Quote:
Originally Posted by Holly Martins
It must be possible to produce a programme which would check scansion as well as spelling, highlight inversions, duff rhymes, repetitions, non sequiturs etc, but it wouldn't be as good as the Sphere.
|
Holly
You introduce an interesting point. I have made an intensive study of automatic textual analysis and I can tell you that there is no known algorithm for reducing text to
syllables (in English at least), let alone "scanning". Partly because there are no universally accepted rules, but, moreso, because English has its roots in so many other languages and inherits their varying stress patterns.
I made considerable progress on the matter (and I am no mean programmer) by analysing a number of available electronic dictionaries (there are a fair few) but eventually gave up because my own ear was undoubtedly better. One has to hard code so many exceptions to any rules one cares to dream up.
There are some reasonable heuristics but they fail often enough to make them not worth the candle.
I think I might be saddened had I been successful.
Philip