I'm not sure this rises to the level of "patently obvious" or even "probable." How would this work better on a practical basis?
However, it is patently obvious that both the queue and lottery systems I outlined are more accessible, fair and egalitarian than either the standard phone system or the Ko online system.
Let's look at the queue model: According to your proposal, they would open up the reservations queue at some date, and people could load it up. Your specific example proposes offering a reservation slot nine months in the future. How is this egalitarian? This seems to unduly favor not only people who were "in the know" about Ko to the tune of 9 months of more in advance of other people, but unduly favors metro-NYC residents, since most visitors to the city are unlikely to know about their plans 9 or 6 months in advance -- especially where they might like to go to dinner. Would there also be some mechanism in place to prevent someone from placing his name in the queue 500 times and then picking and choosing (or giving away) the reservations they want? This would be simple to do for anyone with a knowledge of email forwading aliases and proxy servers (this favoring the technologically savvy). A queue system would either be too short a queue to offer any benefits over the current system, or would be so long as to be about as egalitarian as the queue to buy season tickets for the Green Bay Packers at Lambeau Field, which currently favors those whose parents guessed they might like season tickets some ten years before they were born.
A lottery system, performed on a daily basis, seems a bit more egalitarian, but still has many potential problems. Presumably the system would have some way of accounting for reservations for parties of two or one, which might involve some fairly tricky weighting math depending on how egalitarian you wanted to make the lottery (Does a reservation for two get one chance or two chances in the lottery? How do you account for the fact that some chances take away two slots whereas others only take away one?). And, of course, the technologically savvy should be able to game a lottery system to their advantage fairly easily. Presumably they're not going to assign staff to visually check the lottery entrants on a daily basis, and anyone with imagination and technological savvy could fool an automated computer system using proxy severs (to fool IP address checking), email forwarding aliases (to fool email address checking) and variations of, say, Steven Shaw, Steve Shaw, S. Shaw, Steven A. Shaw, Steve A. Shaw., S. A. Shaw, Ellen Shapiro, etc. It's even possible to get one-time credit card numbers for use online.
Either system has the problem that you don't know right away whether you got your reservation, which is a major advantage of the current system.
Now, it could be possible to make these models more "game-proof" and egalitarian (the lottery system much more so than he queue system) but it would require a much more sophisticated system, and most likely one that requires a certain amount of human monitoring. Meanwhile, the advantages over "first X reservations to click through starting at 10AM two weeks out" seem slight for the work and complexity that would be required to make another system work. The question is whether the system is reasonably fair and egalitarian, and the answer seems to be that it is. It doesn't seem reasonable to invent an entirely new system with much greater complexity and security, that is more burdensome to most customers, in order to solve the problem that some people seem to have a faster connection to the reservations server than others.
One possible solution might be to set aside a small number of reservations each day (say 4 seats) for a small lottery in order to accommodate those who may not have been able to reserve due to a slow internet connection to the Ko server. Even this would carry some significant technological burden for Ko, however, and may simply not be worth that amount of extra trouble.