> On 2012-05-10 09:24:26 +0000, Jumpin' Jehoshaphat said:
>> Classic paranoia. Where is Mozilla's alleged financial interest in
>> making people buy the latest and greatest CPU, mobo, and GPU?
>> Kickbacks from Intel, AMD, ATI, Asus, and nVidia? What's your evidence
>> in support of your outlandish conspiracy theory, Madhu?
> You know, there's some kind of horror movie script idea here, where it
> turns out that some terrible virus of the mind is spreading over the
> internet, the victims getting trapped in a loop where they repeat a
> small number of stylised phrases, eventually starving to death, to be
> found, wormy and decaying, over their keyboards. How will our hero stop
> the spread of the plague? Stealing a nuke and destroying the internet
> via EMP seems the only option.
What does that have to do with Lisp, Bradshaw?
> Oh: who do you think fund Mozilla? Where does their income come from?
> Gosh, what a susprise!
I very much doubt it's conditioned on intentionally bloating their software, Bradshaw.
>> Why, drag and drop. You have icons on the desktop for all possible
>> characters,
>> and then you just drag them one by one to the script icon.
> You think that's a joke, but have you seen any of the maths editors
> people use? A lot of them are just like that.
Probably because no standard store-bought keyboard has the alphabet, the digits, the commonplace punctuation, the Greek alphabet, an integral sign, dot-in-circle, aleph, square-root sign, del, grad, +/-, squiggle-equals, not-equals, intersection, union, element-of, ... symbols and a bunch of extra shift keys for overbar and the like.
So it's either toolbar buttons, an emacsesque bestiary of impossible-to-memorize-them-all control-key combinations most of which have no mnemonic connection with what they insert, or typing in verbose markup like TeX code. (TeX editors sometimes give you the choice between the toolbar buttons and manually typing in TeX markup. Useful if you memorize the most commonly needed markup but can then just dig in the toolbar drop-downs for less-often-used symbols instead of having to task switch to a cheat-sheet, read some possibly-long string, switch back, and type it, or copy and paste it.)
>> As mentioned in another post, this can be useful for data entry on an
>> old, cruddy, lagged system or a laggy oldfashioned telnet connection.
> You didn't understand my point at all.
That is incorrect. I just didn't agree with it.
> The bottleneck is often in human sensory input.
Good FPS players prove you wrong.
> But it helps with those 'lagged systems' too. Maybe in your ideal world
> computers never go slow, but in reality it isn't the case. Even most
> modern systems can go thrashing if something goes awry, and in my
> experience it's better to use text console at that time.
In my experience it's better to fix whatever the problem is (e.g. nice a process that started hogging the CPU, or restart it if it's not supposed to occasionally do that, or even debug it and recompile it) than to just try to carry on regardless.
> Also, if you just happen to need to work with server in California from
> Europe you have ~300 ms roundtrip for each packet just due to the speed
> of the light.
Input to your locally-running browser will not be affected by this. Stuff typed into a web form will echo promptly and mouse movement won't be affected. It may just take a while for the response to get back to you after you click "submit" or a link or whatever.
> No matter how fast your network is and how fast your
> computer it. In command prompt it is noticeable, but not problematic at
> all. With GUI it might be more of an issue.
You're hypothesizing somehow puppetting a GUI that's running on *the remote machine* rather than using a local client or web browser to communicate with the server. That would be ... weird. Mostly you use a web interface to a remote machine these days. If for some reason you can't (usually because you're the admin trying to remote-fix the machine after the httpd died) you grit your teeth, ssh into it, and use a command prompt. So when there's a GUI its mouse/keyboard feedback loop is local rather than stretching over the network.
>> For command entry it's just asking for trouble. Enter a directory and
>> select the third file to be deleted sight-unseen? Er, I don't think so.
>> What if the third file isn't what you assumed it would be? You could be
>> misremembering, or other users (if any) may have added or removed or
>> renamed files in there.
> You're strawmanning shit out of this.
Classic unsubstantiated and erroneous claims.
> Not all commands are deleting files.
That was an example.
> Shell commands reference file name, but not position.
The example was using a program more like dired rather than a command prompt.
> I can type `cd .. <enter> ls` without checking that cd was correct. In
> the worst case I'd have to type it again. It isn't like cd will
> accidentally turn to `rm -rf` a.
Actually, if there's line noise you never know. ;)
> When I'm typing commands I'm very aware of what they do. So I need
> different levels of attention, and I can do cd/ls stuff much faster than
> rm -rf. And that's actually great because maybe 90% of commands are like
> cd/ls but not rm.
You could still find yourself wasting time, at the very least, if you wind up taking a wrong turn somewhere while overdriving your headlights, even if you manage not to crash.
> rm -rf frobla.txt <look that it's correct> <enter>
And if you're in the wrong directory and there are frobla.txts you don't want to delete as well as ones you do, in different parts of the filesystem ...
> In reality it's more like
> 1. rm -rf fro<TAB>
> 2. look whether it was completed, there is a list of options or ...
> 3. when everything looks OK press enter.
The GUI equivalent would be to click a file and hit del. You'd have to look to see that it had acknowledged your click and selected the file before hitting del, but that's the same as your having to look before hitting enter.
Where you went off the rails is comparing keystrokes to mouse clicks. Everything but the "rm" and "enter" parts correspond to one single mouse click to select the file. The del key supplies those parts (and gives a confirmation prompt in any decent GUI file manager to boot, as one added level of safety).
If you're deleting all files of that name under some directory, if you're using a recent Windoze you'd type frobla into the search box, wait for it to find everything, then hit ctrl+A DEL. Again one wait.
> Please learn how to use shell before criticizing it.
Classic erroneous presupposition.
>>> No, you're just using straw man shamelessly here.
>> Don't be ridiculous. I'm just pointing out an inconsistency in the
>> pro-keyboard-only position.
> And who holds that pro-keyboard-only position?
You, Madhu, and Kylheku, at the very least.
>> As do I, for the most part; also for some actions that involve selecting
>> from large numbers of options (you can acquire from a 2D array much
>> faster than from a 1D list and much faster with a random-access-ish
>> selection mechanism than with an arrow-your-way-around mechanism, so
>> unless the items have names you can select via short unique prefixes...)
> If information is textual chances are that incremental search would have
> worked better.
Information is often not textual, or not solely textual. Textual information often lacks short unique prefixes on each item, or one doesn't quickly remember the one desired. It's faster to recognize a particular icon, if that's unique to your target, than to recognize a particular short snippet of text, even if that's also unique, since the latter is made of nonunique individual letters.
>> and actions in a graphics-manipulation context (Photoshop, mainly).
> Except that graphics professionals use pen tablets. You know why?
> They don't require feedback from screen. Pen tablets have absolute
> positioning unlike mouses which have relative positioning. This means
> that pen tablet reproduces your motion exactly, you don't need to follow
> cursor to draw a curve you want.
That's when drawing freehand rather than doing something else, such as dragging pre-created image fragments around and assembling them in some manner or applying filters, etc.
There's more sorts of graphics work than you seem to be assuming.
Also, pen tablets are expensive. The last time I saw any in a store the cheapest was 4-5x the cost of a decent mouse. Not everyone is going to spring for something like that. Particularly among those who work with graphics a fair bit but *not* in a capacity where they get paid to do so, or at least to do so very often.
>>> Writing a novel is just so much typing; so automate it!
>> That's data entry, not command inputs.
> You have a problem pal!
Classic unsubstantiated and erroneous claim.
> Code = Data !!!
And indeed typing in code is a form of data entry. But directly issuing single-gesture (control-key, mouse-click) commands to be executed immediately is not really comparable when it comes to the whole "is it a good idea to do it blind/while the interface display lags behind?" issue.
> Huh? At school level people work with concrete numbers, but on more > advanced level there are barely any numbers in formulas.
> I don't know much above physics above school level, but I've got M.Sc. > applied math degree without being particularly good at arithmetics.
Well yes. It turns out that in physics numbers do matter. A very important skill (which people are explicitly taught) is to be able to have a feeling for whether something is "reasonable" which can mean various things, but often means "is it about the right size?" and that requires, among other things, the ability to get numerical estimates of stuff very quickly. And although you are almost never doing literal arithmetic (you're typically just trying to get an answer which is right to an order of magnitude and some kind of notion of how rough the answer is), you really can't do this without being pretty competent at arithmetic, it turns out.
If you want a really good example of this there are several in Feynman's autobiographical books. Obviously he was better than most.
It's very tempting to say "oh, none of this matters now, you just use a computer" but that turns out not to be right for reasons I'm too tired to explain now, but are, I think, well-known.
>> I know this because at a slightly different level it's what happened to
>> me: I was lazy, didn't practice enough boring mathematical methods
>> stuff, and that lack of fluency came back and bit me later.
> Math =/= arithmetics.
I am aware of the difference. That's why I said "at a slightly different level".
> In reality it's more like
> 1. rm -rf fro<TAB>
> 2. look whether it was completed, there is a list of options or ...
> 3. when everything looks OK press enter.
And even that is not what people do: they have typed and executed the whole command before they have time to think. I do this (as I've discovered), and the interesting thing is that, after a while, although you do execute things you should not have, you (almost) never do it in a state where it matters.
On 2012-05-10 12:53:08 +0000, Jumpin' Jehoshaphat said:
> So it's either toolbar buttons, an emacsesque bestiary of > impossible-to-memorize-them-all control-key combinations most of which > have no mnemonic connection with what they insert, or typing in verbose > markup like TeX code. (TeX editors sometimes give you the choice > between the toolbar buttons and manually typing in TeX markup. Useful > if you memorize the most commonly needed markup but can then just dig > in the toolbar drop-downs for less-often-used symbols instead of having > to task switch to a cheat-sheet, read some possibly-long string, switch > back, and type it, or copy and paste it.)
I'd guess that you've never typed any significant amount of maths. I have, and TeX is just enormously faster than any kind of pick-and-drop thing. Even now, nearly 30 years after doing any large amount of stuff, and something like 20 years after really even using TeX much, I can look at even quite complex expressions and type them, and very often not have to fix anything. It was designed by someoene who typed maths, and that really shows (troff may be as good, I just never learnt more than you needed to make a man page).
On 2012-05-10 13:45:41 +0000, Jumpin' Jehoshaphat said:
> And indeed typing in code is a form of data entry. But directly issuing > single-gesture (control-key, mouse-click) commands to be executed > immediately is not really comparable when it comes to the whole "is it > a good idea to do it blind/while the interface display lags behind?" > issue.
See my earlier message. I make my living typing commands at machines, in a context where mistakes can be expensive. I type blind all the time, and so do other people I know (I only discovered this recently by observing what I do and then asking other people doing the same job).
> On 2012-05-10 11:32:37 +0000, Alex Mizrahi said:
>> In reality it's more like
>> 1. rm -rf fro<TAB>
>> 2. look whether it was completed, there is a list of options or ...
>> 3. when everything looks OK press enter.
> And even that is not what people do: they have typed and executed the
> whole command before they have time to think. I do this (as I've
> discovered), and the interesting thing is that, after a while, although
> you do execute things you should not have, you (almost) never do it in a
> state where it matters.
Perhaps, but when it's sometimes "rm -rf", "almost never" isn't going to be good enough.
> On 2012-05-10 13:45:41 +0000, Jumpin' Jehoshaphat said:
>> And indeed typing in code is a form of data entry. But directly
>> issuing single-gesture (control-key, mouse-click) commands to be
>> executed immediately is not really comparable when it comes to the
>> whole "is it a good idea to do it blind/while the interface display
>> lags behind?" issue.
> See my earlier message. I make my living typing commands at machines, in
> a context where mistakes can be expensive. I type blind all the time,
Why? Are you using old, obsolete, slow equipment or an old, obsolete, slow network protocol that makes the input/feedback loop take a round trip to the server and not just complete commands?
> On 2012-05-10 12:53:08 +0000, Jumpin' Jehoshaphat said:
>> So it's either toolbar buttons, an emacsesque bestiary of
>> impossible-to-memorize-them-all control-key combinations most of which
>> have no mnemonic connection with what they insert, or typing in
>> verbose markup like TeX code. (TeX editors sometimes give you the
>> choice between the toolbar buttons and manually typing in TeX markup.
>> Useful if you memorize the most commonly needed markup but can then
>> just dig in the toolbar drop-downs for less-often-used symbols instead
>> of having to task switch to a cheat-sheet, read some possibly-long
>> string, switch back, and type it, or copy and paste it.)
> I'd guess that you've never typed any significant amount of maths.
Then you'd guess wrong.
> I have, and TeX is just enormously faster than any kind of pick-and-drop
> thing.
It is if there's a convenient way to get at infrequently-used stuff you haven't got memorized. Ideally, a visual way so you can find the symbol you want by sight and click it. And then if you're starting to use that one a lot, or a lot in that particular document, you note what it pasted when you did that and type it the next time.
But if you have to task switch to some cheat sheet (likely a giant bloated pdf file open in giant bloated Acrobat Reader, bletch) every time you want to find the code for some symbol you don't remember it will be much slower than if you can open some palette and click it in that when you don't remember the code to type.
On 2012-05-10, Jumpin' Jehoshaphat <jj1058672...@gmail.com> wrote:
> On 10/05/2012 4:46 AM, Alex Mizrahi wrote:
>>> ALL actions require feedback! I don't want to blindly issue commands to
>>> my computer while not even looking to see if it's gone into Sorceror's
>>> Apprentice mode deleting the wrong files or something.
>> The difference is that with mouse you need feedback for each action, you
>> need feedback immediately as it blocks you.
>> But with keyboards you can issue multiple keypresses at once and look
>> for results later. So this check can be postponed.
> As mentioned in another post, this can be useful for data entry on an > old, cruddy, lagged system or a laggy oldfashioned telnet connection.
You missed the part about there being lag that cannot be eliminated.
Even if your computer's interactive response time is a nanosecond, there is lag in the brain.
Anything that is blocked on the human having to make a conscious decision is
looking at a 500 millisecond response (Libet's half second delay).
Training to do things unconsciously brings that down, but not to zero.
> On 2012-05-10, Jumpin' Jehoshaphat<jj1058672...@gmail.com> wrote:
>> On 10/05/2012 4:46 AM, Alex Mizrahi wrote:
>>>> ALL actions require feedback! I don't want to blindly issue commands to
>>>> my computer while not even looking to see if it's gone into Sorceror's
>>>> Apprentice mode deleting the wrong files or something.
>>> The difference is that with mouse you need feedback for each action, you
>>> need feedback immediately as it blocks you.
>>> But with keyboards you can issue multiple keypresses at once and look
>>> for results later. So this check can be postponed.
>> As mentioned in another post, this can be useful for data entry on an
>> old, cruddy, lagged system or a laggy oldfashioned telnet connection.
> You missed the part about there being lag that cannot be eliminated.
No, *you* missed the part about the proficient FPS players whose mouse use is not hampered by that lag.
On 2012-05-10, Alex Mizrahi <alex.mizr...@gmail.com> wrote:
>> ALL actions require feedback! I don't want to blindly issue commands to
>> my computer while not even looking to see if it's gone into Sorceror's
>> Apprentice mode deleting the wrong files or something.
> The difference is that with mouse you need feedback for each action, you > need feedback immediately as it blocks you.
I'm afraid we have been, as I suspected, wasting time responding to Seamus
MacRae, a.k.a. Oxide Scrubber, Series Expansion, etc.
The wacky pseudonym, low intelligence, and overall style give it away.
On 2012-05-10 14:52:33 +0000, Jumpin' Jehoshaphat said:
> Why? Are you using old, obsolete, slow equipment or an old, obsolete, > slow network protocol that makes the input/feedback loop take a round > trip to the server and not just complete commands?
Sorry, what I meant was "type and enter commands without long enough to become consciously aware of mistakes". The systems can echo fast enough, I just can't proces the information.
On 2012-05-10 14:54:50 +0000, Jumpin' Jehoshaphat said:
> It is if there's a convenient way to get at infrequently-used stuff you > haven't got memorized.
Right, we call that The TeXbook (or the AMSTeX manual, or whatever). I guess I would have referred to it once a day or so when I typed maths seriously, may be less. For anything you use that seldom it hardly matters what it is.
