I see the point of some people argumenting towards put the origin to the circle center.
However, I have the following arguments against that:
1.
That origin will be far away out of the bounding box of this part.
When using this part in a tool, and clicking "rotate", a rotation about such a far away origin
can easily rotate the part out of the current window, and you have to tediously find it again.
2.
That origin is only a useful one when you are building your tracks of 45° and 90° curves.
If you use more complicated and arbitrary sequences of track, e.g. use 1 curve, then some straight,
another curve, some straight, another curve etc., like you can do with train track layout software BlueBrick,
the circle center far, far away isn't really of interest for you. You want to rotate the part in-place
to properly connect to the existing track.
3.
Consistency with existing train track curve parts for me here is a very important matter.
To me it makes no sense at all to make 1 of the train track kinds behave _differently_ than the others.
All curved track parts should follow the same principle to place the origin.
So for the reasons above, I like your current choice of origin, i.e., in the middle of the central sleeper part.
That's a better choice than "at one end" IMHO, because that would introduce an a-symmetry.
Why should 1 end of the track be "more special" than the other?
The part center is a good choice.
I also object to changing the origins/placements of all currently existing train track parts.
If you dislike their placement, you can easily create a single one-liner file on your own (i.e., not part of the library)
which relocates/rotates the part in question to a more convenient location for you.
In your models, you then use _that_ file, if it is easier for you to position.
-----> my opinion: keep the current origin in the center of the center sleeper.
However, I have the following arguments against that:
1.
That origin will be far away out of the bounding box of this part.
When using this part in a tool, and clicking "rotate", a rotation about such a far away origin
can easily rotate the part out of the current window, and you have to tediously find it again.
2.
That origin is only a useful one when you are building your tracks of 45° and 90° curves.
If you use more complicated and arbitrary sequences of track, e.g. use 1 curve, then some straight,
another curve, some straight, another curve etc., like you can do with train track layout software BlueBrick,
the circle center far, far away isn't really of interest for you. You want to rotate the part in-place
to properly connect to the existing track.
3.
Consistency with existing train track curve parts for me here is a very important matter.
To me it makes no sense at all to make 1 of the train track kinds behave _differently_ than the others.
All curved track parts should follow the same principle to place the origin.
So for the reasons above, I like your current choice of origin, i.e., in the middle of the central sleeper part.
That's a better choice than "at one end" IMHO, because that would introduce an a-symmetry.
Why should 1 end of the track be "more special" than the other?
The part center is a good choice.
I also object to changing the origins/placements of all currently existing train track parts.
If you dislike their placement, you can easily create a single one-liner file on your own (i.e., not part of the library)
which relocates/rotates the part in question to a more convenient location for you.
In your models, you then use _that_ file, if it is easier for you to position.
-----> my opinion: keep the current origin in the center of the center sleeper.