[SEMI-OT] On The (Non) Value of Programming Language Research

[jbritt]

There have been periodic requests for a formal specification of the Ruby
language. As it is, the C code itself is the most complete spec.

Erik Meijer [0] has an interesting entry on his web log [1] about why
smart language designers do not write formal specifications.


James Britt




[0] http://research.microsoft.com/~emeijer/
[1]
http://blogs.gotdotnet.com/emeijer/PermaLink.aspx/c15c9d4c-e0f7-44b6-8de6-e857910940f3

 

[Nathan Weston]

When people asked for a spec of Ruby, I always thought they meant
something like the Java language spec
(http://java.sun.com/docs/books/jls/second_edition/html/jTOC.doc.html).

In the article, OTOH, when he says "formal" he means something along
the line of an operational/denotational semantics (which for the most
part is comprehensible only to professional PL researchers).

Or maybe I'm wrong, and everyone's clamoring for an operational
semantics of Ruby...

Nathan