I have every basis for a debate, see my comments above. you both are just in denial.
the reason why i say about regular usage is cos you can use something once or twice for messing about and it can work fine, but under regular use and also stress testing, you can see problems after constant use. It's a common thing to do in any environment from software design to the production industry, if you don't know about daily use testing or stress testing, that explains alot....
obviously you 2 cannot listen to reason and logic so on that note, i'm outta this thread until some common sense appears.....
First, there are many (in my view, more) plausible explanations for Apples non-inclusion than that performance deemed sub-par after extensive testing.
Second, i have personal experience of software-testing, have you?
Third, as no one uses Siri constantly, that use-case is quite useless. Also, as their is no concurrent processing, and each load essentially the same, behavior is extremely predictable.
Fourth, i have a masters degree in systems development (and currently work within academia as an information systems researcher).
p.s.
I'm still not arguing that Apple may not have capabilities in place to make a full roll out feasible. However, that point is moot. Nothing other than Apples decision to not include non-4S devices is hindering the inclusion of e.g. 3GS. At least, no-thing that has a technical nature.
----------
Can't we at least accept that there are technical resources necessary "somewhere"? For example, there needs to be x amount of servers available to process requests from y amount of devices. Now add orders of magnitude more devices to y, what happens to x?
x is scaled up accordingly. financial, not technical limitation. technically, scaling is pretty much a non-issue; one can simply buy computation as a service*.
* technically, "total computation" is capped. however, one would not be anywhere close to hitting the ceiling by allowing non 4S users to run Siri