As I contemplated what "efficient"<sup>a</sup> pre-scanning would look like, I realized it would need to handle input such as this:<br><code> create table " Travails of a Verbose Identifier Afficionado (Or, ""How I came to avoid obtuse names."")" ( " When I learned to program, identifiers could contain up to 6 dark characters, with any amount of embedded whitespace.", " For larger programs, I sometimes kept a symbol table; There detailing what names variables would have if not for such limitations. /* Crucial, but too much work! */", " Now, in SQL(ite), identifiers may have virtually any length. (Hooray!)", " My main problem today is that complex queries no longer fit on a screen and so they are harder to read in their totality.", " -- (Funny looks from my coworkers are tomorrow's problem.)" ); </code>.<br>This dampens my enthusiasm somewhat. ---- a. I put "efficient" in jest quotes because it is unclear what the measure of success should be.