Laudator Temporis Acti

Excerpts from J.R.R. Tolkien (1892-1973), "Prefatory Remarks," Beowulf and the Finnesburg Fragment: A Translation into Modern English Prose, by J.R. Clark Hall, revised by C.L. Wrenn (London: G. Allen & Unwin, 1940), pp. ix-xliii (I haven't seen the original):

But you may be engaged in the more laudable labour of trying actually to read the original poem. In that case the use of this translation need not be disdained. It need not become a 'crib'. For a good translation is a good companion of honest labour, while a 'crib' is a (vain) substitute for the essential work with grammar and glossary, by which alone can be won genuine appreciation of a noble idiom and a lofty art.

[....]

The primary poetic object of the use of compounds was compression, the force of brevity, the packing of the pictorial and emotional colour tight within a slow sonorous metre made of short balanced word-groups. But familiarity with this manner does not come all at once. In the early stages — as some to whom this old verse now seems natural enough can doubtless well remember — one's nose is ground close to the text: both story and poetry may be hard to see for the words. The grinding process is good for the noses of scholars, of any age or degree; but the aid of a translation may be a welcome relief.

[....]

But no translation, whatever its objects — a student's companion (the main purpose of this book), or a verse-rendering that seeks to transplant what can be transplanted of the old poetry — should be used or followed slavishly, in detail or general principle, by those who have access to the original text. Perhaps the most important function of any translation used by a student is to provide not a model for imitation, but an exercise for correction. The publisher of a translation cannot often hedge, or show all the variations that have occurred to him; but the presentation of one solution should suggest other and (perhaps) better ones. The effort to translate, or to improve a translation, is valuable, not so much for the version it produces, as for the understanding of the original which it awakes. If writing in (one's own) books is ever proper or useful, the emendation or refinement of a translation used in close comparison with a well-studied text is a good case for the use of a careful pencil. The making of notes of this sort is at any rate more profitable than the process more popular (especially with those reading for examinations): the inter-linear glosses in the text itself, which as a rule only disfigure the page without aiding the diffident memory.