Page 1 of 2

why does the axis to move something towards you not work ?

Posted: Wed Mar 02, 2005 3:53 pm
by Money_YaY!
If I am in front view and press g and an axis constraint all work except the axis to move towards your view or away from your view.

Example: front view, Z and X work but not y. I find this very inconvenient, when logicaly I would expect it to look like it is resizeing but in fact is comeing closer or further,, but I still would know this espicially in perspective view.

Posted: Wed Mar 02, 2005 4:05 pm
by SamAdam
i think it can do it, but lets see your cursor move out of the screen!
how would you control it? move the pointer right to come closer or farther away?

Posted: Wed Mar 02, 2005 4:26 pm
by theeth
Like SamAdam said, it is quite logical that you can't do it since you can only move your mouse cursor on the screen, it can't represent a motion on an axis that is straight out of the screen.

That's like compaining that since you can't use a pen to draw a line that points straight out of a paper sheet, it must be the pen's fault...

Martin

Posted: Wed Mar 02, 2005 4:30 pm
by Arathorn
Would it be possible to fake this by letting moving the mouse down and up correspond with moving the mouse "out of and into" the monitor?

Posted: Wed Mar 02, 2005 4:45 pm
by theeth
Arathorn: Everything is possible, the question is: would it be consistent?
And the answer is: no.
If that's done only for axis going out of the screen, it has no consistence whatsoever with the rest. And if all axis constraints work that way, it looses all its logical interactivity (move mouse, object goes there).

Martin

Posted: Wed Mar 02, 2005 5:39 pm
by Money_YaY!
??? eh ? I cant even follow you ideas of this. pencil something

twist the view just a little with middle mouse click and it works fine albeit a little slow but it works. It is just a silly axis what is so wrong about giving a user the power to move something on the axis even if it is in front of there face ?

further more, blender is not a 2d app it is a 3d app. There for if I move something on an axis i would expect it to move on that axis as it does for the other two. I mean even size works, so why not move ?

Posted: Wed Mar 02, 2005 5:55 pm
by theeth
Ok, I'll try to simplify again.

The transform system is based on the concept that the selection moves like the mouse cursor moves.

When constraints are involved, it follows the mouse motion as best. For planar constraints, that means moving the object along the plane mapping the 2D motion of the mouse (if you start with the cursor at the center of the object, it will follow it closely when you move the mouse). For axis constraints, it's the same thing, it tries to find the closest point on the axis as it is displayed on the screen.

That's where the problem lies. If the axis is exactly perpendicular, it's a single point on screen. There's no way to know where you want to move it on that axis, since the whole axis is always nearest to the cursor at once.

Uhm.. I feel like that wasn't as simple as I thought it would be.

Martin

Posted: Wed Mar 02, 2005 6:04 pm
by MNME
theres always the mousewheel.... i know its steps.. no real fine control, but could do 1 unit or 0.1 with shift, 5 and 1 degree and 0.1 and 0.01 size (same as mouse move with ctrl held)

Posted: Wed Mar 02, 2005 6:04 pm
by Money_YaY!
theeth wrote:Ok, I'll try to simplify again.

The transform system is based on the concept that the selection moves like the mouse cursor moves.

When constraints are involved, it follows the mouse motion as best. For planar constraints, that means moving the object along the plane mapping the 2D motion of the mouse (if you start with the cursor at the center of the object, it will follow it closely when you move the mouse). For axis constraints, it's the same thing, it tries to find the closest point on the axis as it is displayed on the screen.

That's where the problem lies. If the axis is exactly perpendicular, it's a single point on screen. There's no way to know where you want to move it on that axis, since the whole axis is always nearest to the cursor at once.

Uhm.. I feel like that wasn't as simple as I thought it would be.

Martin
ok this one made way more sence thank you. I am still however perplexed then why does resizeing the mesh on each axis work. It defies what you just explained to me.

Posted: Wed Mar 02, 2005 6:18 pm
by Money_YaY!
MNME wrote:theres always the mousewheel.... i know its steps.. no real fine control, but could do 1 unit or 0.1 with shift, 5 and 1 degree and 0.1 and 0.01 size (same as mouse move with ctrl held)
hell I would be happy with that

Posted: Wed Mar 02, 2005 6:21 pm
by theeth
Money_YaY!: I explained what it does for Translations (Grab). For Resize, the pointer motion is mapped differently before applying the constraint so the restriction does not apply. I won't go into the differences since there's a very high chance that you won't get it. Others can just look at the code.

MNME: Mousewheel is used to resize the PET area of influence.

Martin

Posted: Wed Mar 02, 2005 6:32 pm
by Money_YaY!
theeth wrote:Money_YaY!: I explained what it does for Translations (Grab). For Resize, the pointer motion is mapped differently before applying the constraint so the restriction does not apply. I won't go into the differences since there's a very high chance that you won't get it. Others can just look at the code.

MNME: Mousewheel is used to resize the PET area of influence.

Martin
ok ok i understand... Sooo as an idea then, someone could remap the one axis method of resize to that move axis then eh... whatever I asked a question and gave a request that is all I can do, from a user point of view it is incorrect feedback on the 3d view. From a coders view it is logic. You know where I stand...

Posted: Wed Mar 02, 2005 6:41 pm
by theeth
Money_YaY! wrote: ok ok i understand... Sooo as an idea then, someone could remap the one axis method of resize to that move axis then eh...
err no, that just wouldn't work for obvious reasons.
Money_YaY! wrote:whatever I asked a question and gave a request that is all I can do, from a user point of view it is incorrect feedback on the 3d view. From a coders view it is logic. You know where I stand...
From a logic point of view period it makes no sense. That's like trying to draw a line that goes outside of a sheet of paper (towards you).

Actually, it does work *a bit* in the latest cvs version (in what I have here anyway, maybe not all is commited).

Martin

Posted: Wed Mar 02, 2005 9:41 pm
by Delta
i still dont see how this can work.

if measurement is relative to what we see, we cant gauge how far we are moving the object (unless you're in perspective mode)

even then, how will you be able to accurately distinguish if youve moved your object 0.5 blender units or 500 blender units?

Regards,
~Delta

Posted: Thu Mar 03, 2005 12:13 am
by Money_YaY!
Delta wrote:i still dont see how this can work.

if measurement is relative to what we see, we cant gauge how far we are moving the object (unless you're in perspective mode)

even then, how will you be able to accurately distinguish if youve moved your object 0.5 blender units or 500 blender units?

Regards,
~Delta
For something like an armature it doea really matter to have this, since the arm will not go further than it can go but it is acward to have to align the arm twice in top and side view to just move a hand forward. Whe the front view would be perfect.