An interesting discussion and a good example of how the practical doesn't always agree with the theoretical. Years ago when wooden 5 bar gates were made of real wood in reasonable cross sectional areas the brace was (from memory) invariably in compression. Nowadays with gates made from paper thin sections, putting them in tension stops any droop caused by bowing sideways of the invariably too thin cross brace or unseasoned wood. However if you look at the current best gates when the brace is in tension the hinge post is usually extended upwards and the suspension point of the brace is above the top rail thus providing some additional support for the top rail as well. Alternatively they are triangulated with the first half of the gate span braced in compression and the outer half in tension.
Just observations of the practical application, I'm not getting involved in the boring theoretical stuff. Of course it does highlight the crap that is set as exam questions these days; perhaps it is to find out if the train driver knows which one to aim his train at in an emergency - they have got steering wheels on trains haven't they :-)))
Keith