this is a collection of feature requests that I think would be usefull. I know that some of them are posted already, but I want to post them again as a <<ME TOO>>.
1.When rendering halos it is possible to render rings and lines OVER the halo, but it is not possible to switch the halo off and only render rings and/or lines. Would be cool being able to switch the halo of just like the lines or rings.
2.A grabbing/scaling relative to the faces direction (like X/Y/Z after a grab/scale relative to the faces direction). Directly after an extrusion (no grab/scale/rotate) the shrinken/fatten moves the face without scaling it. But when the face is already a bit away, it is scaled. My idea would be to use the XYZ keys to tell the shrinken fatten command in what direction (relative to the faces direction) to move. Perhaps this feature is allready implemented but I didn't found it, neither in blender, nor in the manual or the forums. A combination with number 4 would be ultra cool.
3.An extrude along path function for meshes not for nurbs or beziers. I know Dupliframes and Dupliverts but they(as far as I know) don't extrude the first object but just duplicate the mesh and place the copies along a path.
4.A scale/rotate individual faces. When more than one face is selected, blender rotates/scales around the center of all these faces. A feature that rotates/scales each face individually would be cool.
5.When grabbing/scaling/rotating the mouse still moves on the screen, so when the mouse hits the sides, top or bottom of the screen you have to finish scaling/grabbing/rotating and move the mouse again to the middle of the screen and start again. A cool solution would be to stick the mouse to the middle of the screen when scaling/rotating/grabbing. The UnrealEd (Unreal Editor http://udn.epicgames.com/Two/UnrealEngi ... me22262001) for example already has this feature. This would speed up work flow extremely.
6.Scaling two axis at once. When pressing X/Y/Z it is possible to scale one axis but when I want to scale along two, for example along X and Y but not Z, I have to either scale X and Y individualy or scale all and then Z back. Perhaps like pressing Z once scales only the Z axis pressing it twice scales X and Y but not Z.
7.In the transform properties it is possible to enter exact values. I think three more fields for entering absolute X/Y/Z values would be usefull. For example when moving certain points to the Z-Axis for rotation
. Or when I want to get some vertices of an object to the same height.
8.When two objects are joined to one or even when verticies are duplicated and seperated from a mesh (no connecting faces) but still in the same object the Catmul-Clark Subsurf function subsurfs the objects to the individual, means to two different object centers. I think a feature to subsurf them to the same object center would be usefull. I think this this is hard to implement, but it would be very usefull.



without subsurf, with subsurf, subsurf method I proposed
The following points arose from the idea of enriching Blender with more Game Level Editor like functions.
9.The possibility of absolute viewport movement, something like in 1st-person camera games. The mouse moves the view direction and the arrow keys move the view forward, backward, left and right. I know about the
Shift-F function when the camera is selected. But that function only works when the viewport is switched to camera view and the camera is selected. And this movement is not really usable for moving through level like areals. The zoom function is not usable for that at all, it only zooms in to a certain degree and not further. It is perfect for moddeling a single object, but not for creating levels because they are made up of a lot of different objects. Perhaps when the middle mouse button is pressed the arrow keys work as movement keys. A real turntable viewrotation would be usefull in this case.
10.The possibllity to turn off free object movement, means forced snap to grid. This is very usefull when creating the base structure of a level. Like in GTK-Radiant (the official quake-engine map editor http://www.qeradiant.com), where all cuboid objects are created just by pulling the mouse over the grid (which is always horizontal) and the cube occurs with the length and width the mouse pulled but always snaped to the grid. A forced snap to grid is very important when creating the basic structure of a level, because the structuring objects (outer hull of level) have to enclose the level without a gap. It is easily much more simple to create a level with a fixed snap to grid. Also the snap to grid should be adjustable, means to which level of gridlines should be snaped to, perhaps adjusted by the zoom factor (more zoomed out means snap to bigger grid, zoomed in means snap to tight grid).
11.A boolean operation that considers the insides of a mesh. When two cubes are aligned with their faces touching, the boolean union sometimes keeps parts of the inner faces or creates new ones inside. This makes the boolean operation useless for creating the outer hull of a level out of different objects by unioning them and selecting onesided and recalculate normals inside. This method would be so easy and even much better than in GTK-Radiant.
12.The decimator tool should have the possibility to automatically decimate the mesh until it changes shape. This would be very usefull after a boolean operation, which creates alot more vertecies than necessary.
Hope everything is understandable, if not post. Please give critics, solutions and ideas.