Scheme Scripting for Language Teachers
and Other Kinds of Poets

• Introduction
• Online Textbooks
• Easy Code Snippets

Nowadays, new scripting languages are created everyday. Each seems to duplicate the functionality of the last and none seems to communicate with any of the others. In the middle of this sea of change lies Scheme and C, languages that have been around forever growing slowly and cautiously since the early 1970's. Scheme is a dialect of Lisp whose origins date back to very invention of computers. Scheme is a very small but sophisticated language. It's a medium for research and innovation in computer science. It is also used to introduce new computer science students to programming at MIT (Massachusetts Institute of Technology) the world's foremost technical university. The most classic and respected textbook on programming uses Scheme.

Because of the volume of material available online Scheme is also the ideal language for those who want to educate themselves. There is even a tradition of implementing Prolog the most important language for computational linguistics in Scheme and Lisp.

Nowadays who has the time to put their life on hold and go back to school to learn. There are also millions people around the world who want to learn but who have no access to books. (This problem is particularly acute in the country that's my home away from home, Myanmar.) MIT is trying to remedy this problem by making the material for its courses available free of charge over the internet. For those who want to learn and who are motivated to learn, Scheme is the answer.

The idea behind this page is to point you to the numerous online textbooks for learning Scheme and to supplement these textbooks with small snippets of code that you can use to extract information out of texts that you can use in the classroom. The information you extract out of texts might be the ways in which a word is used, or it might be a particular grammatical construction. Some people are even starting to extract instances of language functions (e.g. polite requests or refusals, ways to ask for information,...) out of texts.

People often complain that scripting languages are slow. Well, that's where C comes in. If there's something you have to get done quickly (like looking for occurences of a set of words in a text) you can rewrite that part in C (the Aho-Corasick algorithm would be the algorithm to use in the example above) and hook it up to the slower scripting part with something known as a "Foreign Function Interface."

C is the language that first pulled programmers out of their assembly code without sacrificing efficiency. Scheme has been a primary vehicle for research, innovation, and education in computer science for decades, before the internet was even a twinkle in computer scientists eyes. Could there be a better language combination for applying new innovative techniques to the solution of practical scripting problems?

One final note about computational linguistics: There are a lot of exciting things happening in computational linguistics right now that are relevant to English teaching:

1. Wide coverage computer grammars of the English language can automatically parse sentences to help teachers explain syntax and "focus on form" with students.
2. Tools like Penn Tree Bank's tgrep are being used to find grammatical patterns in large language corpuses. This would be a great complement to the use of corpus based dictionaries like Collin's Cobuild or Cambridge International.
3. Chatterbot or conversational robot technology is improving so it might be possible to have use a computer as a surrogate conversation partner in the near future.

Unfortunately, most of these innovations are trapped in university research departments and are only available for large amounts of money. Let's try to free them up so we can use them!

• Structure and Interpretation of Computer Programs (SICP) The classic.
• How to Design Programs Dr Scheme's Textbook
• Teach Yourself Scheme in Fixnum Days Easy to follow. Short.
• Extensive introduction to Scheme
• How to use regular expressions in Scheme.
• Great for understanding macros.
• The Scheme Programming Language Sophisticated, perhaps a little too difficult for the beginner.
• SRFI's: (Scheme Requests for Implementation) Standard ways to do common things that haven't officially been added to the language yet.
• Oleg's page of insightful Scheme code snippets

Prolog

• Schelog: Prolog in Scheme. Prolog is the premier language for computational linguistics (the study of using computers to do useful things with language).
• SWI-Prolog My favorite Prolog.

• Balzac Text Data Mining: How to pick apart Balzac's Comedie Humaine with a computer.
• Balzac's Noses: A Rich Vein of Collocations from the Balzac Corpus
• Collocations in the Original French

Natural Language Parsers

• Apple Pie Parser This parser was a PhD project at New York University. It's reasonably accurate for short sentences and uses the Penn Tree Bank tag set for describing grammatical constituents.
• Parsing as Deduction (A paper + Prolog code)
• Prolog and Natural Language Analysis - Pereira and Schieber
• Natural Language Processing for Prolog Programmers
• Scheme Earley Parser
• Semantics These two online books show you how to write Prolog programs to handle natural language semantics.

The following simple examples use PLT Scheme:

Regular Expressions (using Sitaram's pregexp):

• findall1.scm find all regex matches in a string (like Python's 'findall')
• listtolinks.scm HTML list to HTML list of links
• regexptext2.scm Extracting HTML tags from strings.
• pregexp7.scm Extract all lines in a file that match a regex and write them to another file
• pregexp9.scm Search for all instances of a last name in the works of Balzac.

Text Processing:

• filline1.scm wrapping text into fixed length lines
• split1.scm split a delimited string into substrings
• stringpad1.scm - pad (numeric) string to the left with spaces to make up a string of fixed size (simple version of the srfi for string libraries).

Awk Examples:

• awk2.scm Adding line numbers to the left margin of a file (with proper line number formatting (using string-pad))
• awk1.scm Using Dr Scheme's awk macro (modelled on Scsh) to do the same match as pregexp7.scm above in only a few lines of code

Basic Data Structures

• sort1.scm simple basic sort example.
• hash1.scm simple hash table example
• vectormap1.scm vector mapping functions, timing a scalar multiplying a thousand long vector by 2 (10,000 repetitions) finding average time in milliseconds, (note: should look at the frequency distribution seems to alternate between taking no time and significant amount of time)