I think that IR would be something like first order predicate logic, to which SQL and the relational calculus is closely related. Now that we have WITH and recursive queries, you've basically got a bottom-up evaluation of the declarative subset of Prolog (if you ignore issues relating to logic variables).
On Wed, Feb 18, 2015 at 1:13 PM, Roger Binns <rogerb at rogerbinns.com> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 02/18/2015 11:43 AM, Richard Hipp wrote: > > but I think the truth is we are probably stuck with SQL for a while > > yet. > > In theory there could be an intermediate representation form (like > compilers do) that is publicly available, with the (now optional) SQL > part producing IR, as well as any other query language > implementations. The LLVM project is an example of doing a design > like this. > > In practise the result wouldn't be very Lite, and would constrain > future development. > > Roger > > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1 > > iEYEARECAAYFAlTlAGsACgkQmOOfHg372QRP+ACgqoP3Ss8ZgvO95M8IVHhLRDbo > itEAoMhJKWIKiiYjsAqNUGl/cpv/e+fp > =Z+np > -----END PGP SIGNATURE----- > _______________________________________________ > sqlite-users mailing list > sqlite-users at mailinglists.sqlite.org > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users >
