-=-= a b o u t - h e r e =-=-

A bother.

Something that really bugs me is trying to figure out uncommented code.  And it bugs me even more when I wrote the code but neglected to comment it.

One Issue with OCaml.

Anyone who has programmed in OCaml for any length of time knows the pain that comes from accidentally using types incorrectly in lists.  Since OCaml is (1) statically typed and (2) determines types at compile-time, you have to be consistent when you're using a particular value.  That is, if you have something that's a type (i.e., a record) with {time : float; value : float}.  You can't start using that same record as a tuple (float * float) or else you get an error.  That makes sense.  What often doesn't make sense is that OCaml chooses to tell you that you're using some value incorrectly in the strangest of places--not where you're using it incorrectly, but where it's determined the type. 

So if you use the value in some function in one location incorrectly, it might tell you that you're using it incorrectly.  However under many common circumstances it will tell you that the problem exists in the call to the function rather than in the function itself.  That can get very frustrating when debugging compile-time errors.

Meanwhile.

I find continually switching between writing code in different programming languages enlightening. 

The beautiful OCaml with keyword.

The with keyword has several uses. 

OCaml Macro-ish facilities.

As I learn more about OCaml and the tools available for it, my hunger for more information heightens.  I likely should have known about this stuff years ago.

There's a great way to extend OCaml using metaprogramming via Camlp4.  You can include domain-specific constructs or generate a completely new language that translates into OCaml.  I plan to learn the specifics when I have time (which I hope will be within a month or two).

What's wrong with a really big string?

Is there anything inherently wrong about creating a really big string then outputting it to a file?  Can any general statement be made about this practice?  Does it mean that I "don't have good program design"?

OCaml mode called Tuareg for (X)Emacs.

Often people do well with puns. 

Tuareg works on XEmacs and Emacs as a mode for Objective Caml (OCaml) editing.  It performs much better than the built-in stuff in my experience.  When I'm ready, I'll move on to some of the attempts at IDEs for OCaml.  (Consider Cameleon, for instance.)

Very nice OCaml introduction.

Looking for some overview information on certain constructs in OCaml, I ran across this book called "Objective Caml for Scientists". It's got a great, freely available introduction to OCaml included. (I suggest anyone diving into the world of OCaml read it--avoid letting the title of the book frighten you.) Direct links to the pdf and web versions of the introduction.

Article on type checking (includes OCaml discussion).

Here's a great article comparing static and dynamic type checking. 

There are some obvious weaknesses and biases, but I think it gives a good introduction for those considering switching to OCaml or Ruby on Rails from Java.

Software architect opinion.

There's a great article @ Dr. Dobb's Journal's web site.  An "online exclusive".  It's called "Why I'm Not An Architect".  If you write software in any capacity (e.g. work, research, fun, open source projects, to feel smarter, to avoid or solve personal problems), there's a chance you'll enjoy reading this article.

Why I'm Not An Architect