The "options are costly" argument is of course valid, but even then it does not need to be as complex as the article describes. As long as it is clear when an option is "advanced" one, they can be pretty low effort. Look at firefox's about:config for example -- no localization, no help, some options might be no-op, break browsing experience, or not implemented.. and yet those options were extremely helpful occasionally.
I agree. Software should be designed for advanced users, and also be sufficiently documented (both of these are true (in my opinion) whether or not this involves adding additional options/features). A separate menu (or file, if it is documented) for advanced options would be OK, though. (These things are especially true of web browsers, but apply to all modern software.)
It is true that adding options is not always the way to do it, but sometimes it is. There are other ways, including allowing users to specify external commands, using standard I/O with pipes, generalizing an existing option or feature, etc.
Of course there are other considerations like the article mentions (and that you do, too), but I still think that it is better to include options if they would help.
It is also good to, when adding features (or options), add good ones instead of bad ones.
Additionally, apparently SQLite is commonly used in many programs. If that is the case for your program, then the capability for the user to use SQL queries will allow to do many things.
