Re: Lilypond error behaviour

[Andrew Bernard]

I now understand, which I had not previously, that lilypond will make its best effort to continue at all times, and this is a core value of the design philosophy – thanks Simon for pointing this out. But nobody has been able to explain to me how classing the error in question as fatal and then going on to produce good output is consistent with normal English usage. Since it is not consistent with normal English usage it is a source of confusion, as others have also pointed out. One is loth to pull out a dictionary in discussions. but the Oxford English Dictionary 2nd Edition gives the definition or denotation of fatal as:

fatal: 6. a. Producing or resulting in death, destruction, or irreversible ruin, material or immaterial; deadly, destructive, ruinous.

Apart from this common language usage, it is my praxis and experience in all spheres of programming that a ‘fatal error’ means the program cannot continue, even if it would like to.

Surely by any standard usage a recoverable error cannot be reported as fatal. There is no ruinous outcome. The file is produced. This is what I am referring to. It certainly confused me. I would say there is a class of errors that lilypond throws that are clearly recoverable, and should be reported in a more fine grained manner.

If certain people choose to think this topic is obsessive on my part, then apart from the fact that that is personally offensive, then why bother discussing subtleties of programming and usage at all? Semantics matters, as does also, for example, the semantics of fine engraving of musical notation. Clear semantics leads to clearer understanding. If lilypond had instead reported ‘recoverable error at or near line x’ this thread would never have arisen.

Andrew