrary to the very
name of the key "escape" just for good measure.

No, actually it isn't. Someone less experienced than I might actually
read that and believe you. :P

If it is not, then you have just codemned the entire GUI standard that
you are spending so much time championing. If Files->Open File... is
/not/ easy to find for a mouse-fiddler, then Notepad is equally
byzantine to the beginner.

??? Non-sequitur. We were discussing learning all of those arcane
things in emacs, not using Windows and Notepad.

And yet, despite the utter impossibility, I keep posting to this
thread.

Memorization savant, or just plain liar; at this point I hardly care
which.

I find emacs to be entirely consistent. Highlights, search, paste and
C-g all work the same every time I try them.

And C-w, C-y, M-y, undo(!), and backspace(!!) don't. And none of them
are consistent with *other apps*. Any other apps. At all.

Plus the need to track a lot of complex state mentally,

Complex to some perhaps.

Perhaps not to memorization savants and other very abnormal brains.

memorize the whole document being worked on so that you can use search
for long-range navigation, and suchlike.

The only person in the world who thinks he needs to memorize documents
in order to work on them, is you.

I work on documents every day without exact memorization. YOU are the
one who implicitly claims to memorize whole documents exactly, by
claiming to primarily navigate using methods that are not very
efficient unless you have.

But analogies are your bread and butter! The very core of your being!

Only ones that actually make sense.

One or two seconds, but it's still a lot of work compared to a couple
of clicks, and I *really* don't see this one being stumbled upon,
given that an exact sequence of twelve whole keystrokes is needed.

I don't know why you would want to stumble upon it.

Neither do I, but if it were a feature of something other than emacs,
then I might want to know that it exists. :P

Ah, you mean, "only appear when the software thinks you need them" -
which turns out to mostly show them when you don't want them, or hide
them when you do need them.

I never claimed that GUI apps could not be badly designed. That's
hardly a valid condemnation of the whole idea of a GUI app, though.

[insults and BS deleted]

None of the nasty things that you have said or implied about me are at
all true.

Submitting a command generally involves hitting enter. If you can
point out where I have said otherwise I will address it.

At least one command was submitted some other way; I just can't recall
which.

.

Quantcast