Looking up Software Development Bibliographic Items

Purpose of Document

This defines how to specify searches for my Common Gateway Interface lookup to select bibliographic entries.

Summary

An entry is selected if there is a line in it that contains a substring that matches the pattern. The search is carried out by the UNIX awk program and so the pattern must be expressed in the awk form of regular expression. This means that certain symbols including periods, parentheses, brackets, slashes, asterisks and vertical bars have special meanings and will need a "\" in front of them.

By a fluke irregular operations can also be expressed. [ Irregular Patterns ]

Example Patterns

• Any item which has the string "Hoare" in it.

   Hoare

• Any item with either "help" or "Help" in it.

   [hH]elp

• Items with various spellings of "Parnas" with the second a replaced by any character

   Parn.s

• String like bananas, banananas, bananananas etc.

   bana(na)+s

• Items with the strings "regular" and "expression" on the same line

   regular.*expression

• Items with either regular or expression

   regular|expression

• Items with CSG or CFG in them.

   C[FS]G

• Items with either OO or "Object Oriented" on one line.

   OO|([oO]bject( |\-)[oO]riented)

• Items referring to context free or sensitive grammars with any combination of capitalized initial letters.

   [cC]ontext.*([fF]ree|[sS]ensitive).*[gG]rammar

• To search for characters other than letters or digits put a "\" in front of them.

   S\/SL

Formal Definition

1. Item::=@Line. Each Item is essentially a set of Lines -- because the order is not important to the search.

2. Line::=#char. A given line either matches or does not match a pattern... the set of matching Lines is written M: M:: pattern->@Line. The lines that match a given pattern

   An item is selected by a pattern if any of its lines matches the pattern:

3. For Item I, pattern p, selected(I,p)::= for some L:I(L in M(p) ).

   The lookup cgi use awk for searching. Typically M(p) is any string that contains a string that matches p as defined below.

4. pattern::= See http://csci.csusb.edu/dick/samples/regular_expressions.html

   Irregular Patterns

   Awk excepts Boolean operators in its searches:

   	awk '/string1/&&/string2/'

   So by including "/&&/" and "/&&!/" you can search for two strings in any order (but on one line):

    Program/&&/Law

   • Items with a line with both "Program" and "Law" in some order on it.

      Program/&&!/Law

   • Items with a line with "Program" but without "Law" on it.

      ACL/&&!/ORACLE

   • Items referring to ACL but not ORACLE

   Lexemes

5. char::= See http://www.csci.csusb.edu/dick/samples/comp.text.ASCII.html#char, -- the set of ASCII characters.
6. empty::@#char= {""}, -- the set with a single empty string in it.

. . . . . . . . . ( end of section Lookup) <<Contents | End>>