There's a space of texts which are grammatical English.
There's a subspace of text which have meanings
There's a further subspace of texts which have a regular structure, the poems.
Various subspaces of poems. Relevant structures different between languages.
English poems have stress patterns, rhyme, alliteration, para-rhyme.
Latin poems have syllable length patterns, no rhyme (or alliteration?)
French poems have syllable count and rhymes, and rules are strange (silent e's sometimes count, according to antique pronunciations and liaisons not used in ordinary French)
Also caesura rule (word may not cross half-way boundary)
Attempts once made to do syllable-length poems in French, but not fashionable now.
Take a particular text from the space of French poems.
There is a meaningful English text whose meaning is closest to the French poem.
It is unlikely to be an English poem.
There may be a poem which has a reasonably close meaning
Meaning metric, goodness as poem measure, goodness as text measure.
What is translator wishing to achieve?
Best poem within a certain meaning distance?
Best poem according to some combination of goodness as poem and closeness of meaning?
Writing poems is very hard. Unless translator is accomplished poet, poem will not be worth reading as poem.
Why do poetic translation at all?
Try to find closest meaning, or go for good English prose with close meaning? Same problem to define good English prose. Is bad English which best captures original meaning the goal?
Do we want an English version which will help a non French reader read the French?
Do we want to capture the French rhythm in English?
Do we want a version which will help a French reader with the words and idioms that he does not know?
Do we want to produce good English poems by using the ideas from the French poem?
Is this translation or (laudable) plagiarism?