Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[eba]

At the moment, MC's standard view doesn't ever go quite full screen, i.e. if you click in the very top right corner, rather than clicking on MC, you'll click on whatever's behind it.  I often find this means that I'll close whatever is behind MC rather than MC itself, as in the case just now I lost my Interact forum reply.

This never happens with standard windows!  If they are full screen, you cannot click on anything behind without minimizing that window first (or otherwise switching to the other window).

I tried editing the skin myself some time ago to move the close button, but was no good...right in the Top right corner is a permanently non-MC area...and the same with the top left.

Thanks

[gappie]

yeah.. lol.. you are right. never happend to me, but you are right.

but you are lucky today, my new skin does not have that problem (i guess because it has side borders from just 1 pix.  ). its on the third party board..


 
gab

[eba]

my new skin does not have that problem

Really?  I thought it was impossible to change via skinning, as like I said I'd tried moving the buttons myself...but moving the actual borders was a tad complicated for me!

Heading over to try your skin now
(and was going to say before but didn't actually get round to posting, I do love the bikes )

[gappie]

thanks..

i just looked a bit at the noire skin. its not the side borders, but the small magenta spots in the upper left and right corner of the mainframe_border.bmp. when i make those two black, the problem disappears.

 
gab

[Matt]

I wonder if the border shape should be ignored, and filled with black when maximized?  This way, we could catch a click at the top right.

[eba]

That sounds like it would make sense

Personally I think it should actually catch it as the close button too (as do standard windows) but doesn't really matter.

Didn't realise it was actually so few skins that it actually affected, I thought it was all of them, but now testing a few it's only the few with curved corners.

Cheers

[MrHaugen]

That was funny. I have thought of it before, but never actually tested it.

In fullscreen I would very much prefer if the skin filled the whole screen. If you have a very dark skin and something white behind MC, the 1 or 3 pixels is very noticable!
When it's not in fullscreen I would prefer to have the rounded corners. Just looks a bit prettier. So, ignoring border shape at maximised, and fill with color sounds good.

[eba]

Bump

[Matt]

In build 103 and later:
Changed: When maximized, any rounded edges of a skin are removed and filled with solid black so that clicking in the screen corner won't send the click to a window behind Media Center.

[eba]

Thank you

[ZoFreX]

This has always been the most annoying aspect of Media Center for me, so I'm pleased to see this getting attention but feel that this fix is unsatisfactory.

1. (minor) It's not actually necessary to fill the corners black, I don't think [verification needed ], to make them "block" clicks, and maybe skin designers wouldn't want this to happen always?

2. (major) There's something in human computer interaction study called "Fitt's Law" which is highly regarded and briefly, explains how long it takes a user to move to a target, based on how far away they are and how big the target is. The upshot of this law is that moving to targets at the edge of the screen, or in a corner, can be done in effectively zero seconds. Windows leverages this to give quick access to the close button on applications and the start button (note that in Windows Vista & 7, the start button does not actually extend to the bottom left of the screen, yet clicking the bottom left pixel still activates it).

For this reason I think the top-right-most pixel in MC should be registered as a close button when the application is maximised, so users get back the expected functionality: Moving the mouse to the top right of the screen and clicking closes the application.

[MrHaugen]

1. (minor) It's not actually necessary to fill the corners black, I don't think [verification needed ], to make them "block" clicks, and maybe skin designers wouldn't want this to happen always?

I have to disagree. Why people would want rounded cornes in full screen I can't understand. To fill in a solid like black might not allways be the best choise, but it's a good quick fix solution. Ofcource, it would be better to fill it with the color of the actual skin, but that might be hard to do. Hardly any skins are so light that the black pixel(s) will be visible. But if you have a black skin, and a white desktop backgroud, it shows really well. That's why black is the best color when a fixed one have to be chosen.

[ZoFreX]

Well, I think skin designers might want things like that left up to them. I don't know, I'm not a skin designer!

On that front though, this problem could also be fixed by people not making skins where the top right isn't a close button.

In case anyone is interested in why I'm so insistent that the top right corner should be the close button, I've whipped up a quick explanation of Fitt's law, primarily for interface & skin designers I guess:

Fitt's Law
Fitt's Law is, according to Wikipedia:

"a model of human movement which predicts the time
required to rapidly move to a target area, as a function of the
distance to the target and the size of the target"

Wikipedia also gives the mathematical equation which represents it:

Well that's great ZoFreX, but what the hell does that mean?

Well, let's first simplify the equation. a and b are constants that are fit to the data (i.e. these are modified to represent the person / input device / how drunk the person is), so let's just get rid of them. If x is bigger than y then log(x) is bigger than log(y) so for understanding the impact of D and W, we can get rid of the logarithm for now (we will revisit it later).

So now we have:

This is completely wrong mathematically of course. But all we need to know is, that as D gets bigger, T gets bigger, and as W gets bigger, T gets smaller.

But lets get rid of the equation completely, because they rarely help actually understand something. What we have is simple: The further you need to travel, the longer it takes, and the bigger the target you need to hit, the less time it takes. Wow, that turned out to be a pretty obvious law. Let's take a look at it in practice:

^{Fig 1: About to move the mouse cursor to the button marked "target"}

^{Fig 2: The distance to the target and width of the target (width measured against the axis of motion)}

The red arrow is the distance, and the fuschia arrow is the "width" of the target. It makes sense that if we're closer, we will take less time, and if the target is bigger, it's easier to hit.

That's Very Interesting, But What Does That Have To Do With Buttons In The Corner Of The Screen?
Well, consider a target that's at the very edge of the screen:

^{Mouse cursor poised to move to the top right corner of the screen}

^{Distance to target}

The distance is straightforward, but how wide is the target? As we're headed to a target area 1 pixel big, you might be tempted to say W = 1, but you'd be wrong. Instead of thinking of W as width, think of it as error tolerance: If the user aims for the center of the target, then W/2 is the amount they can mess up by. Obviously, larger error tolerance = faster user operation, as they can be less precise. Well, how far wrong can you go and still be at the edge of the screen? As far as you like! You can keep moving the pointer and it goes nowhere, so W is effectively equal to infinity:

^{Target width is effectively infinite}

What does this imply for how fast the user can get there? Well, we showed earlier that the larger the target area, the faster they can move. You can't get larger than infinity, and so it is faster to move to the edge of the screen than to any other position on the screen.

Appendix A: You said you would revisit that logarithm later
Yes. When you see a binary logarithm in algorithms, you know the problem is being divided up in a 50/50 strategy. For an example, the way computers (and, on an interesting HCI note, humans!) search through a list of objects that is in some order (say, 1, 3, 4, 5, 7, 12, 24, 25, 30). If you want to find the number 25, it would be ineffecient to start at the first item in the list and work your way through. Instead:

Jump to the middle of the list Is the number there what we're looking for? If yes, we're done! If the number is less than what we're looking for, then all the numbers higher than this form the list we want to search in If the number is more than what we're looking for, then all the numbers higher than this form the list we want to search in Repeat on this new list

In Fitt's law, this division comes in how users move the mouse. While the distance is still large, they move quickly. As they get closer to the target, they slow down to gain accuracy. This can be crudely thought of as, in each given time period, the user halving their distance to the target:

Appendix B: Further Reading
Wikipedia article on Fitt's Law - In depth, well written, and has a much tighter mathematical demonstration of where the logarithm comes from

I hope that's helpful / interesting to someone, and if you have any questions just ask

[rick.ca]

Wow. I'm sold. But shouldn't we be using that corner for something more useful, like a mathematical expression editor or something?