Earlier today I found some Gnome 2 packages for Debian woody so I have now installed them.
The first thing I notices is that Gnome 2 bucks the trend of later versions of software to be slower than their predecessors - Gnome 2 is noticably faster than Gnome 1.4.
Then I discovered an interesting little problem. On my desktop machine all works correctly, but on my portable some of the items from the menu that pops up from the Gnome foot are missing, mostly those in the "Desktop Preferences" and "System Tools" categories. As both machine were configured with the same packages this seems very strange.
Does anyone know where the config file and/or directory tree is for those menus in Gnome2 ?
Steve Fosdick fozzy@pelvoux.demon.co.uk wrote:
Then I discovered an interesting little problem. On my desktop machine all works correctly, but on my portable some of the items from the menu that pops up from the Gnome foot are missing, mostly those in the
Maybe the portable hasn't run "update-menus" for some reason? See if it helps. Possibly the gnome2 packages aren't including menu scripts yet.
I find it far more annoying that control-a's default binding has changed from "start of line" in GTK+ 1.x (same as in bash, emacs, ...) to "select all" in GTK+ 2.0 (not used by other common X programs AFAIK). This makes for a very quick way to nuke the contents of the input window when you just wanted to add a bit at the start of the line. It's fixable via a dot file, but that's an amazingly dumb default. :-/
I find it far more annoying that control-a's default binding has changed from "start of line" in GTK+ 1.x (same as in bash, emacs, ...) to "select all" in GTK+ 2.0
<snip>
It's fixable via a dot file, but that's an amazingly dumb default. :-/
Perhaps they're trying to make it less confusing for ppl switching from windoze... Not a dumb default for me, since I didn't know that ctrl-a was a binding for start of the line in emacs etc! ;-)
Ricardo
N3tw0rkAng31B0y ricardo@corez23.com wrote:
Perhaps they're trying to make it less confusing for ppl switching from windoze...
That seems to be exactly what they are doing, from reading their release notes. So screw all of you X users, mmmkay?
Not a dumb default for me, since I didn't know that ctrl-a was a binding for start of the line in emacs etc! ;-)
Do you never use any of the shortcuts (^U, ^K, ^Y, ^T, ^A, ^E) in bash et al? Yikes.
On Mon, 25 Nov 2002 12:04:48 GMT MJ Ray markj@cloaked.freeserve.co.uk wrote:
N3tw0rkAng31B0y ricardo@corez23.com wrote:
Perhaps they're trying to make it less confusing for ppl switching from windoze...
That seems to be exactly what they are doing, from reading their release notes. So screw all of you X users, mmmkay?
Not a dumb default for me, since I didn't know that ctrl-a was a binding for start of the line in emacs etc! ;-)
Do you never use any of the shortcuts (^U, ^K, ^Y, ^T, ^A, ^E) in bash et al? Yikes.
I went for some time without knowing about these. Then I learnt emacs and, of course, before long I would accidentally type the emacs key for something in a completely different app and find it worked! The biggest surprise for me in that respect was FrameMaker (at work).
Steve.
On Sun, 24 Nov 2002, MJ Ray wrote:
I find it far more annoying that control-a's default binding has changed
...
but that's an amazingly dumb default. :-/
CTRL-A selects all in most of my X apps: nedit, galeon, openoffice.org, etc. I'd suggest the "Home" key would be a saner binding for "start of line".
Andrew.
Andrew Savory lists@andrewsavory.com wrote:
CTRL-A selects all in most of my X apps: nedit, galeon, openoffice.org, etc.
Brokenness of other applications is no reason to continue it. Actually, isn't Galeon GNOME/GTK+2, which is what I was complaining about? openoffice.org are still sorting themselves out. nedit should be better ;-)
I'd suggest the "Home" key would be a saner binding for "start of line".
Home is a very poor key for such a common function, being over on the navigation keypad, making you relocate your hands. "Start of Line" is something you want to do a lot while in the flow of typing. "Home" is a lot better for "start of buffer" or "home", depending on app, but the functions which are normally bound in the main keypad: start/end of line; delete backwards and forwards; transpose characters; mark/copy/kill/yank region; are there for a reason.
On Mon, 25 Nov 2002, MJ Ray wrote:
That seems to be exactly what they are doing, from reading their release notes. So screw all of you X users, mmmkay?
Or how about "Microsoft and Apple plough millions into usability studies, perhaps they have a point", or "if 95% of the population expect CTRL-A to select all, perhaps we should be consistent", or "current keybindings date back to limited capability terminals, it's time to move on"?
As long as it's reconfigurable, I don't see a problem.
Brokenness of other applications is no reason to continue it.
Brokenness by what definition?
Actually, isn't Galeon GNOME/GTK+2, which is what I was complaining about?
Galeon is a mozilla derivative, so presumably has the same keybinding as Mozzilla, across _all_ platforms.
Home is a very poor key for such a common function, being over on the navigation keypad, making you relocate your hands.
Only one hand, and only one keypress rather than an RSI-inducing key combo.
"Home" is a lot better for "start of buffer"
CTRL-Home
but the functions which are normally bound in the main keypad: start/end of line; delete backwards and forwards; transpose characters; mark/copy/kill/yank region; are there for a reason.
<comment type="inflammatory common sense"> I'd suggest they are there for *nix hackers and are a legacy from the limited-capability terminal days. They are not friendly to the mass audience. Let's dump the elitism and start doing what usability suggests is best for a change (as Red Hat and Gnome appear to be doing). After all, the hackers can always reconfigure. </comment>
Andrew.
Andrew Savory lists@andrewsavory.com wrote:
Or how about "Microsoft and Apple plough millions into usability studies, perhaps they have a point"
They are not immune from making goofs (and many books have been written about their goofs) and just adopting their decisions ignores our own past work.
or "if 95% of the population expect CTRL-A to select all, perhaps we should be consistent",
That is not true, though.
or "current keybindings date back to limited capability terminals, it's time to move on"?
Some have survived for a reason. Others have not.
As long as it's reconfigurable, I don't see a problem.
Having to reconfigure it everywhere is a real pain and GTK seems not to use the X Resource Database, so it's not even consistent on the one display until you edit everywhere that you run programs. Having the new default cause injury to users used to the old default *of the same application* is also very bad.
Home is a very poor key for such a common function, being over on the navigation keypad, making you relocate your hands.
Only one hand, and only one keypress rather than an RSI-inducing key combo.
Much more RSI is caused by relocating your hands continuously, IIRC. We type combinations in most sentences. Where do you get the idea that control-A runs the risk of RSI more than relocating twice?
"Home" is a lot better for "start of buffer"
CTRL-Home
That's making a simple thing a hard-to-reach combo.
<comment type="inflammatory common sense">
Inflammatory? Yes. Common-sense? No.
There is no usability study evidence to support injuring users. There is hardly any to support ^A as select all, even on Windows. Unless, of course, you know otherwise?
On Mon, 25 Nov 2002, MJ Ray wrote:
Andrew Savory lists@andrewsavory.com wrote:
Or how about "Microsoft and Apple plough millions into usability studies, perhaps they have a point"
They are not immune from making goofs (and many books have been written about their goofs) and just adopting their decisions ignores our own past work.
Oh, ok -- I wasn't aware that there had been extensive usability tests on the keybindings you mentioned?
or "if 95% of the population expect CTRL-A to select all, perhaps we should be consistent",
That is not true, though.
Ok, whatever current percentage use Windows or MacOS and are aware of keyboard shortcuts. It's still very safe to assume "the majority".
Having to reconfigure it everywhere is a real pain and GTK seems not to use the X Resource Database, so it's not even consistent on the one display until you edit everywhere that you run programs. Having the new default cause injury to users used to the old default *of the same application* is also very bad.
For sure it's a pain, and switching bindings in a program is bad. But does that mean we must stick with the errors of yesterday? I don't think so. It would not be difficult to imagine a new version of a program warning the user: "Keyboard shortcuts have changed: <show changes> <use old shortcuts> <use new shortcuts>". Not doing that is the bug, not changing shortcuts to follow "the norm".
Much more RSI is caused by relocating your hands continuously, IIRC. We type combinations in most sentences. Where do you get the idea that control-A runs the risk of RSI more than relocating twice?
Depends on where your hands are when at rest (which I guess depends on whether you've had 'formal' keyboard training or not), and just how often you need "start of line". It's all pretty circumstantial. What may be best for you in emailing/coding won't be best for someone typing essays. What's best for someone typing essays won't be best for someone surfing, etc. All we can hope for is consistency (across platforms) and configurability (for deviations from "the norm").
There is no usability study evidence to support injuring users. There is hardly any to support ^A as select all, even on Windows. Unless, of course, you know otherwise?
Nope, but then I've yet to see any evidence in favour of ^U, ^K, ^Y, ^T, ^A, ^E, either. All I have is the casual observation that to most people I've spoken to, a key labelled "Home" makes MUCH more sense than ^whatever. *That* is the common sense bit.
Andrew.
I usually bind ^A to autodial Pizza Express!! :O)
Keith ____________ BORE, n. A person who talks when you wish him to listen. Ambrose Bierce - The Devil's Dictionary
Andrew Savory lists@andrewsavory.com wrote:
Oh, ok -- I wasn't aware that there had been extensive usability tests on the keybindings you mentioned?
The (obvious?) point was that neither have such backing.
[...]
Much more RSI is caused by relocating your hands continuously, IIRC. We type combinations in most sentences. Where do you get the idea that control-A runs the risk of RSI more than relocating twice?
Depends on where your hands are when at rest (which I guess depends on whether you've had 'formal' keyboard training or not)
Do you rest your hands over the nav' keys?
and just how often you need "start of line". [...]
More than I need to select all, for sure. Control-A isn't far from Shift-A and no-one complains about that being particuarly bad for RSI.
[...] All we can hope for is consistency (across platforms) and configurability (for deviations from "the norm").
This is not the same as "let's junk all our past work and injure existing users," though.
Nope, but then I've yet to see any evidence in favour of ^U, ^K, ^Y, ^T, ^A, ^E, either.
These are fairly language-independent: A is the first letter, moving you to the first letter K are the scissors cutting forwards Y is the pasting tool T is similar to a proof mark for transpose
I admit that I can't remember whether U and E were intended in this pictographic way, or just accidents that caught on. It's not possible to justify "select All" other than in English, is it?
MJR
Facetious remarks aside, the problem with key bindings is that different people have different preferences depending on their past experiences of IT HCIs and their own personal preferences.
IMO well designed software will allow users to set up their own key bindings to suit themselves.
There is an issue about the initial defaults, and there are several different arguments about what these should be but, in the end, they are all arbitrary. For example what about left handed people like me? I use my left hand for the mouse so the left hand end of the keyboard is very inconvenient. I prefer the really old bindings that use Ctrl or Shift in combination with the Insert/Delete/Home/End/Page Up/Page Down block above the cursor movement keys.
I would tend to design software with configurable key bindings and set the initial default to 'none' so that users do not inadvertently use a combination that does something other than they expect. Perhaps also supplying a number of presupplied alternative configurations they can select.
Keith ____________ Do not fear the arising of thoughts - only be concerned lest your awareness of them be tardy. Zen Saying
Andrew Savory lists@andrewsavory.com wrote:
On Mon, 25 Nov 2002, MJ Ray wrote:
That seems to be exactly what they are doing, from reading their release notes. So screw all of you X users, mmmkay?
Or how about "Microsoft and Apple plough millions into usability studies, perhaps they have a point", or "if 95% of the population expect CTRL-A to select all, perhaps we should be consistent", or "current keybindings date back to limited capability terminals, it's time to move on"?
Just because its survived that long in the windows world does not make it a sane keybinding. How often do you actually need to select all? rarely. how often do you want to get to the beginning of a long line, or the end, or delete a word, often.
As long as it's reconfigurable, I don't see a problem.
How about actually because the keybinding should be correct by default and configurable to be broken. Select All is definately a broken default.
Brokenness of other applications is no reason to continue it.
Brokenness by what definition?
Actually, isn't Galeon GNOME/GTK+2, which is what I was complaining about?
Galeon is a mozilla derivative, so presumably has the same keybinding as Mozzilla, across _all_ platforms.
yes, moz does have the same broken keybinding. but how often in a graphical web browser are you wanting to select all? especially in X, when all you're going to get in the clipboard is the text. At the same time, you're probably not going to want to go to the beginning of the line though, so web browsers are not a prime example. And I think that you'll find that the keybindings for moz derivatives can be set by the client app, they only basically use the moz rendering engine.
Home is a very poor key for such a common function, being over on the navigation keypad, making you relocate your hands.
Only one hand, and only one keypress rather than an RSI-inducing key combo.
ROFL. guessing that Shift is also a bad modifier then... and that you don't often change desktops with the keyboard, or have any useful external bindings. Home is HOME, not "Start of line".
"Home" is a lot better for "start of buffer"
CTRL-Home
obviously you travel to the beginning of lines more than the start of the buffer. personally I do both quite often so find it handy when they're bound to sane places.
but the functions which are normally bound in the main keypad: start/end of line; delete backwards and forwards; transpose characters; mark/copy/kill/yank region; are there for a reason.
<comment type="inflammatory common sense"> I'd suggest they are there for *nix hackers and are a legacy from the limited-capability terminal days. They are not friendly to the mass audience. Let's dump the elitism and start doing what usability suggests is best for a change (as Red Hat and Gnome appear to be doing). After all, the hackers can always reconfigure. </comment>
I'd suggest that a majority of nix hackers are 'lazy', I don't mean that in the way of they just don't do anything, I mean that from the point of view that they can do things in as short a time as possible. I'd suggest that if a hacker that is needing these type functions lots of the time, finds that these key combos are useful, then maybe they actually are. 30+ years of use would suggest that they are far from non intuitive, and once learnt a great asset. Just because Billy Boy assigned a different set of keys for things, and carried that on through the series, does not mean that he was right in the first place. Familiarity is the only thing driving the argument, its certainly not common sense. There are by sheer marketing power more windows users out there than nux users. If everyone changes all the nux keybindings to be "Windows User Friendly" that's 30+ years of research and use of certain keybindings down the toilet. That's bloody ridiculous.
Oh well,
Just my 2ps worth...
(not that andrew will bother reading it)
Cheers,
Brett.
Andrew Savory wrote:
Or how about "Microsoft and Apple plough millions into usability studies, perhaps they have a point", or "if 95% of the population expect CTRL-A to select all, perhaps we should be consistent", or "current keybindings date back to limited capability terminals, it's time to move on"?
As long as it's reconfigurable, I don't see a problem.
<snip>
Galeon is a mozilla derivative, so presumably has the same keybinding as Mozzilla, across _all_ platforms.
mozilla mail so I just noticed uses Ctrl+a as the beginning of line shortcut key (at least it does on linux) and most certainly doesn't select all. Alt+a selects all in moz-mail.
Adam
On Sun, 24 Nov 2002 22:30:21 GMT MJ Ray markj@cloaked.freeserve.co.uk wrote:
Steve Fosdick fozzy@pelvoux.demon.co.uk wrote:
Then I discovered an interesting little problem. On my desktop machine all works correctly, but on my portable some of the items from the menu that pops up from the Gnome foot are missing, mostly those in the
Maybe the portable hasn't run "update-menus" for some reason? See if it helps. Possibly the gnome2 packages aren't including menu scripts yet.
Thanks.
I was pointed to a solution by Gustavo Noronha Silva, the debian guy who did the back port. The cause was left-over junk in the /etc/gnome-vfs-2.0/vfolders directory and the solution was to remove all files in that directory that were not part of a debian package by running:
dpkg -S /etc/gnome-vfs-2.0/vfolders/*
then removing all those files about which dpkg complains.
I find it far more annoying that control-a's default binding has changed from "start of line" in GTK+ 1.x (same as in bash, emacs, ...) to "select all" in GTK+ 2.0 (not used by other common X programs AFAIK). This makes for a very quick way to nuke the contents of the input window when you just wanted to add a bit at the start of the line. It's fixable via a dot file, but that's an amazingly dumb default. :-/
I saw that Gnome 2 was supposed to have benefited from extensive user interface research. I hope that doesn't mean simply copying windows in many respects because, expensive research or not, there are many features of windows which do not make for a convenient interface. A couple that spring to mind - my pet hates are:
* Word oriented select - where a hand-dragged selection gets extended to a word boundary.
* Text boxes which don't allow the insert point to be positioned with the first click of the mouse but which select the whole text instead.
I have noticed the first of these mis-features somewhere in Gnome 2 though I forget where. I like the feature of gnome-terminal and others to select a whole word by double clicking but that was there in Gnome 1.4 too. If I drag by hand I want the selection to end where I pointed even it is is half a word as the next thing I am likely to do is paste the selection in somewhere and the change the way that word ended.
Steve.
-
Adam Bower -
Andrew Savory -
Brett Parker -
Keith Watson -
MJ Ray -
N3tw0rkAng31B0y -
Steve Fosdick