[usability] mouse-1 for performing actions?

[usability] mouse-1 for performing actions?

[Simon Josefsson]

Why does emacs use mouse-2 for activating hyperlinks, buttons etc?

I haven't seen any other application doing this, and altough I don't
remember what I thought when I started using emacs, I could imagine
that a new user would be surprised by this.  Does anyone else agree?
Does anyone here talk with newbie users and have more insight?

I recall one argument; that mouse-1 is used to position point and is
therefor already taken.  I don't understand the argument, many GUI
applications have something similar as "point" and they don't have a
problem with using mouse-1 for two things.  Another argument on the
same theme is that mouse-1 is used to mark text.  Again, I don't see
the argument, emacs is able to separate the two cases where mouse-1 is
simply clicked and when it is used to mark text.  Many GUI
applications behave differently depending on how you click mouse-1.

Proposal: Make mouse-1 do the same as mouse-2 in Info (section links),
Help (customize link, back link, variable links, etc), and similar
buffers.

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Why does emacs use mouse-2 for activating hyperlinks, buttons etc?

Because these buffers are usually read-only, which means that the
normal definition of mouse-2 (yank) is not applicable.

The normal definition of mouse-1 is perfectly meaningful in these
buffers.  It seems like a bad idea to replace it with something
incompatible and make the normal definition of mouse-1 unavailable.

Do you have a precise proposal that you think is ok?

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

Simon Josefsson <jas@extundo.com> asked:

   Why does emacs use mouse-2 for activating hyperlinks, buttons etc?

Just now I clicked on a URL using my middle mouse button to activate
the Galeon Web browser.  The site did not appear.  So I marked a
portion of the URL, the site name, using the first mouse button, and
copied that name so I could run ping.  I could not do that if mouse-1
were to activate the browser.

    ... emacs is able to separate the two cases where mouse-1 is
    simply clicked and when it is used to mark text.

I do not understand:  I click mouse-1 to mark the beginning of a
region, move the mouse cursor, and then click mouse-3 to mark the end
of the region.  If the two commands are on the same button, and used
in the same context, how can Emacs distinguish between a click to mark
the beginning of a region and a click to follow a hyperlink?

    Proposal: Make mouse-1 do the same as mouse-2 in Info (section links),

When telling people where to find documentation, I often copy Info
links.  I write text such as, `go to File: emacs,  Node: Mouse Commands'

Incidentally, you cannot copy Info links in a plain vanilla Emacs 21
any more.  The new default makes the links be a part of Emacs that you
cannot copy.  For me this is a mistake, and I always set the value of
`Info-use-header-line' to nil so I can copy the links.  

It looks to me that the proposal, if adopted, would be like this new
default; it would make it harder to copy text, which is to say, the
proposal would lead to more typing, when copying would be easier,
quicker, and more accurate.

-- 
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

> Simon Josefsson <jas@extundo.com> asked:
>
>    Why does emacs use mouse-2 for activating hyperlinks, buttons etc?
>
> Just now I clicked on a URL using my middle mouse button to activate
> the Galeon Web browser.  The site did not appear.  So I marked a
> portion of the URL, the site name, using the first mouse button, and
> copied that name so I could run ping.  I could not do that if mouse-1
> were to activate the browser.

Marking text is (by me, at least) usually done by down-mouse-1 at the
start point and up-mouse-1 at the end point.  This would still work
even if clicking on mouse-1 would activate the browser.

>     ... emacs is able to separate the two cases where mouse-1 is
>     simply clicked and when it is used to mark text.
>
> I do not understand:  I click mouse-1 to mark the beginning of a
> region, move the mouse cursor, and then click mouse-3 to mark the end
> of the region.  If the two commands are on the same button, and used
> in the same context, how can Emacs distinguish between a click to mark
> the beginning of a region and a click to follow a hyperlink?

I didn't know about this feature.  When I mark text, I do it by
holding mouse-1 down until I reached the end of the marked text.  Ok,
so it is difficult to change the behaviour in a backwards compatible
way.

However, in some modes activating buttons is probably more common than
marking text in this way, would it make sense to have the
click-mouse1-to-activate-button in such buffers?

Btw, do ANY other application have this mouse1 + mouse3 mark text
feature?  I haven't seen it before.  IMHO dropping this feature in
favor of activating buttons and links with mouse-1 is more user
friendly (except for old time Emacs users) as it is consistent with
how other applications behave.

>     Proposal: Make mouse-1 do the same as mouse-2 in Info (section links),
>
> When telling people where to find documentation, I often copy Info
> links.  I write text such as, `go to File: emacs,  Node: Mouse Commands'
>
> Incidentally, you cannot copy Info links in a plain vanilla Emacs 21
> any more.  The new default makes the links be a part of Emacs that you
> cannot copy.  For me this is a mistake, and I always set the value of
> `Info-use-header-line' to nil so I can copy the links.  
>
> It looks to me that the proposal, if adopted, would be like this new
> default; it would make it harder to copy text, which is to say, the
> proposal would lead to more typing, when copying would be easier,
> quicker, and more accurate.

No -- you should still be able to mark text in a button by dragging
mouse-1 over the area you want to mark.

I agree that the Info header line should allow the user to mark text
in it.  It should be possible to mark text =everywhere= IMHO.

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

   Btw, do ANY other application have this mouse1 + mouse3 mark text
   feature?  I haven't seen it before.

You must have been using poorly designed applications.  The two button
marking technique often makes it easier for people to mark a specified
region than dragging.  (Sometimes dragging works OK -- it depends on
the situation.)

   IMHO dropping this feature in favor of activating buttons and links
   with mouse-1 is more user friendly ..

We should never reduce the quality of the Emacs user interface just
because other applications use worse interfaces.

Please suggest to the developers of other applications that they
improve their work.  (Besides improving the methods for copying,
please ask these developers to implement incremental search, which is
a highly comfortable and efficient way to navigate.)

--
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: [usability] mouse-1 for performing actions?

[Stefan Monnier]

>     Why does emacs use mouse-2 for activating hyperlinks, buttons etc?
> 
> Because these buffers are usually read-only, which means that the
> normal definition of mouse-2 (yank) is not applicable.
> 
> The normal definition of mouse-1 is perfectly meaningful in these
> buffers.  It seems like a bad idea to replace it with something
> incompatible and make the normal definition of mouse-1 unavailable.
> 
> Do you have a precise proposal that you think is ok?

I understand the above reason, but I must agrre that newbies find
it surprising that nothing happens when the click with mouse-1
(most of them rarely use mouse-2 since most applications only ever
use mouse-1 anyway or mouse-3 occasionally for a contextual menu).

We should try and come up with some way to have our cake and eat it too.
For this we need to figure out what do people use "mouse-1 click on
mouse-sensitive text" for.  In my case, it's mostly for cut&paste, which
means that drag-mouse-1 should definitely prevent the hyperlink activation.
Maybe I also sometimes us mouse-1 to place the cursor just before or
just after the hyperlink (the just-after case is OK because it doesn't
risk activating the hyperlink, but the just-before might require special
handling).


	Stefan

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

>    Btw, do ANY other application have this mouse1 + mouse3 mark text
>    feature?  I haven't seen it before.
>
> You must have been using poorly designed applications.  The two button
> marking technique often makes it easier for people to mark a specified
> region than dragging.  (Sometimes dragging works OK -- it depends on
> the situation.)

If this is a good idea (I don't know--I've never used the technique)
perhaps it should be suggested to the GNOME people.  Right now the
GNOME Human Interface Guidelines [1] says something that doesn't fit
well with how Emacs behaves:

Recommendations:

+ The left button should be used for selecting, activating components,
  dragging, and the display of drop-down menus.

+ The right button should be used to display and select actions from a
  popup menu.

+ The middle mouse button should be used to paste the current PRIMARY
  (usually the last-highlighted) selection at the pointer position, as
  follows:
...

[1] http://developer.gnome.org/projects/gup/hig/draft_hig/userinput.html#mouse-buttons

>    IMHO dropping this feature in favor of activating buttons and links
>    with mouse-1 is more user friendly ..
>
> We should never reduce the quality of the Emacs user interface just
> because other applications use worse interfaces.
>
> Please suggest to the developers of other applications that they
> improve their work.  (Besides improving the methods for copying,
> please ask these developers to implement incremental search, which is
> a highly comfortable and efficient way to navigate.)

The goal should IMHO be consistency within the GNOME desktop since it
is the GNU desktop, and Emacs can be a desktop application.

Re: [usability] mouse-1 for performing actions?

[Francesco Potorti`]

Simon Josefsson <jas@extundo.com> writes:
   Btw, do ANY other application have this mouse1 + mouse3 mark text
   feature?

gpm does it.

But apparently this is not the standard with X applications.

rxvt (similar to xterm) does not use this technique.  Rather, it uses
dragging mouse1 to select a region.  But once a region is selected, you
can extend it on both sides using mouse3.

Netscape, as far as I can see, only uses dragging.

Probably, Emacs should use the same conventions as the Gnome
applications.  Anyone knows if there is such a standard?

   I haven't seen it before.  IMHO dropping this feature in
   favor of activating buttons and links with mouse-1 is more user
   friendly (except for old time Emacs users) as it is consistent with
   how other applications behave.

Emacs is oriented towards editing, not navigating.  Having mouse1 on a
link jump to that link is not what commonly one would expect from an
editor.

Re: [usability] mouse-1 for performing actions?

[Alan Shutko]

Francesco Potorti` <pot@gnu.org> writes:

> Simon Josefsson <jas@extundo.com> writes:
>    Btw, do ANY other application have this mouse1 + mouse3 mark text
>    feature?
>
> gpm does it.

FWIW, gnome-terminal does, most Gnome apps do not.

-- 
Alan Shutko <ats@acm.org> - In a variety of flavors!
There are new messages.

Re: [usability] mouse-1 for performing actions?

[Karl Eichwalder]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

> You must have been using poorly designed applications.

No, they are just different.

> The two button marking technique often makes it easier for people to
> mark a specified region than dragging.  (Sometimes dragging works OK
> -- it depends on the situation.)

In Gnus or Info buffers URLs or other highlighted elements are atoms
the user wants to activate with mouse-1.  If you are interested in a
part of a URL only, marked it as a whole, paste the whole thing in an
edit buffer and change it as wanted -- tedious, but tolerable since
this is an exection.

> We should never reduce the quality of the Emacs user interface just
> because other applications use worse interfaces.

For a casual user Emacs' mouse usage is clumsy (to put it mild); it
just does not fit in the desktop environment.

> (Besides improving the methods for copying, please ask these
> developers to implement incremental search, which is a highly
> comfortable and efficient way to navigate.)

Sure, but that's not the point.  Some of us want to use the mouse, but
the mouse has to DTRT.  Also default Emacs lacks the possibilty to drag
highlighted text to another location.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

"Robert J. Chassell" <bob@rattlesnake.com> writes:
> Incidentally, you cannot copy Info links in a plain vanilla Emacs 21
> any more.  The new default makes the links be a part of Emacs that you
> cannot copy.  For me this is a mistake, and I always set the value of
> `Info-use-header-line' to nil so I can copy the links.  

Of course you could just use the `Info-copy-current-node-name' command,
which is in the Info menu (and also now bound to `c').

-Miles
-- 
"I distrust a research person who is always obviously busy on a task."
   --Robert Frosch, VP, GM Research

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Karl Eichwalder <keichwa@gmx.net> writes:
> Also default Emacs lacks the possibilty to drag highlighted text to
> another location.

Eh?  It works just fine when I try it (using emacs -q)...

-Miles
-- 
Ich bin ein Virus. Mach' mit und kopiere mich in Deine .signature.

Re: [usability] mouse-1 for performing actions?

[Karl Eichwalder]

Miles Bader <miles@lsi.nec.co.jp> writes:

> Eh?  It works just fine when I try it (using emacs -q)...

Okay, I must read again how drag & drop is considered to work; copy-drag
& drop using mouse-2 works (changing the mouse pointer while dragging is
still missing).  Of course, a new user would expect doing the action
with mouse-1.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

Simon Josefsson <jas@extundo.com> writes:

> "Robert J. Chassell" <bob@rattlesnake.com> writes:
>
>> I do not understand:  I click mouse-1 to mark the beginning of a
>> region, move the mouse cursor, and then click mouse-3 to mark the end
>> of the region.  If the two commands are on the same button, and used
>> in the same context, how can Emacs distinguish between a click to mark
>> the beginning of a region and a click to follow a hyperlink?
>
> I didn't know about this feature.  When I mark text, I do it by
> holding mouse-1 down until I reached the end of the marked text.  Ok,
> so it is difficult to change the behaviour in a backwards compatible
> way.
>
> However, in some modes activating buttons is probably more common than
> marking text in this way, would it make sense to have the
> click-mouse1-to-activate-button in such buffers?
>
> Btw, do ANY other application have this mouse1 + mouse3 mark text
> feature?  I haven't seen it before.  IMHO dropping this feature in
> favor of activating buttons and links with mouse-1 is more user
> friendly (except for old time Emacs users) as it is consistent with
> how other applications behave.

Note that clicking mouse3 twice cuts the text.  Very nifty.  By
default, this feature is not available when dragging mouse-1.

But with mouse-sel enabled (I think that's the package), you can
press mouse-1, drag the mouse, then _also_ click mouse-3, then
release mouse-1, to effect a cut operation.  Even niftier :-)

Hm.

XTerm has this feature: if something prevents you from using mouse-1
for its normal function (for example when running Emacs with
xterm-mouse-mode enabled), then you can still use S-mouse-1 to get
the function.  I wonder if this could be translated to Emacs?  (Of
course, there is a binding for S-mouse-1 already.  This needs to be
dealt with.)

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Andreas Schwab]

Francesco Potorti` <pot@gnu.org> writes:

|> Simon Josefsson <jas@extundo.com> writes:
|>    Btw, do ANY other application have this mouse1 + mouse3 mark text
|>    feature?
|> 
|> gpm does it.
|> 
|> But apparently this is not the standard with X applications.
|> 
|> rxvt (similar to xterm) does not use this technique.  Rather, it uses
|> dragging mouse1 to select a region.  But once a region is selected, you
|> can extend it on both sides using mouse3.

You can also mark text without dragging in both terminals.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE GmbH, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

Miles Bader <miles@lsi.nec.co.jp> writes:

> Karl Eichwalder <keichwa@gmx.net> writes:
>> Also default Emacs lacks the possibilty to drag highlighted text to
>> another location.
>
> Eh?  It works just fine when I try it (using emacs -q)...

AFAIK, Emacs does not support drag-and-drop, only
cut/copy-and-paste.  I think Karl is speaking of drag-and-drop.

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
> AFAIK, Emacs does not support drag-and-drop, only
> cut/copy-and-paste.  I think Karl is speaking of drag-and-drop.

Well, emacs supports the `select with the mouse, and then insert
elsewhere with mouse-2' which is basically `drag-n-drop for text'
(and fairly standard across X apps).

Is he referring to some specific drag-and-drop protocol?

-Miles
-- 
Is it true that nothing can be known?  If so how do we know this?  -Woody Allen

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

Miles Bader <miles@lsi.nec.co.jp> writes:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>> AFAIK, Emacs does not support drag-and-drop, only
>> cut/copy-and-paste.  I think Karl is speaking of drag-and-drop.
>
> Well, emacs supports the `select with the mouse, and then insert
> elsewhere with mouse-2' which is basically `drag-n-drop for text'
> (and fairly standard across X apps).

Cut-and-paste is functionally equivalent to drag-and-drop, but the
interface is different.  I just tried nedit, it supports
drag-and-drop: you mark something with mouse-1, then press mouse-2 on
it, then drag the mouse and you will see the text move in real-time,
as you move the mouse.  Releasing the mouse button then finishes the
move.

Drag-and-drop gives the user more feedback about what happens.

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Francesco Potorti`]

   |> rxvt (similar to xterm) does not use this technique.  Rather, it uses
   |> dragging mouse1 to select a region.  But once a region is selected, you
   |> can extend it on both sides using mouse3.
   
   You can also mark text without dragging in both terminals.

From the rxvt manual:
       Selection:
              Left click at the beginning of the region, drag  to
              the  end  of the region and release; Right click to
              extend the  marked  region;  Left  double-click  to
              select  a  word;  Left  triple-click  to select the
              entire line.

Re: [usability] mouse-1 for performing actions?

[Karl Eichwalder]

Miles Bader <miles@lsi.nec.co.jp> writes:

> Is he referring to some specific drag-and-drop protocol?

Implicitly.  Like GNOME supports Emacs' kill-and-yank features, Emacs
sould be modified/enhanced to match the mouse behavior on GNOME.

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai writes:

> Cut-and-paste is functionally equivalent to drag-and-drop, but the
> interface is different.  I just tried nedit, it supports
> drag-and-drop: you mark something with mouse-1, then press mouse-2 on
> it, then drag the mouse and you will see the text move in real-time,
> as you move the mouse.  Releasing the mouse button then finishes the
> move.
>
> Drag-and-drop gives the user more feedback about what happens.

Drag-and-Drop is especially useful when you want to move files from one
directory to the other or messages from one folder to the other.  Or
URLs from an Emacs buffer to the X root window.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

Karl Eichwalder <keichwa@gmx.net> writes:

> Of course, a new user would expect doing the action with mouse-1.

Really?  Are there other programs which do this?

nedit uses drag-mouse-2...

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Andreas Schwab]

Francesco Potorti` <pot@gnu.org> writes:

|>    |> rxvt (similar to xterm) does not use this technique.  Rather, it uses
|>    |> dragging mouse1 to select a region.  But once a region is selected, you
|>    |> can extend it on both sides using mouse3.
|>    
|>    You can also mark text without dragging in both terminals.
|> 
|> >From the rxvt manual:
|>        Selection:
|>               Left click at the beginning of the region, drag  to
|>               the  end  of the region and release; Right click to
|>               extend the  marked  region;  Left  double-click  to
|>               select  a  word;  Left  triple-click  to select the
|>               entire line.

The manual obviously doesn't tell the whole truth.  Tested with rxvt
version 2.7.8.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE GmbH, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [usability] mouse-1 for performing actions?

[Francesco Potorti`]

   The manual obviously doesn't tell the whole truth.  Tested with rxvt
   version 2.7.8.

Maybe it does (look at yours), my version is older: 2.6.4, and behaves
as the manual says.

Re: [usability] mouse-1 for performing actions?

[Andreas Schwab]

Francesco Potorti` <pot@gnu.org> writes:

|>    The manual obviously doesn't tell the whole truth.  Tested with rxvt
|>    version 2.7.8.
|> 
|> Maybe it does (look at yours), my version is older: 2.6.4, and behaves
|> as the manual says.

The manual is unchanged in this part.  But it seems to be buggy at least,
because it does not work correctly in the first screen line.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE GmbH, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Incidentally, you cannot copy Info links in a plain vanilla Emacs 21
    any more.  The new default makes the links be a part of Emacs that you
    cannot copy.  For me this is a mistake, and I always set the value of
    `Info-use-header-line' to nil so I can copy the links.  

I will change Info so that does not hide the header line.

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Btw, do ANY other application have this mouse1 + mouse3 mark text
    feature?  I haven't seen it before.

This interface follows that of Xterm.

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:
>     Incidentally, you cannot copy Info links in a plain vanilla Emacs 21
>     any more.
> 
> I will change Info so that does not hide the header line.

Info doesn't `hide' the header line, in fact, it's quite a bit more
visible when it uses the emacs header-line feature, because it doesn't
move out of sight as soon as you scroll.

I would suggest getting more opinions before turning off that feature.

-Miles
-- 
o The existentialist, not having a pillow, goes everywhere with the book by
  Sullivan, _I am going to spit on your graves_.

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Probably, Emacs should use the same conventions as the Gnome
    applications.  Anyone knows if there is such a standard?

There is a good argument for this.  At the same time, it seems that
the current Emacs mouse commands are superior.

Perhaps GNOME should adapt its conventions to allow editors to work
the way Emacs does.

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Of course you could just use the `Info-copy-current-node-name' command,
    which is in the Info menu (and also now bound to `c').

The drawback of that method is that you have to remember it.
I must have known about it once upon a time, but I forgot it.
Bob apparently didn't know about it either.

The simple change I'm going to install will make it easy to
copy this from the text in the usual way.

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Cut-and-paste is functionally equivalent to drag-and-drop, but the
    interface is different.  I just tried nedit, it supports
    drag-and-drop: you mark something with mouse-1, then press mouse-2 on
    it, then drag the mouse and you will see the text move in real-time,
    as you move the mouse.  Releasing the mouse button then finishes the
    move.

    Drag-and-drop gives the user more feedback about what happens.

It would be interesting to try implementing this in Emacs; I don't
think it should be terribly hard.

Re: [usability] mouse-1 for performing actions?

[Francesco Potorti`]

       Probably, Emacs should use the same conventions as the Gnome
       applications.  Anyone knows if there is such a standard?
   
   There is a good argument for this.  At the same time, it seems that
   the current Emacs mouse commands are superior.
   
   Perhaps GNOME should adapt its conventions to allow editors to work
   the way Emacs does.

In order to do this, we need some GUI expert from the Gnome community
to join this discussion, and possibly set up a small workgroup to work
towards this end.  I am saying this in order to give some ideas, but I
personally don't know anyone who fits that description.

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

Francesco Potorti` <pot@gnu.org> writes:

> Probably, Emacs should use the same conventions as the Gnome
> applications.  Anyone knows if there is such a standard?

Yes:

http://developer.gnome.org/projects/gup/
http://developer.gnome.org/projects/gup/hig/
http://developer.gnome.org/projects/gup/hig/draft_hig/

The conventions I quoted from that document does not map well to how
Emacs work right now.

Re: [usability] mouse-1 for performing actions?

[Karl Eichwalder]

Richard Stallman <rms@gnu.org> writes:

> It would be interesting to try implementing this in Emacs; I don't
> think it should be terribly hard.

Note, using mouse-2 for dragging is "wrong" -- maybe, Motif does it this
way, but the majority of all potential user are used to drag objects
with mouse-1.

BTw, mouse-2 is basically a no-op on many a lot devices, esp. on such a
device like my notebook.  And there are mice with only 1 button!

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

Karl Eichwalder <keichwa@gmx.net> writes:

> Richard Stallman <rms@gnu.org> writes:
>
>> It would be interesting to try implementing this in Emacs; I don't
>> think it should be terribly hard.
>
> Note, using mouse-2 for dragging is "wrong" -- maybe, Motif does it this
> way, but the majority of all potential user are used to drag objects
> with mouse-1.

Programs which use mouse-1 for dragging the selected text:  How do
they enable the user to make a new selection?  This is presumably also
done with dragging mouse-1.

> BTw, mouse-2 is basically a no-op on many a lot devices, esp. on such a
> device like my notebook.  And there are mice with only 1 button!

You mean it's hard to hit?  Maybe yes.

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai writes:

> Karl Eichwalder <keichwa@gmx.net> writes:
>
>> Richard Stallman <rms@gnu.org> writes:
>>
>>> It would be interesting to try implementing this in Emacs; I don't
>>> think it should be terribly hard.
>>
>> Note, using mouse-2 for dragging is "wrong" -- maybe, Motif does it this
>> way, but the majority of all potential user are used to drag objects
>> with mouse-1.
>
> Programs which use mouse-1 for dragging the selected text:  How do
> they enable the user to make a new selection?  This is presumably also
> done with dragging mouse-1.

Yes.  If you drag mouse-1 inside the current selection you get
drag'n'drop, if you drag mouse-1 anywhere else it marks.

>> BTw, mouse-2 is basically a no-op on many a lot devices, esp. on such a
>> device like my notebook.  And there are mice with only 1 button!
>
> You mean it's hard to hit?  Maybe yes.

I agree.  I don't have mouse-2 on my laptop.  I can use all GNOME &
KDE applications just fine, but Emacs is almost unusable until I plug
in an external mouse.

The mark/drag behaviour above isn't the only problematic thing in
Emacs if you don't have mouse-2 -- GNOME and KDE have RMB (mouse-3)
menus in input fields which contain "Paste" and "Copy" which are very
useful if you don't have mouse-2.

I would offer to work on this if I had time, but making user
interfaces usable isn't my area of expertise so I'll probably just
waste my time.  Still, as a user I certainly have opinions on how the
application should work though. :-)

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

   ... there are mice with only 1 button!

In 1984, I found that the `1 button' Macintosh mice were actually
multi-button devices, with the second button on the keyboard.  I
think they are still configured like this, with a second button on
the keyboard.

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

Simon Josefsson <jas@extundo.com> wrote:

    I don't have mouse-2 on my laptop.  I can use all GNOME & KDE
    applications just fine, but Emacs is almost unusable until I plug
    in an external mouse.

I am very confused:  how can Emacs be `almost unusable' without an
external mouse?  What I am trying to get at is, what user-model are
you presuming? 

Clearly, a user-model that always presumes access to a 3 button mouse
produces a quite different user interface requirements than a
user-model that presumes access to a 3 button mouse only some of the
time.

(Similarly, a user-model that presumes the user is always going to
remain a novice leads to a different user interface than a user-model
that presumes the user starts as a novice but can learn.)

(You don't need a mouse for Emacs.  Indeed, about a half hour ago, I
was connected, via an ssh hop to a machine and then from there via
telnet, through a slow connection (perhaps 1200 baud; traceroute
indicated some sort of trouble in Newark, NJ) to a third machine,
using Emacs without a mouse.)

-- 
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

> Simon Josefsson <jas@extundo.com> wrote:
>
>     I don't have mouse-2 on my laptop.  I can use all GNOME & KDE
>     applications just fine, but Emacs is almost unusable until I plug
>     in an external mouse.
>
> I am very confused:  how can Emacs be `almost unusable' without an
> external mouse?  What I am trying to get at is, what user-model are
> you presuming? 

Ok, I'm probably not a typical Emacs user, since I use Gnus more than
the regular Emacs user.  And Gnus, like most interactive Emacs modes,
makes heavy use of buttons and hyperlinks where mouse-2 is quite
tricky to use.  Fortunately the keyboard work, but some operations is
alot faster when using the mouse.

> (You don't need a mouse for Emacs.  Indeed, about a half hour ago, I
> was connected, via an ssh hop to a machine and then from there via
> telnet, through a slow connection (perhaps 1200 baud; traceroute
> indicated some sort of trouble in Newark, NJ) to a third machine,
> using Emacs without a mouse.)

Congratulations. :-)   I would call such an environment unusable from a
usability point of view.

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Simon Josefsson <jas@extundo.com> writes:
> Yes.  If you drag mouse-1 inside the current selection you get
> drag'n'drop, if you drag mouse-1 anywhere else it marks.

Note that this particular behavior can be seriously, seriously annoying.

For instance, when you select a region, and get the start wrong by 1
character, you can't just redrag it, you have first get rid of the
region by clicking (outside the region!), and _then_ redragg.

I get stung by this constantly...

-Miles
-- 
`Cars give people wonderful freedom and increase their opportunities.
 But they also destroy the environment, to an extent so drastic that
 they kill all social life' (from _A Pattern Language_)

Re: [usability] mouse-1 for performing actions?

[Karl Eichwalder]

Miles Bader <miles@gnu.org> writes:

> For instance, when you select a region, and get the start wrong by 1
> character, you can't just redrag it, you have first get rid of the
> region by clicking (outside the region!), and _then_ redragg.

No, just tried with galeon.  Click inside without moving this mouse
(yes, that's possible).

> I get stung by this constantly...

You must not use Emacs that extensively ;)

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.suse.de/~ke/                                  |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

   > ... a slow connection (perhaps 1200 baud; traceroute
   > indicated some sort of trouble in Newark, NJ) to a third machine,
   > using Emacs without a mouse.)

   Congratulations. :-)   I would call such an environment unusable from a
   usability point of view.

Many programs are no good at all for remote operations on a slow
connection.

QWEST had some sort of trouble in New Jersey and does not properly
re-route packets the way the Internet was originally designed.  I had
to create a different route in order to get hold of my email.  I
agree, a slow connection makes for a poor interface, but fortunately,
with Emacs, it is basically the same interface I always use, only
slower.

The point of this is that a good user interface takes into account not
only good circumstances, such as I am enjoying as I type these words,
but other circumstances as well, such as remote connections over poor
lines, or being without a mouse on a laptop, or being situationally
blind, as in a car, listening to email.

In many ways, designing for varied conditions is the same as designing
for varied users: on the one hand, the goal is to design an interface
that works when you lose speed, a mouse, and a meta key.  On the other
hand, the goal is to design an interface that novices can not only use
readily (for which I think menus and arrow keys are fine) but, at the
same time, that novices can learn to work more efficiently than is
possible with menus and arrow keys.

(Your descriptions of GNOME and KDE suggests to me that their
interfaces are designed for people who are novices and who are not
expected to learn to become more efficient; I don't whether this is
true.)

Moreover, designing for varied users includes designing an interface
that can be modified for people who type whole sentences or paragraphs
before editing them (Viper mode) as well as being designed for people
who edit each word or clause (Emacs in general).

The question is what defaults are suitable?

As a rule of thumb, for example, Emacs is oriented towards people who
edit as they go along, rather than write for a while and then edit, as
vi users do.  (For someone who edits as they go along, the shift in vi
from insert mode to edit mode and back requires too many keystrokes;
for someone who goes into edit mode once a paragraph, the shift is
hardly noticeable.)

Similarly, Emacs is oriented towards people who prefer to cut, copy,
and paste rather than type, and who operate on 50 or more different
buffers at one time.  Emacs is oriented towards people who navigate by
doing incremental searches, since that is easy, quick, and intuitive.
It is less intuitive and efficent to make nonincremental searches, so
the default Emacs keybinding has an extra character in it.

(The Edit menu provides a string search that is easy for novices and,
being in a menu, is intrinsically slow.  The menu also provides
incremental search, but it goes without saying that no one in their
right minds will use incremental search off a menu more than the once
or twice it takes to learn, since that interface is so slow and
inefficent.)

   Ok, I'm probably not a typical Emacs user, since I use Gnus more than
   the regular Emacs user.  And Gnus, like most interactive Emacs modes,
   makes heavy use of buttons and hyperlinks where mouse-2 is quite
   tricky to use.

That is weird:  I have used Gnus without the mouse.  And I am using
mail mode now.  Often, I use text, Texinfo, and Emacs Lisp mode
frequently.  All of these are interactive.  None of these make much
use of buttons or hyperlinks.  I use the mouse to move point to a
non-near random location and to mark regions that commands like `M-h'
don't handle.  

Only in read-only buffers do I use the mouse much for clicking on
buttons, and that only when that is efficient.  (In the RMAIL-summary
buffer, for example, I tend to use the `d' and `n' keys alot.  I know
I can mouse click on a line and do sometimes, but that is infrequent
compared to the number of times I press the `d' and `n' keys.)

Obviously, a mouse is useful in some circumstances.  I am not saying
that one should not use a mouse; I use one when able.  I am being
puzzled at the degree you use the mouse for things other than marking
and copying regions text or for changing the location of point.

   ... heavy use of buttons and hyperlinks where mouse-2 is quite
   tricky to use.

How is the use of mouse-2 trickier than the use of mouse-1 or mouse-3?

--
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Info doesn't `hide' the header line,

It does two things:

* Set up a header line.
* Make the line of links invisible.

I plan to change the second one only.

Re: [usability] mouse-1 for performing actions?

[Robert J. Chassell]

   Miles Bader <miles@gnu.org> writes:

   > For instance, when you select a region, and get the start wrong by 1
   > character, you can't just redrag it, you have first get rid of the
   > region by clicking (outside the region!), and _then_ redragg.

Karl Eichwalder <keichwa@gmx.net> writes:

   No, just tried with galeon.  Click inside without moving this mouse
   (yes, that's possible).

I just tried this Gnome galeon 1.2.0 and it did not work.  I click on
a spot and that becomes either the beginning or the end of the region.
When I reclick inside the marked region, I still cannot extend both
sides.

-- 
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:
>     Info doesn't `hide' the header line,
> 
> It does two things:
> 
> * Set up a header line.
> * Make the line of links invisible.
> 
> I plan to change the second one only.

Well, that's somewhat uglier (since the link-line is then duplicated
while viewing the first page), but it would solve Bob's complaint.

It would be nice to have a variable to select the old behavior though.

-Miles
-- 
I'm beginning to think that life is just one long Yoko Ono album; no rhyme
or reason, just a lot of incoherent shrieks and then it's over.  --Ian Wolff

Re: [usability] mouse-1 for performing actions?

[Eli Zaretskii]

> From: Simon Josefsson <jas@extundo.com>
> Date: Fri, 24 May 2002 20:14:22 +0200
> 
> > (You don't need a mouse for Emacs.  Indeed, about a half hour ago, I
> > was connected, via an ssh hop to a machine and then from there via
> > telnet, through a slow connection (perhaps 1200 baud; traceroute
> > indicated some sort of trouble in Newark, NJ) to a third machine,
> > using Emacs without a mouse.)
> 
> Congratulations. :-)   I would call such an environment unusable from a
> usability point of view.

I think this is a very biased point of view.  Quite a few people have
slow connections and need to run Emacs via them.

Re: [usability] mouse-1 for performing actions?

[Eli Zaretskii]

> From: Miles Bader <miles@gnu.org>
> Date: 25 May 2002 10:27:54 +0900
> 
> Well, that's somewhat uglier (since the link-line is then duplicated
> while viewing the first page), but it would solve Bob's complaint.
> 
> It would be nice to have a variable to select the old behavior though.

I'd say the old behavior should be the default, with an option
available to make the link-line visible.  I doubt many users need to
copy the text from the link line; I don't remember more than a
handful of questions on user forums, and I myself never had any
reason to copy the links anywhere.

Re: [usability] mouse-1 for performing actions?

[Kai Großjohann]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

> How is the use of mouse-2 trickier than the use of mouse-1 or mouse-3?

On a two-button mouse, mouse-2 is achieved by pressing both buttons
at the same time (proper X11 configuration assumed).  On the laptop
I'm using right now, there is quite a bit of distance between the two
buttons, so hitting them both requires some conscious effort.  (The
fact that the buttons are hard to hit adds to this.)

kai
-- 
Silence is foo!

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Yes.  If you drag mouse-1 inside the current selection you get
    drag'n'drop, if you drag mouse-1 anywhere else it marks.

I don't think this would be hard to implement, so I suggest once again
that someone give it a try.  It is not necessary to know anything
special about GUI programming, it is only necessary to change the
command that is bound to down-mouse-1 so that in some cases it moves
the region text to where the mouse is.  It should try to do something
smart with the undo list so as not to accumulate things there
for the intermediate changes that were not really made.

What happens if you click within the current selection?
Does that cancel the selection and move point there?
Perhaps it should; that could help address Miles' complaint:

    For instance, when you select a region, and get the start wrong by 1
    character, you can't just redrag it, you have first get rid of the
    region by clicking (outside the region!), and _then_ redragg.

Re: [usability] mouse-1 for performing actions?

[Simon Josefsson]

Richard Stallman <rms@gnu.org> writes:

> What happens if you click within the current selection?
> Does that cancel the selection and move point there?
> Perhaps it should; that could help address Miles' complaint:
>
>     For instance, when you select a region, and get the start wrong by 1
>     character, you can't just redrag it, you have first get rid of the
>     region by clicking (outside the region!), and _then_ redragg.

Yes, the applications I tried behaved as you thought -- the selection
is canceled if you click anywhere (inside or outside the selection).
But Emacs already behaves this way from what I can tell.

Re: [usability] mouse-1 for performing actions?

[Miles Bader]

Simon Josefsson <jas@extundo.com> writes:
> Yes, the applications I tried behaved as you thought -- the selection
> is canceled if you click anywhere (inside or outside the selection).

The program I usually have problems with is Mozilla, which does the
select-dragging thing -- there, you _can_ click inside the selection to
reduce it to a point, but if your click is a little sloppy and produces
some movement events between the down/up events, it thinks you're trying
to drag instead.

Actually none of the Gnome programs I tried seem to implement `selection
dragging with mouse-1' at all...

-Miles
-- 
`...the Soviet Union was sliding in to an economic collapse so comprehensive
 that in the end its factories produced not goods but bads: finished products
 less valuable than the raw materials they were made from.'  [The Economist]

Re: [usability] mouse-1 for performing actions?

[Richard Stallman]

    Yes, the applications I tried behaved as you thought -- the selection
    is canceled if you click anywhere (inside or outside the selection).
    But Emacs already behaves this way from what I can tell.

Yes, it does, but we were thinking of trying to add this
drag-the-selection feature; the point is that a click in the selection
(or out of it) should deselect it.

Re: [usability] mouse-1 for performing actions?

[Stephen Berman]

On Wed, 22 May 2002 10:44:24 +0200 Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai   Großjohann) wrote:

> Note that clicking mouse3 twice cuts the text.  Very nifty.  By
> default, this feature is not available when dragging mouse-1.
> 
> But with mouse-sel enabled (I think that's the package), you can
> press mouse-1, drag the mouse, then _also_ click mouse-3, then
> release mouse-1, to effect a cut operation.  Even niftier :-)

I wasn't aware of mouse-sel-mode before, but trying it now I don't see
this behavior with Emacs 21.2 and 20.7 on my laptop.  Could this be
because I have three-button emulation enabled under X?  (I.e., when I
simultaneously press mouse1 and mouse3 (using an external PS/2
three-button mouse), and then release either, the region gets yanked a
la mouse2.)

Also, when mouse-sel-mode is enabled, I cannot kill the selected
region by clicking mouse3 a second time, but only by double-clicking
mouse3.  Is this intended?  What's more, when I subsequently disable
mouse-sel-mode again, I still can't kill the selected region by
clicking mouse3 a second time (though -- provided transient-mark-mode
is enabled -- I can kill it by double-clicking mouse3).  I don't know
how to restore the previous behavior (where clicking mouse3 twice
first selects and then kills the region) without killing and
restarting Emacs.  Is there something I'm missing or should I file a
bug report?  (This happens with both Emacs 21.2 and 20.7, both with my
init-file as well as with -q --no-site-file.)

--Steve Berman