1. Thanks for correcting me.
2. Four characters. Very efficient.
3. All to avoid using any control keys. The reason? vi must be able to work with a keyboard that doesn't have any control keys, and vim must work exactly like vi.
Not only do modern keyboards have control keys, but modern computers have mice and other pointing schemes. But vi/vim lives in a permanent time warp, a warp so profound that an old, justly retired editor I wrote in 1977 was far better at exploiting technological advances.
Why was Apple Writer retired? Its limitations quickly became obvious, especially once early Apple computers supported mice (which happened before the Macintosh). No one wanted to keep using such a dinosaur in modern times, including me.
About keyboard, mouse and clipboard support:
> Not only do modern keyboards have control keys, but modern computers have mice and other pointing schemes.
> All to avoid using any control keys. The reason? vi must be able to work with a keyboard that doesn't have any control keys, and vim must work exactly like vi.
> * It cannot take advantage of the modern computing environment -- mouse navigation, different fonts, text styling, page layout, or anything else. Just like vi.
> Keyboards have changed, but vi hasn't. Computers have mice and touch-sensitive displays, but vi can't exploit them. Operating systems have clipboards, but vi can't take advantage of that convention. You know -- don't get me started.
> * No clipboard support -- all copying and pasting operations, apart from being another example of bizarre vi incantations requiring the recall of a raft of magic characters, is within vi/vim and disregards the system clipboard.
> What you see as vim appearing to exploit a mouse and system clipboard is actually Bash and (sometimes) a command-line app (example Konsole) doing so, or equivalent utilities outside vim in a non-GUI level. Those features work with any editor in exactly the same way.
> As to using the system clipboard just fine, you can copy text from vim or any other displayed source of text, but you cannot paste it, because it's not vim that's acting, or is in any way aware that text has been copied.
That's what you wrote. It's all false. First, let's level the playing ground - we're now talking about GVIM, which comes with every Vim installation (unless NO_X11 or something is defined at compile time). Whenever you see the word "Vim" below it means GVIM. To make sure we're really talking about the same thing, here are some screenshots of GVIM in action:
http://screenshots.en.sftcdn.net/en/scrn/92000/92299/gvim-7.... http://www.kianryan.co.uk/wp-content/uploads/2012/09/Screens... http://labnotesh.files.wordpress.com/2008/12/gvim-windows.pn...
This way we can avoid going into terminal emulator and consoles details, and focus on Vim itself rather than on things it can be accessed through.
Ok, so first the keyboard and modifiers keys. It has been told you many times already that Vim supports those. It should be evident from the fact that Vim manual describes default keybindings with modifiers in them. For the example look here: http://vimdoc.sourceforge.net/htmldoc/windows.html#window-mo...
This makes your claim false - apparently Vim doesn't have to "avoid all control keys" (or "work exactly like Vi", but that's another matter).
Second, mouse. Of course it is supported, as explained here: http://vimdoc.sourceforge.net/htmldoc/usr_09.html#09.2 and here: http://vimdoc.sourceforge.net/htmldoc/gui.html#gui-mouse
It is even mentioned that the mouse works in console version, but let's leave it for other time and keep the argument simple. The point is that you can use the mouse for everything you would use it in other editors out of the box, and if not, then you can easily remap it to do something else.
Which makes you claim about Vim not supporting mouse obviously false.
The systems' clipboard support is of course available too. Actually, it is exposed by default via two special purpose registers, named 'star sign' (the thing that makes text italic here...) and '+', which is documented in http://vimdoc.sourceforge.net/htmldoc/gui_x11.html#x11-selec... Note that on Windows the latter is an alias for the former, because the system has only one clipboard. There are commands for placing text in registers, appending to registers and pasting from registers and they work in exactly the same way with 'native' registers and with the ones representing OS clipboard. So it's, of course, completely false that Vim doesn't support clipboard.
One thing of note, you wrote:
> Four characters. Very efficient.
in response to me pointing out how to get the text into clipboard. Well, it's because by default Vim doesn't use OS clipboard as a default register, but you can of course make it do so. If you really want.
On efficiency and modes:
> Some of them might even think it's natural to have to use twice as many keystrokes as are required on a modern editor, even another command-line editor, to switch from editing to navigating (as just one example)
> No, to switch from entering text, to deleting text, to searching for text, to copying and pasting text, to saving a file, to loading a file, all require switching modes. Let's be honest.
> Having to enter additional keystrokes, a throwback to a computer keyboard that either didn't have control keys or couldn't exploit them, makes editing "faster and more powerful"? Nice reasoning.
> * To perform multiple operations, like moving the cursor around, entering text, deleting text, searching the document, saving and loading files, requires you to switch modes. Just like vi.
> It is less efficient. But you are now obviously trolling, posing arguments that no rational person could possibly accept. "So, a few more keystrokes for each command, repeated for each such command. So what? Are you in a hurry to get somewhere? Do you want to live forever?"
Well, you basically say that having modes is less efficient than not having them. The problem here is that you don't present any evidence to back this claim. You say that you need to press more keys when working with Vim, but despite multiple people asking you to, you don't give even one realistic scenario where it would be less keystroke to perform an action in other editor. The few examples you provided were unnecessarily verbose and you were given shorter versions, which require at most the same number of keystrokes as other editors. And it's without custom mappings.
Despite this - and despite the fact that I even included <esc> for changing the mode and still got shorter commands - you're convinced that having modes necessitates more keystrokes to perform an action. I'm at a loss here, how could I ever convince you otherwise? You didn't even post any more examples of what you think takes more time in Vim as if you thought that the previous ones are enough to prove your claim. It's silly - and false. Of course.
The truth is that having modes allows for not using modifier keys, except for one. Without modes you need to use them, because you have no other way of differentiating text input and commands. And that's all modality does, by itself - replaces a few modifier keys with a single one. It has no impact on average count of keystrokes.
For example, copying a full line of text in other editors goes like this (well, using a mouse is an option, but until it's proved or disproved to be fatser than keyboard for text editing I'm not going to include it):
Home, Shift+End, Ctrl+C
In Vim, thanks to modality, the command doesn't involve modifier keys and is equally long (or if you count pressing modifiers as keystrokes - shorter):
<esc>yy
For replacing text inside parens, starting from inside:
go to one of the parens, Shift+Alt+arrow, arrow in opposite direction (to deselect matching paren), and start writing (inserting deleted paren)
In Vim:
<esc>ci(
Moving a paragraph below another one:
go to the beginning of the paragraph, press shift+arrow down enough times, ctrl+x, move to the end of next paragraph, ctrl+v
(No wonder using mouse is more efficient than this!)
In Vim:
<esc>dip}p
And so on. Of course, you can shorten the modeless commands by introducing additional, special purpose commands, so don't take them seriously, but that's not the point - do you really think you can shorten them any more than what Vim does? If so, please show us. No mouse please.
Anyway, if keystrokes count remains the same, why bother with modality? I don't know, to be honest. Vim proponents claim that (with <esc> mapped to CapsLock key) they never leave home row. I'm not a touch-typist and so I don't feel it's a great difference. But at the same time I can't see why is it a bad thing. The keystrokes count stays the same. Chaining is easier, especially when using numerical prefixes. There is a bit of an overhead in that you need to learn of various commands for changing modes - and everyone using Vim admits this - but that does not hamper efficiency in any way once learned.
I wanted to, but I think I won't comment on the rest of what you wrote, because it carries no substance whatsoever. We're really not interested in ed, ex and vi early history, we're talking about current versions of Vim, which is, contrary to what you say, != vi. So here you go, your claims at least commented at length. I hope it's enough and that I will be able to just paste the link to this post under your Vim related comments in the future.
It's too bad you're unable to find any examples to support this claim. On that topic, here's just one of your outright lies:
[ my remark ]
>> All to avoid using any control keys. The reason? vi must be able to work with a keyboard that doesn't have any control keys, and vim must work exactly like vi.
[ your reply ]
> That's what you wrote. It's all false.
As a matter of fact, more key entries are always required in vi and vim compared to a modern editor, and two, the control scheme in vi and vim arises from the requirement that it avoid the use of control keys. That was a design requirement, because the original machines it was designed for did not have control or navigation keys. How do I know this? I used those computers, and I used the early versions of the programs under discussion.
Since then, the machines have changed, but vi and vim have not changed.
Stop lying.
> I wanted to, but I think I won't comment on the rest of what you wrote, because it carries no substance whatsoever.
I already used repeated and substantial evidence to prove that that you are lying, and you just lied again.
> And that's all modality does, by itself - replaces a few modifier keys with a single one. It has no impact on average count of keystrokes.
I and others have already proved that this lie is, in point of fact, a lie. Mode-limited editors require more keystrokes than those without modes. Mode-limited editors require you to be aware of, and to change, modes. Those mode changes require extra keystrokes.
* Vi/vim, switch from moving the cursor to editing text: Esc, Esc, :, i, fist text entry. Four additional keystrokes. Newer versions have s shortcut in which you can just press i, to switch modes, one additional keystroke. But all version of vi and vim always require additional keystrokes, compared to a modern editor.
* Modern editor, switch from moving the cursor to entering text: Just start entering text, no additional keystrokes required.
You are lying.
> We're really not interested in ed, ex and vi early history
First, based on the positive votes on my posts, there is no "we" that lacks interest, and second, one of the the reasons your post is so misguided is because you don't know anything about the history you feign disinterest in. The other reason is that, like a religious True Believer, you cannot be persuaded to argue in good faith.
> Anyway, if keystrokes count remains the same, why bother with modality?
One, I just proved that keystroke count doesn't remain the same, and two, the reason for modes is to avoid use of control keys, for the reason that those keys didn't exist on the keyboards for which vi and vim were designed.
The designers of vi and vim had no choices about modes, it resulted from a limitation of 1970-era computers and terminals, and such modes have been abandoned by all modern editors without exception, as a clear hindrance to efficient use of a computer.
Stop lying.
The original docs for vi, written by Bill Joy and Mark Horton, are online. You can see quite a few places where the docs explicitly refer to control key combinations. For example:
http://docs.freebsd.org/44doc/usd/12.vi/paper-8.html#section...
vi was developed in 1976 on the ADM-3A dumb video terminal, and that terminal had a ctrl key. I remember because I had one. You can see the ADM-3A keyboard here:
http://www.catonmat.net/blog/why-vim-uses-hjkl-as-arrow-keys...
The author of that page also explains why Bill Joy chose hjkl as movement keys. Note that those keys have arrows printed on them to indicate that they move the cursor when used with the ctrl key.
Although vi was never intended to work on paper terminals as you claim (having the line-oriented ed embedded as a mode does not make vi equivalent to or derived from ed), the most common paper-oriented terminal in use back then was the Teletype Model 33, and it had a control key. You can clearly see it here:
http://www.baudot.net/gil/tty-machines/pics/M33-ASR-Gil-Keyb...
The ADM-3A gave way to video terminals from DEC, HP, Wyse, Hazeltine, etc., and they all had control keys. I've worked on a lot of terminals in my time, and I have to go back to keypunch machines to find something that didn't have (or need) a control key.
So what machine did you use vi on that didn't have any control keys?
If such a terminal indeed existed and you were forced to use it at NASA, I'm sorry. The experience obviously left you bitter and scarred. But please stop making claims about vi and early terminals that are simply and obviously false.
I don't know if Bill Joy modded his ADM-3A, but for anyone who wonders why vi uses esc instead of caps lock to switch modes it's because there was no caps lock key back then. I believe Bill Joy and the vi cartel deliberately suppressed the caps lock feature on early terminals to make vi users miserable in the future.
First, I never said either of those things anywhere. But, having sampled the level of personal responsibility among the people in this thread, I won't bother asking you to try to locate your source for those false claims.
The reason vi doesn't use control keys is because many keyboards in those days didn't have control keys to use. Which word don't you understand? The authors couldn't exploit keys that didn't exist, so they avoided use of an entire keyboard filled with control keys, an unfortunate necessity.
This is not to say there were no keyboards with control keys, only that the designers didn't want to lock out users having keyboards without control keys. That's why vi used letter keys for arrow keys, and shifted modes by means of something other than a control key, to make control functions possible in the absence of control keys.
> But please stop making claims about vi and early terminals that are simply and obviously false.
What? I should defend statements that I never made, that you invented and then tried to hold me responsible for? If you will stop inventing quotes, I will stop calling you out on them. Now, as just one example among many, locate where I said "vi is lame" ANYWHERE, FOR ANY REASON, IN ANY CONTEXT. And when you discover that you invented it, you will apologize.
I've given examples with supporting evidence. What evidence have you given to back up your claims?
I didn't invent any quotes. The quotes I posted -- using the widely-accepted quotation mark convention of English (assuming your keyboard has those!) -- are clearly indicated. The "vi is lame" statement is my own summary of your opinions expressed in this thread, it is not attributed to you as a quote but as a sentiment. Thus the lack of surrounding quotation marks.
But maybe let that bone go and try to produce any evidence supporting the meager meat of your arguments about early terminals and the history of vi.
---
lutusp 2 minutes ago | link
> You repeatedly claim that vi is lame because it has to work with keyboards with no control keys. What? "vi is lame because it has to work with keyboards with no control keys"? I never said that anywhere. You will now try to find your source for this lie, then you will apologize for lying. For God's sake -- how can you people expect to have useful conversations while making stuff up? If you had said, "I think I heard you say..." or something like that, I would cut you some slack for being stupid and irresponsible, but not under these circumstances. FIND THE QUOTE. THEN, REALIZING THAT IT DOES NOT EXIST, APOLOGIZE. I won't hold my breath for you to accept adult responsibilities.
---
OK, I'll take the troll bait. You wrote: "I never said that [vi is lame because it has to work with keyboards with no control keys] anywhere. You will now try to find your source for this lie, then you will apologize for lying."
I'm not going to apologize to you for anything, and you show yourself to be quite a jerk telling me that I will apologize for pointing out your factual errors. You don't even bother to address the facts I posted. But showing you where you wrote what you claim you never wrote is easy. You can find all of these quotes in your own comments:
* "The reason? vi must be able to work with a keyboard that doesn't have any control keys..."
* "... the control scheme in vi and vim arises from the requirement that it avoid the use of control keys. That was a design requirement, because the original machines it was designed for did not have control or navigation keys. How do I know this? I used those computers, and I used the early versions of the programs under discussion."
* "And do you know why this arrangement exists? Because there were teletype terminals and early glass terminals that didn't have control keys. No control keys, no way to exploit their existence."
* "I wrote an incremental improvement over vi, one that exploited the existence of control keys..." [Implying that vi did not or could not use the control keys, which didn't exist back then anyway according to you.]
* "It's false when comparing mouse use against a modern keyboard with control and function keys, and it is certainly false for the limited keyboards for which vi/vim was designed, those keyboards that result in vi/vim not being able to exploit control characters."
* "Having to enter additional keystrokes, a throwback to a computer keyboard that either didn't have control keys or couldn't exploit them, makes editing 'faster and more powerful'? Nice reasoning."
* "Vim still imitates vi, and vi still imitates 1970. Extremely inefficient, steep learning curve in order not to press readily available control keys."
* "As soon as someone invented the control key and its purpose, and well before Douglas Engelbart invented the mouse, the retirement clock started ticking for vi and vim." [I guess not since vim is installed on all Unix and Linux distros and comes with OSX].
I provided links showing the ADM-3A and Teletype Model 33 keyboard layouts -- both with control keys. And that was the standard keyboard layout back then. Show me a terminal that was in widespread use when vi came out in 1976 that didn't have a control key. Not that it really matters if you find some weird example, because vi was specifically designed for the ADM-3A and similar terminals available in 1976 that had control keys. That's from Bill Joy.
Your repeated claims that vi was designed for paper terminals, or that vi is derived from ed (rather than the reality that vi incorporated ed) are contradicted by knowledge of vi but also by the authors of vi. You wrote:
* "vi, which had to be able operate with a 'paper terminal', essentially a roll of paper as a display..."
* "I can say this in vi/vim's favor: it's a step up from ed, its predecessor."
Both obviously false.
Read much? You claimed that I said "vi is lame". In point of fact, I never said any such thing anywhere. Prove me wrong -- locate a quote that does not exist. You're inventing arguments to have with yourself.
I just went through your list of quotes, and everything in the list is true. "... there were teletype terminals and early glass terminals that didn't have control keys ..." Care to dispute this truth?
> I'm not going to apologize to you for anything ...
Of course not. You also won't avoid making things up as you have done in this and prior posts. Which leads an intelligent person to wonder what the point is of trying to discuss things with you.
You can dispute the presence or absence of control keys on early terminals, sure. But find one that was in widespread use after vi was released in 1976. I asked you already to show an example of a terminal that didn't have a control key that you used vi on. It's irrelevant, really, since vi was designed specifically for terminals like the ADM-3A that DID have a control key. Bill Joy's original vi docs support that statement.
The possible existence of some early terminal that lacks a control key doesn't change the fact that vi wasn't designed for such terminals, so it shouldn't be surprising that vi didn't work well on such things. Supposing they existed. I posted links and screen shots. What evidence do you have?
I haven't made anything up. I posted your own quotes verbatim and cited my sources for early terminals and the history of vi.
Give it a rest. No one is interested in this anymore and you are just making a fool out of yourself.
While I respect the research done concerning modal UI and the associated consequences of that in the context of HCI, in practice I do not find that there is ever any confusion about what mode I am in when using Vim after using the editor for many hours, and if one was supremely concerned with avoiding such confusion UI changes per state could be implemented so that the changes were unavoidably obvious (such as by changing the entire background to a different color on a context switch).
The design has some nice benefits elsewhere too with how Vim uses it.
However, I do not find Emacs chording emphasis particularly demented either though, especially once the full array of modifiers are used.
The amount of chords is only really burdensome due to how many domains Emacs covers though IMO, as it makes collisions with short chords much more frequent as the amount of keybinds necessarily expands.
I find the criticisms of Vim usually extremely exaggerated and generally indicative of ignorance concerning its design, some of which you have correctly noted and addressed in your response.
That is not to say that Vim is without its imperfections, but the design is well thought out, coherent, and well documented which makes the learning curve bearable.
I find a mixture of chords and modes in Emacs quite nice using Evil-mode as I can choose to either avoid changing states or not, unlike in Vim.
As I have noted elsewhere though, there is no intrinsic reason why Emacs must use chording over a modal design or some other means of using individual key sequences to invoke commands.
It is merely a matter of redesigning how commands are invoked across Emacs, something which some people are working on and may be interesting in the future.
I do wish there was something up to date that was similar to Practical Vim but for Emacs though as that was a very illuminating and fun text.
Do you know of some documentation that elucidates CEDET more than the documentation on the site does? I have read that documentation a few times and have yet to grok what it really does and how one would use it in practice. It seems extremely powerful but I do not see how an end user takes advantage of it or how an extension writer would use it to compose an IDE for a new language.
It's been 5 months since I switched to Emacs. My .emacs.d is actually bigger than .vim_runtime was, but it's because of more configurable features found in Emacs plugins, so it's natural. My init.el is also longer than my .vimrc, but this is because of much friendlier Emacs API and my love for Lisps (and dislike for vimscript). On the whole I'm quite pleased with the switch, although making Emacs into my editor took a comparable amount of time and code to Vim. I don't use Evil or anything like it, of course.
Emacs is better in many respects. CEDET is nice, SpeedBar and imenu with it's syntactic extensions work wonders - TagBar in Vim comes close to SpeedBar, but there's no imenu. auto-complete is better than omni-complete and easier to extend. Eshell works better than conque. Icicles, ido and smex make working with minibuffer a pleasure to the extent that Vim could never match - although it has basic completion of commands. Of course there is more, but the important thing to note here is that, while Emacs implements many things a bit better, Vim has them too. And it's perfectly reasonable to prefer Vim versions. So what I'm saying is that I found Emacs better but Vim is still really great.
Ok, so now to the modality. Emacs lacks modes, Vim has them. Last time we talked you were saying that modes are inefficient and hard to use. I could argue for modality then, but not against it, because I lacked experience with non-modal editor. Now I can compare the two.
The truth is that I was suspicious of non-modal editing in Emacs. I heard and read many rants about long sequences of modifier keys you need to press in order to do something. Turns out these rants where right for the most part. Some examples: C-- C-4 C-x <tab> performs hard dedent by four spaces. C-s C-w C-s ... performs incremental search (forward) for current word (Esc # in Vim). C-4 M-f goes four words forward (4w in Vim). There are also tricky ones, like C-x d opening dired and much easier to type C-x C-d displaying a list of files in current directory in a buffer.
In practice, however, with some amount of remapping, this is not an issue. But wait, isn't pressing escape instead of some universal prefix (C-c for user commands in emacs) not a problem in practice, too? After using Emacs and Vim I can say that the complexity of commands in both is almost the same. In Vim I remapped Esc to CapsLock - in Emacs I have Ctrl in place of CapsLock (although I'm contemplating placing Alt there and Ctrl in place of Alt). The only real difference I feel is that in Vim I could just hit the first modifier key (esc) while in Emacs I need to keep it pressed. Emacs uses Alt (Meta), while Vim does not, but that's a little difference, because Alt is usually in a place that's easy to reach.
In conclusion, while I like Emacs very much, I still find your claim that "modal editors" are for sad, handicapped individuals who just don't know any better full of bullshit. Modality is not any worse than modeless model in itself and Vim implements modality in a way which makes using it as convenient as Emacs chords. While there are other things I like in Emacs better, bashing Vim because of modes alone is pure stupidity; bashing it because of it's ancestry is an equivalent of saying that "your mama is fat" and saying it has no X feature is most likely false anyway.
PS. Elsewhere in this thread you're saying that Vim can't use mouse and clipboard. Of course it can - <esc>"=y copies to system clipboard instead of local one. I admit that Emacs support for mouse events is by far superior, and of course typesetting - colors, sizes and styles of fonts - are well supported in Emacs and not in Vim. But reaching for the mouse is bad anyway and if you have to, Vim supports all three mouse keys, which only makes the latter a valid claim. It's up to everyone to decide whether they truly need different fonts in their editor.