Toolbox Flash Simulation, R2

The interface, modeling, 3d editing tools, import/export, feature requests, etc

Moderators: jesterKing, stiv

alt

Post by alt »

Nice work!

but..

a)
Top-level menus are in strange angles. It could be easier if they were arranged in compass-like fashion (as said before).

b)
When list-menus pop up they hide their parent menu. It is not fun since it makes mousing slow hiding other options user may want to look at. These menus should open outside radial menu and not over it. Maybe they could open centered in same direction as their parent is in radial menu. This way list menu would grow in two directions halving (on my planet) selection time.

ding-a-ling

odyssey
Posts: 18
Joined: Wed Oct 23, 2002 5:58 am
Location: 3rd Rock from the Sun

Post by odyssey »

I like the idea.
'Passed it around for show to the others in the group and we all agree.

Well done. :D

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Post by thorwil »

To bring this up again:
Second Flash version.
Addressed layout and top level menu label issues.

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Post by thorwil »

I'm sorry for this, but there is something wrong with this forum and I'm testing.

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Post by thorwil »

testing 123

Pablosbrain
Posts: 254
Joined: Wed Jan 28, 2004 7:39 pm

Post by Pablosbrain »

I dislike the change from the top level radial layout to the submenu linear layout. It's just not consistent. I would rather see something more similar to the way maya does the radial menu if you are going to use any sort of radial menu design.

And as for the advantage comment on your site "It's an original design, like no other application has ;-)"... being original isn't always good... just as copying something else isn't always good. Lets make sure we learn from what the others have done instead of trying to venture out and learn from scratch ourselves. Just a note.. take it with a grain of salt... or pepper or whatever.

LukeW
Posts: 0
Joined: Mon Mar 03, 2003 1:14 pm

Vertically Centered Submenus

Post by LukeW »

This is about something I was talking to thorwil about in an email:

My first email was about my suggestion to have vertically centered submenus, rather than some that begin from the top, and some beginning from the bottom.

Here are some of his replies and my replies:
-------------------------------------------------------------

> The average vertical distance in reality could only be halved by
> putting most used items in the middle.
My point is not that the average vertical distance is halved, it is that it is substantially reduced... e.g. to about 2/3 or something. (In my first email I gave examples which gave exact figures, though it assumed that the items were used equally frequently)

>But that would be very bad for scanning the lists.
I think people can get used to it quite easily, and also, I think it is easier to navigate using a vertically centered submenu...

e.g. say there are 7 items.
non-centered:
1 <- (the start of the submenu)
2
3
4
5
6
7

centered:
1
2
3
4 <- (the start of the submenu)
5
6
7

Now let's assume that the toolbox is always consistent - so item 2 is always in the same place. Let's say you wanted to go to item 5 a lot. If the menu wasn't vertically centered, you could estimate where item 5 is but you might have to read it to check to see if it is the right item. When the submenu is centered, it is easier to find item 5. In the centered menu (with 7 items), all of the items are at most 1 position away from a "landmark". Items 1 and 7 are at the ends. Items 2 and 6 are one item away from the ends. Items 3 and 5 are one item away from the middle of the submenu were the triangle and the highlighted parent menu item is, and item 4 is directly in line with the triangle. On the other hand, in the non-centered menu, some of the items are quite a distance away from "landmarks" (maybe there is a better term) which means that users have to pay more attention and it would slow them down a bit. (One of Blender's aims seems to be about productivity for hardcore users after all)
A similar concept is this:
Image
Basically the icons here are put into groups (non-logical ones in this example) which are fairly visually distinct. Users wouldn't have to rely on recognizing the icons very much... they could just go for the middle icon in the second group... or the last icon in the third group... rather than looking at the approximate place in the series and trying to identify the icon amongst 2 or 3 that are in that approximate place.

> If we asume the lists can be fully optimized by ranking the items
> by frequency of use, then having the list extend in one direction
> should be optimal. Only with many items of nearly equal frequency of use
> there would be an improvment by using vertical centered submenus.
I think you're wrong about that.
For an example, the frequencies will vary a lot. (none are nearly equal)
Say there are 5 items and the trial involved 31 uses of the menu.
item 1 - 16 times <-
item 2 - 8 times
item 3 - 4 times
item 4 - 2 times
item 5 - 1 time

This is the layout if it was vertically centered:
item 4 - 2 times
item 2 - 8 times
item 1 - 16 times <-
item 3 - 4 times
item 5 - 1 time

Now the frequencies with the distance from the start of the menu:
0 | 16 <-
1 | 8
2 | 4
3 | 2
4 | 1
If you multiply and add those things together you get... 0+8+8+6+4 = 26. So the average distance is 26/31 = 0.84

The vertically centered menu:
2 | 2
1 | 8
0 | 16 <-
1 | 4
2 | 1
If you multiply and add those things together you get... 4+8+0+4+2 = 18. So the average distance is 18/31 = 0.58

And if the frequencies were more similar there might be an even better distance reduction when using a vertically centered menu.

I guess there is an exception to the rule though...
say there were 2 menu items....
(the distances to the submenu start)
non-centered
0 (used 99% of the time) <-
1 (used 1% of the time)

centered
0.5 (used 99% of the time)
<-
0.5 (used 1% of the time)

Now for the non-centered submenu, the average distance is 0.01. For the centered menu, the average distance is 0.5. I didn't take horizontal mouse movement into account though, which would change the results a bit.

Anyway, except for that example I think vertically centered submenus would always have a lesser (or maybe the same?) average vertical distance.

And due to more visual cues or "landmarks" (there would now be 3 rather than 2) it would be easier for experienced users to rapidly navigate the toolbox. And I thought that's what blender is partly about - speed for experienced users.

> My current implementation cuts down on space usage, the proposed
> no-direction-changing would eliminate abrupt direction changes.
> Verticaly centered submenus would do neither.
> You know, I have already played with vertical centering of submenus.
ok. But maybe the speed advantages for experienced users (which I think I've explained) when having vertically centered submenus is more important than those things.

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Post by thorwil »

Pablosbrain wrote:I dislike the change from the top level radial layout to the submenu linear layout. It's just not consistent. I would rather see something more similar to the way maya does the radial menu if you are going to use any sort of radial menu design.
It's just one special case for the top level. Standart menubars use a special case for top level, too, and nobody complains. You imediatley see how it works. Now you dislike that change, but you should ask yourself how this change affects the user. Have you tested the menu? Did it work for you, or not? That is what's realy interesting.

There was already much talk about why it isn't all radial. But feel free to propose a menu organization that would allow it to be radial ...
And as for the advantage comment on your site "It's an original design, like no other application has ;-)"... being original isn't always good... just as copying something else isn't always good. Lets make sure we learn from what the others have done instead of trying to venture out and learn from scratch ourselves. Just a note.. take it with a grain of salt... or pepper or whatever.
I put that smiley behind it because it wasn't meant to be taken all to serious.

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Re: Vertically Centered Submenus

Post by thorwil »

Luke, about vertically centered submenus:

There can be only an advantage if you organize the items acording to frequency just like in your example.

The result will seem to be illogical to the user, a complete mess.

You normaly read/scan lists in one go from top to bottom. Reading/scanning reverse direction is acceptable if you have a good reason, I think. But for my design there's the option to have thirdlevel menus always pointing down, if this should be a problem.

With v-centered submenus I'm pretty sure the user will start to scan the list at the top. So he will loose time if he's going for an often used item.

And your eyes will initialy follow the mouse cursor. But with Ton's new Toolbox I find myself always searching for the top of the submenu. There's a conflict, what realy happens can only be found out with eye-tracking.

This problems affects my current design, too, but to a much lesser extent, I think.

However, thank you for your efforts and the calculations. It's interesting nontheless. Perhaps the menu engine system will allow to test things like this with less effort (and by more people).

But I would still like to read how others think about this (I sure hope some people do read even the long posts...)

Carnivore
Posts: 0
Joined: Thu Jan 30, 2003 6:27 pm

Post by Carnivore »

I didn't read through anything in this thread regarding the 2nd concept, but I hate it. I think that your first concept was great, even if ti did need a little reorganising. This new concept scared me instantly. It's much harder to select anything if you miss by a millimeter (you have to think!!) and there is no clear definition of where one menu subject begins and another ends. Bring back the original concept...

Just my 20 millicents

thorwil
Posts: 0
Joined: Tue Sep 09, 2003 10:30 am

Post by thorwil »

Carnivore wrote:I didn't read through anything in this thread regarding the 2nd concept, but I hate it. I think that your first concept was great, even if ti did need a little reorganising.
The old design has the problem that some people just go for the labels and don't recognize the sections. (At least that was my impression).

That's why I switched to visualizing the optimal angles instead of the sections. And I placed the labels more close to the center, so that when you aim at the labels the chances are high that you don't pass the active area of other items.

The underlying scheme hasn't changed. It's realy strange that you now have problems you didn't have before.

Personaly I think the new Toolbox is easier to navigate. But the problem is, since I designed it, I'm the wrong person to evaluate it.

What do the others think: Was the old design showing the sections better?
New:
http://wrstud.urz.uni-wuppertal.de/~ka0 ... i/toolbox/
Old:
http://wrstud.urz.uni-wuppertal.de/~ka0 ... x/old.html

Please ignore the differences in submenu placement, since that's an idependent issue.

LukeW
Posts: 0
Joined: Mon Mar 03, 2003 1:14 pm

Re: Vertically Centered Submenus

Post by LukeW »

thorwil wrote:Luke, about vertically centered submenus:

There can be only an advantage if you organize the items acording to frequency just like in your example.

The result will seem to be illogical to the user, a complete mess.
Not if the top, middle and bottom are logical groups... so of the items that are rarely used, they could be shared between the top and the bottom so that the ones at the top seem to go together and same with the ones at the bottom - rather than the decision as to whether they go at the top or bottom be completely random for every rarely used item.
You normaly read/scan lists in one go from top to bottom. Reading/scanning reverse direction is acceptable if you have a good reason, I think. But for my design there's the option to have thirdlevel menus always pointing down, if this should be a problem.

With v-centered submenus I'm pretty sure the user will start to scan the list at the top. So he will loose time if he's going for an often used item.
No, if they use the item a lot they wouldn't need to scan through the whole list, beginning from the top... they'd just remember the approximate place on the list... well assuming they are fairly intelligent they would. With y-centered menus it would be easier to find some of the items compared to non-y-centered menus since there are more visual cues that the person can use to rapidly orientate themselves (to get their bearings)... there is always a distinct top, bottom and center. For non-centered menus, the center isn't distinct (assuming there are no additional lines in the menu). When the person knows where to expect the item (which would be in exactly the same place every time - unless you keep on reorganising the menu orders due to frequency for some crazy reason)... so they'd just jump to the center or top or bottom of the menu and look around the appropriate place.
And your eyes will initialy follow the mouse cursor. But with Ton's new Toolbox I find myself always searching for the top of the submenu. There's a conflict, what realy happens can only be found out with eye-tracking.
It isn't that hard to find the top of the submenu... and your eyes would travel about the same distance. If the submenu began from the top and you wanted to look at everything, you'd begin looking at your mouse at the top, then read through the menu, top to bottom, then go back to looking at the mouse pointer. If it is y-centered you'd begin looking at the center, then read from top to bottom, then look back at the center. So assuming you don't move the mouse, your eyes are moving the same distance. If you need to use the mouse to read things you'd move the mouse more when things are y-centered if you wanted to move from the top to the bottom.
However, thank you for your efforts and the calculations. It's interesting nontheless. Perhaps the menu engine system will allow to test things like this with less effort (and by more people).
This menu engine system sounds interesting... it would be good if it is flexible enough to do things like vertically-centered submenus... while not being to difficult to use (I'm not really good enough to be able to compile Blender).
But I would still like to read how others think about this (I sure hope some people do read even the long posts...)
Well I think your latest toolbox is better than the earlier ones.

Post Reply