I have to say that I was at first extremely opposed to the idea, almost in a reflexive sort of way. Nevertheless, I did read through all the walls of text Binski posted, as well as most of the replies. After thinking about this for a few days I am no longer strictly against the proposal. To the contrary, I think that this is an idea well worth considering. Several considerations arise, though. The majority of things I came up with have already been posted, but because this is such a hot topic I can't just stand back and say nothing.
We need to be open to at least testing this idea and I agree with Relation-Ship. Doing this without involving the staff, or implementing any new mechanics should be doable. The proposed system needs to be fine tuned to prevent abuse and to correspond with the established lore, while also not being overly bureaucratic or needlessly complex. Ignoring the problems with complexity and bureaucracy, I can see a few ways to improve upon the basic concept to make it more lore-friendly. These are not well thought out, but whatever.
Only bases belonging to a faction whose ID logged a comparable number of hours to the aggressors' ID should be fair targets. These hours could be counted for all ships at first and inside the contested system once the siege starts. The reason for counting hours inside the contested region is to give the defending faction an option to still play their favorite characters but without having to deal with the siege. In doing so we would use the desire to win to promote sportsmanship; winning by driving the opposition from the server would not be possible. A problem could be the possibility to 'win' by not playing, still, a single station could switch hands even if the opposition stopped logging after losing the first couple of fights. Such a limitation would provide balancing and prevent unpopular factions from becoming victims. Players from the larger faction would have an incentive to balance the fight and keep it fun for both sides. Indeed, even dead factions should see some 'fake' players flying with their IDs - if only to pump up the logged hours in order to open up the bases to siege. I believe such fake activity could be the core around which genuine activity forms as the players would still have to put up the appearance of proper RP.
The cost of besieging a base should depend on the distance from the attacking faction's home-system.
Sieges could only take place during certain hours, admittedly to the detriment of players from some time zones.
Kill ratios should count towards something, perhaps increasing the cost of the next attack or outright automatically giving the initiative to the other party.
Clearly this system cannot be implemented on the whole server at once. It also needs to be limited in scope to certain agreed upon regions. There needs to be a way to limit the consequences of having a faction unpopular with players but strong in-RP. In general the system and the outcomes it generates need to be kept within what is reasonable based on lore and RP. In other words, this system needs to conform to the inner logic of Freelancer universe, else we risk extreme immersion breaking events taking place due to player grudges or just luck. The whole RP universe would then collapse under the resulting inconsistencies. This, I believe, is what drives the majority of negative comments. (Apart from the admittedly combative posture Binski decided to take).