It's funny how much false credit we're giving them. At the time Ko built its reservation system, it hadn't served a crumb. It was declaring its own hotness in advance, and then building a system to cater to it. Of course, I don't grudge them their success, nor the accuracy of their prediction: Chang is, if nothing else, a good self-promoter and a keen judge of the marketplace.Back then, hot restaurants weren't on Opentable. This still tends to be true even now (although less so than then). So, back then, for restaurants as in demand as Ko, the only option was to redial until you were put on hold, and then hope against hope, after 45 minutes or so of phone work, that you hadn't wasted your time and you could get a table.
There's a point you've overlooked. At the time Ko's reservation system came along, Internet reservations were not a novelty: OpenTable existed. Ko didn't change the method of reserving; it reduced the number of methods from two (phone or Internet) to one (just Internet).
I won't rehash the whole Ko thread, and remember, I like reserving on the net, but there are a whole bunch of reasons why people sometimes prefer the telephone, and Ko doesn't offer that possibility.
So they didn't drop one of two available options (which I guess in English are called alternatives). They replaced a dysfunctional one with a functional one.
OpenTable doesn't have any limitations I am aware of. It could have done everything the Ko reservation system does, and in the early days (when the Momo system was buggy), would probably have done it better. The ability to hold back tables (for regulars, telephone reservations, or walk-ins) is a feature of OpenTable, not a bug. Some restaurants (though usually not the hot ones) put ALL their tables on OpenTable; it is simply up to them.
With the economics of OpenTable and the limitations, you can hardly blame a restaurant for building their own reservations system.
Of course, OpenTable charges them for the service (you expected it to be free???). Chang had to pay a programmer to develop his system and had to pay hosting charges to run it. I assume, in the long run, he returned his investment, but up front it probably cost him more. I think Momofuku and Next are the only restaurants in the country that have developed their own online reservation systems. It's both hard and expensive to do, and replicates an already-solved problem, which is why most just use a service that is already out there (OT isn't the only one).