Actually Austra, though I appreciate NSF maybe far more experienced than I on the 'bus, I would be tempted to side with Kirks and largely ignore NSF's post.
If there's "marginal visibility" or "a real chance of a mis-identification of the runway due city features, long roads etc", go down the ILS, other instrument approach or divert- Kirks is right - if you can't see it, don't f
about. The kind of procedure NSF has cited is reminiscent of an instrument NPA - not a visual approach.
1234