It might seem contrary to discuss FTL only after a post on interstellar empire, a concept that pretty much depends on star travel being cheap, convenient, and above all fast. (Yes, there are STL scenarios, but they are so stretchy that violating General Relativity suspends less disbelief.)
But in broader perspective it is our desire for interstellar empire, or other such cool settings, that calls out for the willing suspension of disbelief in the first place. All spaceships, as a commenter on this blog once observed, travel at the speed of plot, but that is especially true when it comes to FTL. Spaceships in the Plausible Midfuture have some of the constraints of real-world vehicles, and plots must work around them, but the only constraint on FTL is that it must sound convincing to readers who want to be convinced, at least while reading the story. (Although FTL systems in SF not infrequently fail even at that.)
Unlike last post, this time I remembered to link the relevant page at Atomic Rockets, which covers or links the relevant physics and pseudo-physics, and offers some useful FTL typologies. Also a link to my observation in the Tough Guide to the Known Galaxy that all FTL tech is equivalent to flipping tarot cards while chanting in Welsh - i.e., effectively it is magic.
And a post here from last year which noted that contemporary physics has just barely left the door open to FTL, subject to certain constraints. The relevant effect of these constraints - as I understand it (and General Relativity is not my field of expertise) - is that you can stay out of temporal trouble so long as your baseline FTL routes do not cross-connect.
To follow the rapid transit metaphor I used in that post, you can have multiple routes out from Earth (via wormholes or whatever), with as many branches and sub-branches as you want, and travel times can be as fast as you want them to be. But each station can only be on one line. If you can take two separate routings to reach a given destination, of differing length, then you have to pay the temporal piper.
For example, from Sol to Sirius is 8.6 light years, while from Sol to Wolf 359 is 9.0 light years. Given tramlines equipped with a suitable array of oscillating hands, you can reach either one in in a month by your onboard clock, or an hour, and return in similar time. In Earth's frame of reference, going outbound you are in fact travelling nearly a decade into the future, and returning you are going nearly a decade into the past. This produces no awkward time-travel-esque results, because you don't return to Earth before you left, and you can't get back to Sirius for a second visit prior to your first one.
From Wolf 359 to Sirius is 7.7 light years, so the roundabout route to Sirius via Wolf 359 is 16.7 light years. If you could simply open another tramline, you could travel to Sirius via the indirect route, arrive there 17 years from now, read some stock quotes, return by the same roundabout route to Earth, then go by the direct route, and invest in Sirian stocks years earlier. Big trouble, and not just for stock markets.
So, to avoid breaking causality, such routings must be ruled out, or at least force you to twiddle thumbs those extra eight years instead of raking in a bundle. I don't begin to understand the required coefficient of jive, but apparently it can be done without violating General Relativity, at the price of making such roundabout routings very inconvenient.
(Commenters who do understand this stuff are welcome - indeed, invited - to step in and expand on / correct these points.)
The hitch is that for story purposes we probably want those roundabout routings - not to travel into the past, but merely to run blockades and the like.
My impression is that there is still a workaround. Simply assume that all FTL travel to Sirius proceeds via Wolf 359, but that there are separate 'local' and 'express' tramlines along the route - the former stopping off at Wolf 359, while the latter passes right through without entering normal space.
If each entry or exit from a tramline involves a cost (e.g., reaching the jump point through normal space), then regular Sol-Sirius travel would follow the express tramline, passing by Wolf 359 nonstop. But the local routing, stopping off in normal space at Wolf 359, would remain available when needed for story purposes.
Yes, this is more than a little strained (and possibly downright wrong), but hey, we're talking about FTL. The alternative, so far as I can see, is simply blowing off a lot of well-establish physics entirely. Remember, it isn't as if Einstein kicked Newton onto the ash heap; old Sir Isaac still gives an approximation good enough for interplanetary travel planning.
Moreover, the above gimmickry can be almost entirely buried out of sight, at least so long as your intended flavor of FTL is jump-oriented, and you don't get deeply into the weeds about the sequence in which routes were established in the first place.
Having said that, this approach should probably be left to physicists, or people who do General Relativity as a hobby, and might as well be physicists.
My more general advice on FTL - which I seem to have arrived at via Wolf 359 - is to say as little about the mechanism as possible. Don't talk about wormholes, don't talk about Alcubierre, or any other present-day speculation. All you will do is date yourself, because I don't need to travel into the future to guess that these particular edge-of-the-envelope speculations will be superceded.
Bury all that theoretical stuff under a couple of toss-off jargon terms - Horst-Milne congruencies, Alderson Drive, whatever - and concentrate on what actually matters, which is how FTL actually operates in a setting: size and cost of the gizmo, apparent travel times, available routings, and so forth.
Some of which I will (probably) discuss next post.
Related links: Atomic Rockets, Chanting in Welsh, and Rapid Transit.
The image comes from a Battlestar Galactica website.