These notes are the original notes provided with the OTA text (U-1936) and were written by Duncan Macrae-Gibson
This file was originally prepared by Greg Hidley, evidently from a collation of an early form of the electronic text of the OE corpus held by the Dictionary of Old English with the printed text of the Anglo-Saxon Poetic Records. The procedures used to generate it are not known. It would seem that it was never quite finalised (from the presence at the start of The Battle of Maldon of an annotation "NEED TO ADD lines 289-294 ...?" - though the lines are in fact present), and it contains many inconsistencies and errors. The present revision rests primarily on a computer comparison with files representing the latest available update kindly made available by DOE. This has been supported by as much visual checking against the printed text, and in some cases other editions and facsimiles, as time has allowed, and where this brought to notice fuller information about manuscript readings it would have seemed foolish not to incorporate it, as in Solomon and Saturn. Finally the text has been rechecked against OCR scans of the printed text procured and scrutinised by Pat Bethel, and against her notes of discrepancies between the scans and her copies of the DOE files. The result should be a very much more reliable text than the original one; it does not purport to be a re-edition, however.
The file identifies poems by a prefixed COCOA reference of the form <t a3.22.10>, where the first numeral normally defines the codex, the second the text, and the third (if present) the item. The codex numeration accords with the ASPR volume number up to 4 (so that the example given defines Exeter Book, first block of Riddles, Riddle 10), but vol. 5 becomes 5 (Paris Psalter, with second number that of the psalm) and 6 (Meters of Boethius, with second number 0 for the proem and then that of the Meter). The texts in vol. 6 (Minor Poems) have separate first numbers, starting at 7 and following the order of the printed text, from <t a7>, Finnsburh, to <t a41>, Brussels Cross; note however that the Hague version of Bede's Death song is <t a33.2>, and precedes the West Saxon version, <t a33.3>, as in the DOE files. The runic texts which would be <t a40>, Ruthwell Cross, and <t a42>, Franks Casket, are omitted. So in the original were <t a43.1> to <t a43.12>, the Metrical Charms, but they are supplied in the revision.
These omissions were probably deliberate; it is presumably by inadvertence that the Exeter Book text breaks off abruptly after Riddle 38. This missing matter is also now supplied; COCOA references accord with those used by DOE, so that <t a3.31.1> and <t a3.31.2> are Riddles 30b and 60, <t a3.34.1> to <t a3.34.35> Riddles 61 to 95.
The character 'thorn' is coded as '*', 'eth' as '=', 'ash' as '$', and 'tailed-e' was originally '-', in each case without distinction between capital and lower case. The half-line break is flagged with '#', which appears after or before an isolated half-line according as it seems best regarded as a first or second half-line (this applies whether a companion half-line is lost or illegible in the MS, presumably lost in copying, or may never have existed); a line without '#' is not even fragmentarily an OE verse line. Line-end was originally represented simply by the end of the computer line. This coding of 'tailed-e' is inconvenient, since the same character also represented a punctuation dash, and a hyphen; in the revision 'tailed-e' is represented '~', and dash '--'. The lack of a specific line-end marker allows confusion between end of poetic line and mere screen or printer wrap-round of a long line; in the revision an end-of-line flag '|' is added to all text lines, with the exception of those (always Latin) which the printed text does not number, where its omission will, it is hoped, act as a flag allowing line equivalence between electronic and printed text to be maintained during computer manipulation.
Words containing emendations or reconstructions of illegible or obscure manuscript have a suffixed '%'. Suffixed '%' was originally used also to indicate a word which is or may be fragmentary, adjoining lost matter, and at times added to a complete word to indicate that it thus adjoins. This clearly rests on the practice of the DOE files; indeed much of the marking seems simply a reproduction of the DOE equivalent. In DOE the intention is merely to indicate that the user should check the manuscript because there is some textual problem; it does not therefore matter that the same code can mark forms complete but unauthenticated, authenticated but incomplete, and both complete and authenticated. It is quite acceptable to include cases where received text rests only on one of the Thorkelin transcripts, or even where the only peculiarity is that a form has been created by subpuncting. The DOE files as available to me, moreover, at times fail to flag even forms unquestionably editorial. The outcome of all this is not satisfactory in a file likely to be used as a basis for computer collection of forms and collocations. An attempt has therefore been made in the revision to regularise the use of '%' to mark all but only cases where text as presented is not reliably authenticated, though it is not to be hoped that it has fully succeeded. For forms authenticated but incomplete or possibly so the revision adopts the DOE '/', but suffixed or prefixed (or both) according as the lost matter follows or precedes. Lost matter as such is indicated (by the code "//") only where there is physical evidence of loss, and then only if there would otherwise be no sign that adjoining words should not be reported as in normal collocation, or to preserve line equivalence with the printed text where this numbers a lost line.
Latin words or numerals are enclosed in square brackets, but this is omitted where Latin words appear to have been adopted into the English text. Runes are transliterated, and enclosed in doubled '%': forms such as %%EA%% represent single runes.
Hidley's basis of choice between the electronic and printed base texts often cannot be confidently inferred; an attempt has now been made to regularise practice as far as possible, but inconsistencies will certainly remain. Where his choice had to be simply a matter of editorial preference it has not normally been disturbed, but it would have seemed perverse not to follow DOE and recent editors in restoring the MS. forms "sunnan", "meowle" (Exodus 109, 580) and "behly*ed" (Riddle 14, 10).
One particular point is worth noting: Hidley normally followed DOE in preserving 'tailed-e', where the printed text relegates it to the Apparatus and substitutes 'e' (unstressed) or 'ash' (stressed) in the main text; this preservation of 'tailed-e' has, it is hoped, been made universal. Punctuation was clearly intended to be that of the printed text; as with other things, an attempt has been made to regularise this.