It seems to be a decent scope, that has a few things going for it, mostly the 500 MHz bandwidth, and the fact that it has 16 channels.
But the downside of only 2 kilo samples of memory depth per channel makes it rather unsuitable in a lot of situations. Not to mention the highly expensive (and a lot of times broken) active probes that one can find on places like Ebay. One can also build up one's own 75 ohm probe, but I doubt one would easily get good performance in terms of linearity.
I myself would like to have one, but they always seems a bit too expensive for what I like to pay for them. (Or just have ridiculous shipping cost....)
But yes, it is an interesting scope, but I would say that it would be far from good specifications compared to anything these days, as the small memory depth practically makes it unusable in a lot of logic analysis situations. And if one is looking at over/under-shoot on digital signals, then one doesn't need more then one channel, maybe a logic analyzer with trigger output if one has a more strange problem.
Or one can sting together a few modern high memory depth scopes with the trigger in/out-put. Downside is that one doesn't have the pattern triggering of the TLS216. Or its sequence triggering feature as well. These two features being probably the biggest reason for why the TLS216 would still hold its weight today. As I have never seen any scope with such functionality. But one could do a "pass/fail" trigger on multiple channels and get the same result. Though, then one still needs 16 channels, and one normally only has 4....