Space station expansion contract leaving station together with my shuttle

[PT]

So I have put a space station around Kerbin, then I got contract to upgrade its capacity. Easy money.

I picked up contract, then for unrelated reasons I undocked my shuttle and sent it for orbital voyage.

And then I see contract attached to shuttle instead of core station, so I had to wait for it to return to dock with extended station...

What steps should I take to ensure that contract will be attached to station core instead of one of its (multiple) modules? (Yea, multiple modules, now I need to wait for one of them to return from Minmus mission...)

[Snark]

5 hours ago, PT said:

What steps should I take to ensure that contract will be attached to station core instead of one of its (multiple) modules? (Yea, multiple modules, now I need to wait for one of them to return from Minmus mission...)

My guess would be, probably, "don't have the modules docked when the contract pops up."    Or perhaps "after you accept the contract, don't detach anything from it until you complete the contract."

Yeah, I know, not especially helpful, but I suspect that's about as good as you'll get.

At a guess, I'd say that you've discovered a weakness in the new "contextual contract" system that gives contracts based on existing ships.  It's a hard problem, from a programmer/designer perspective.  What constitutes the "existing ship"?  If it fragments into pieces (i.e. something undocks), does "the ship" still exist?  Which piece of it does the contract live in?

I assume that what Squad did was to equate "ship identity" with a particular command pod on the ship, so that if the ship fragments, its "identity" sticks with that particular pod.  That's not a bad solution, I suppose, but it raises the interesting question of how to choose which pod it's associated with if the ship has multiple pods.  Clearly, in your case, they made the "wrong" decision from your perspective.

If I put myself in Squad's shoes and try to think how I would code this if the problem were given to me... well, I can think of several possible coding strategies with different advantages and disadvantages, but none of them is a complete slam dunk-- every one will have some edge cases that will foozle it.

[PT]

9 hours ago, Snark said:

My guess would be, probably, "don't have the modules docked when the contract pops up."    Or perhaps "after you accept the contract, don't detach anything from it until you complete the contract."

Will try with undocking prior to acceptance.

9 hours ago, Snark said:

I assume that what Squad did was to equate "ship identity" with a particular command pod on the ship, so that if the ship fragments, its "identity" sticks with that particular pod.  That's not a bad solution, I suppose, but it raises the interesting question of how to choose which pod it's associated with if the ship has multiple pods.  Clearly, in your case, they made the "wrong" decision from your perspective.

I assumed that contracts would pin to vessels just like their names do, seemed pretty logical. But instead I got "Oh wait! We meant your shuttle!".

Next time such contract pop ups I will investigate a bit more. Maybe "proper" undocking sequence would help, eg. control station core, undock, switch to shuttle instead of my usual shortcut of switching control to shuttle first. Or maybe it is set to certain part in save file and simply borks on fetching vessel name.

 

Edit: damn, I'll be downloading all your mods after work ;p

Edited December 9, 2015 by PT

[ineon]

11 hours ago, Snark said:

I assume that what Squad did was to equate "ship identity" with a particular command pod on the ship, so that if the ship fragments, its "identity" sticks with that particular pod.  That's not a bad solution, I suppose, but it raises the interesting question of how to choose which pod it's associated with if the ship has multiple pods.  Clearly, in your case, they made the "wrong" decision from your perspective.

I wonder if a better solution would be to use a pod from whatever has been launched longest ago. They already have that info in the Tracking Station. Or maybe they did. @PT I assume that you launched the shuttle after the station, did you?

Maybe the algorithm for working out what is the root and what are peripheral modules is part count that doesn't span docking ports (i.e. things that are permanently connected together). A space station can easily be made of many small components - none of which have a part count over a few dozen, but a space plane is a complex beast and could easily contain 100's of individual pieces, none of which can be separated off (non-catastrophically).

[PT]

38 minutes ago, ineon said:

I wonder if a better solution would be to use a pod from whatever has been launched longest ago. They already have that info in the Tracking Station. Or maybe they did. @PT I assume that you launched the shuttle after the station, did you?

Maybe the algorithm for working out what is the root and what are peripheral modules is part count that doesn't span docking ports (i.e. things that are permanently connected together). A space station can easily be made of many small components - none of which have a part count over a few dozen, but a space plane is a complex beast and could easily contain 100's of individual pieces, none of which can be separated off (non-catastrophically).

Yep, KSS was up there first, modules joined it later. Everything connected through standard docking ports directly to station core (til yesterday, when I added extension module)