> Making a mistake by allowing non-standard syntax and later complaining that still so many people use this, is not the finest thing to do. I think that is unnecessarily harsh. First, the syntax in question was introduced for a reason that may well have seemed valid at the time (MS SQL server did it). Second, the SQLite project is extremely careful to maintain backward compatibility, more so than many other projects. You may disagree with this stance, but ”not the finest thing” is an unfair characterization. Of course, there are ways to get rid of such warts. For example, one could introduce a pragma to turn off this syntactical misfeature. Using that pragma might simultaneously enable other uses of the square brackets, thus encouraging people to use it. And after a year or two, the default could switch, so that you now have to use the pragma to *enable* the misfeature instead. (It would have to take a boolean argument to allow its usage to remain the same.) Should it be done? That is not for me to decide, but I certainly wouldn't mind.