You can clone with
HTTPS or Subversion.
I'm still at a loss how could an interface for a maneuver node planner work in the first place when all you have is cursor buttons at best.
Maybe like this?
It's tough to sort out the interface with only cursor buttons. It has to be two levels (one to select from the three axes prograde, normal, radial; the other to alter the value. Maybe an alternative to the step control is to have the left/right buttons move the cursor across columns, and the up/down buttons to change the value of the current column. It'd still be pretty klunky to use, but I don't see a better alternative with what we have available right now.
Alter the value can be accomplished without an extra submenu if we have left-right (or +/- or whatever) buttons. They can also be used to select the current step.
We can have left-right buttons, the new prop in ALCOR has them and we can recycle page buttons into globals even with the current prop. :)
I.e. imagine a table:
Up and down buttons move you along the lines of the table, +/-/enter act upon a line.
Is is not possible to have a numpad to interface with the display?
It is. Do you actually want to click it and do we actually want to code cursors and associated messy low-level console support?
I'm going to make a kOS interface plugin using the new prop in ALCOR, and I'm betting you nobody will ever use it to actually type in anything.
The difficulty I see with this arrangement is there are many button presses involved in changing the step. The alternative I had in mind incorporates the step changes on the editor level in the menu, provided we have the additional buttons (we would need six - four cursor keys, plus Enter & Cancel). I don't think I can illustrate with tables very effectively, so maybe I can describe it:
At the top level are all of the entries in your menu, except "Step". The up/down buttons highlight an entry, like the current menus do. Pressing "enter" will activate the row.
For the case of "Snap To", the up/down arrows now cycle between the snap-to options. Pressing "Enter" applies the current option; pressing "Cancel" discards the option.
For the time and vector editors, the most significant digit is highlighted to indicate the current cursor position. Up/Down will change the value under the cursor (cycling 0-9 with no carry when rolling from 9 to 0). The left/right keys will move the cursor position. Pressing "Enter" at any time applies the currently displayed value, while "Cancel" will revert it to its value prior to editing.
I'm not sure that it will be possible to have a good user experience from either of these - we can probably get something that's workable, but it's going to be cumbersome.
The other option is to follow the A10C's setup and add a UFC panel for number entry. It's never going to be pretty, you probably need some kind of dpad type thing . And you have to decide on an active MFD.
The UFC panel feeds numbers etc. to the SOI (selected) MFD
I think a functional implementation of the mechjeb manoeuvre planner would be better all round. Enter what you want to achieve rather than how to achieve it.
Manouver type, and when to do it and out pops a node.
In the meantime, it might provide a lot of desired function (without loads of clicking) to replicate function in the maneuver node editor and rendezvouz planner, specifically hohmann transfer to target, fine tune, transfer to another planet, match velocity at closest approach, get closer, match planes (at AN and DN)
I agree with this...
"I think a functional implementation of the mechjeb manoeuvre planner would be better all round. Enter what you want to achieve rather than how to achieve it.
Manouver type, and when to do it and out pops a node."
I imagine a screen very similar to the current Autopilot screen but with a few items that can be drilled down a bit more.