Yeah, not offering a player-server option where such an option is possible would be unethical. It is basicly an outright statement of "Someday, when you least expect it, I'm going to shut down the server and render your game useless. Also, grind. You're paying me by the hour, after all!"
I'm pretty sure at least some communities would be able to get the resources to host servers.
Still, if BOTH options are offered, I might just get a multiverse subscription anyway. Big server farm, big galaxy, infinite noobs to pwn con interact with.
Oh, one thought.
Tracking and avoidance.
First-degree targeting:
The obvious 'point-and-shoot' turret/etc. targetting would bring the weapon to bear at the target's current position, firing when the target falls within the sights. This can be dodged (assuming non-hitscan weapons) simply by maintaining enough relative velocity in a direction other than straight towards or away from them. (ie, keep moving, circlestrafe, etc.)
Second-degree targeting:
Some so-called "advanced" systems will use the previous scheme operating on a virtual target placed using the distance to real target and relative velocity of real target, along with the speed of the projectile. This is 'lead', and will allow you to actually hit a moving target, at least some of the time. These systems are probably pretty simple to make (though they require at least two "pings" of a target's location, assuming they are only fed position data, meaning a tiny lockon delay), but can be defeated by accellerating. Remember that trick in some old games where you can shake left and right and the enemies will fire at two points beside you? That's what this looks like.
Our "dodging" systems should be prepared for at least this level of targeting, and fast-reaction weapons like point-defence systems should probably use this.
Third-degree targeting:
This system will require a lot more calculation, as well as at least three sensor refreshes per step, to implement; but will probably be able to hit anything out there fairly regularly. Basically, this one adjusts "lead" assuming constant acceleration from ping one through ping three until the point where the shot lands. This should be able to accurately hit anything without a changing engine throttle.
To dodge this, one needs to change acceleration over time.
This is probably enough for main weapons, but for rediculously-long-ranged stuff, more might actually help.
There is better: You can extend this to infinity, but the math (and CPU hogging, execution time, needed sensor readings...) gets pretty rediculous and the countermeasures ever-more precice and time-consuming.
Fourth-degree targeting can be countered by varying how fast you change the accelerator setting. Fifth-degree targeting can be countered by changing the rate of doing that...
This problem is reversed for missiles/evasive computers. Our basic anti-countermeasures missile should probably zigzag, spiral, or sine-curve to avoid hits. No computer is necessary for this, a simple timer (if those are implemented somehow) should suffice.
Up left down right repeat! Left Right repeat! Left coast right coast repeat!
Also, while I don't know much about ship customization in this game, what do we think of modularization? Setting up some semi-standardized missile bays, hardpoints, docking clamps, etc. to be pulled out of the database when designing a new ship?
Cons:
Security concerns (enemies can learn a little about one ship from examining a different one)
Bulk/bloat (not shaving off things because they are required to meet standards)
Pros:
Ease of upgrade (If I develop a new kind of "HABuster Mk IV" tailormade anticountermeasures missile, it fits in all the standards-compliant launchers already.)
Ease of repair (You can ask for a part and get what you meant the first time)
Fleet familiarity ("Hmm... Matz05 is using a Balista heavy cannon. I know it's for long-range use against larger targets, so I'm going to pick off those drones swarming him and let him deal with the mothership")