Jim Langston wrote:
> My main question then becomes, why hasn't anyone done this yet
It sounds like you're in the middle of reinventing SSQLS, so the answer
is, someone already has done this. :)
> UserTable.insert() << MyString << MyFloat << MyInt <<
> MyOtherInt << MyBool;
The only difference between this and SSQLS is that SSQLS is a statically
typed system, fixed at compile time, and this alternative is dynamically
typed. But since SQL is statically typed and is not object oriented in
any meaningful way, I'm not sure there's any real advantage to going
with dynamic typing here. It's a matter of preference, not one of being
able to do something one way that you cannot do at all any other way.
> Does anyone think this would be a Good Thing or a Bad Thing?
Obviously I'm rather dismissive of it, but don't consider this a
complete shoot-down. I don't have much use for template queries,
either, but I'm not going to go and remove them from the library.
If you build this and the implementation seems to be reasonable, there's
a fair chance I'll add it to the library. The only caution I offer to
you is that if you want it to appear before v3.0, it can't break the
library's ABI. That means you cannot change the signature of any
existing functions, and you can't change the inheritance hierarchy;
there are probably other limits I'm forgetting now. Basically, if you
can't do it by just adding new member functions and new classes, it'll
break the ABI, so it'll have to wait until v3.0.