retro html formatted by nom
comments/suggestions to: mjb at nomlang dot org
Its better to know a knot and not need a knot than to need a knot and not know a knot. Scouts Lore
About regular expressions in the nom language (there aren't any).
This could be the shortest document here: There are no regular expressions in the nom language and the pep virtual machine . This may seem a strange and self-defeating design decision, but actually I believe that it is a stroke of genius.
Regular expressions can be slow and they are also quite bulky to implement. I believe approximately 4000 lines of c code is required to implement a regex engine. I note with interest that the inventor of Lua also decided not to implement a full regular expression engine in that language.
But there are other reasons, not to include regexes in the nom language and some of them are psychological. It seems to be a great temptation for programmers to try to make context-free patterns with regular expression patterns, even when they know, or have been told that it is not possible.
Most parser generators (such as ANTLR ) include support for regular expressions during the lexical analysis phase of the parsing and compiling process. Surprisingly regular expressions do not seem necessary for most parsing applications, although sometimes their absence makes the grammar more complex.
For example, the nom script /eg/text.tohtml.pss does (in my opinion) a quite reasonable job of turning plain text patterns into html without the use of regular expressions (this script is the engine that renders the website www.nomlang.org where this documentation is located).
The lack of regular expression support in the “lexing” phase of the nom script means that the implementation of pep and nom is considerably simpler than other tools and this is one of the primary aims of the ℙ𝕖𝕡 🙵 ℕ𝕠𝕞 system: simplicity.