Lecture on 'grep'

Introduction to 'grep'

• 'grep' is a well-known command in UNIX systems, originating in the early 1970s.
• Used for searching arbitrary patterns of text in one or more files.
• Input can come from files or other programs (e.g., UNIX pipelines).
• Allows filtering of large amounts of input, which text editors struggle with.

History of 'grep'

• Name 'grep' has an interesting origin story.
• Developed in early UNIX days (circa 1970-71) on a PDP 11 computer.
• PDP 11 had limited computing power:
  • 32K to 64K bytes of memory.
  • Small secondary storage.
• UNIX software was simple due to hardware limitations and the preferences of developers like Ken Thompson and Dennis Ritchie.

The 'ed' Text Editor

• 'ed' was the standard UNIX text editor, pronounced 'ee dee'.
• Written by Ken Thompson, inspired by the QED editor.
• Operated on paper, not video display terminals.
• Commands were single-letter (e.g., 'p' for print, 'd' for delete, 's' for substitute).
• Line addressing was basic but included features like:
  • '1,$p' to print all lines.
  • '$p' to print the last line.
  • '1d' to delete the first line.

Regular Expressions in 'ed'

• Regular expressions were a significant feature added by Ken.
• Allowed patterns of text to be specified for operations.
• Written in a format like '/pattern/' to find text patterns.

Development of 'grep'

• 'ed' had limitations due to memory constraints (couldn't handle large files).
• Lee McMahon wanted to analyze the Federalist Papers, which required more capability than 'ed' could provide.
• Ken Thompson created 'grep' to handle this need:
  • Searches for regular expressions in one or more files.
  • Based on 'ed's 'g' command (global) for operations on matching lines.
  • The name 'grep' comes from the command structure 'g/re/p' (global-regular expression-print).

Educational Anecdote

• In 1993, the lecturer used 'grep' as a class assignment at Princeton.
• Students received the source code for 'ed' (1800 lines of C) to convert into 'grep'.
• They had advantages:
  • Knew the target behavior of 'grep'.
  • Worked with C instead of assembly language.
  • Disadvantage: none were Ken Thompson, who originally created 'grep'.

Conclusion

• 'grep' is an example of UNIX's capability to handle text processing efficiently.
• It originated from practical needs and was implemented swiftly by utilizing existing tools.