This is some still very basic documentation for using the Web interface to the Index to Sumerian Secondary Literature (ISSL). This interface utilizes a suite of programs collectively entitled SumFind version 4 (SF4). SF4 is written and maintained by Steve Tinney.
Search requests for SF4 are entered in a very simple ASCII character set so no special input methods are needed to communicate with it. For shin, use c or C; for heth, use h or H; for subscript digits just use normal digits.
One useful feature of SF4 is that it uses grapheme aliasing to improve the success of searches. Thus, sag9, sa6 and ca6 are all indexed the same way, and searching for any one of them will return matches to them all. The alias list is always open for improvement, so if you suspect you find an aliasing problem please e-mail me. In combination with the tilde-operator (see next section) SF4 solves to a very large extent the two basic problems with heterogenous transliteration schemes: different renderings of the same sign/semantics combination and different use of hyphens.
The simplest search is just to type in a grapheme, like `ba'. You can use hyphen, underline (_ i.e. `ba_ba') or tilde (~ i.e. `ba~ba') as separators. The hyphen matches graphemes joined by any non-space in the original (hyphen, dot etc.). The underline matches graphemes separated by a space in the original. The tilde matches graphemes that are adjacent regardless of the separator.
SF4 supports three boolean operators: AND, OR and NOT (but you can type them in upper- or lower-case). Thus, one can ask for `ba and bu' or `ba not bu'. The default scope of boolean operators is the rest of the line; scopes can also be explicitly specified as described in the next section.
One can also search for graphemes, or groups of graphemes, which should occur in the same number of signs, words, lines or texts by using the `within' operator. The general syntax of this operator is:
WITHIN NUM SCOPE
In which NUM is composed of an optional + (plus) or - (minus) followed by digits, and SCOPE is one of SIGNS, WORDS, LINES, or TEXTS, all of which may be abbreviated to the first two letters. Thus, a more complicated search might look like this: `a-kal-la and cec-kal-la within +-3 lines'. This would match any locations in which a-kal-la and cec-kal-la occurred within 3 lines of each other, regardless of which came first. At present, SF4 just returns the first line of a match which covers a range of lines, as in the last example. In future this restriction will be lifted.
SF4 has a complete implementation of regular expressions, aka regexps (the regexp routines from GNU grep are built right in to SF4). Some simple pointers should be adequate to get you up and running.
One useful shortcut is to put `0' after the search term, e.g. `kici0', which retrieves all matches to kici regardless of their numerical index.
Note that a regular expression only modifies one grapheme; to search for any ba followed by any ba one must repeat the regexp (`ba0~ba0').