> 
> As a Tcl'er, I reckon that even "intval" and "floatval" are not
> *sufficiently* verbose.  Were this following the principles set out in
> http://tcl.sourceforge.net/engManual.pdf (which have worked very well
> in Tcl development over a long period) I'd be reckoning on calling
> them something more like "intValue" and "floatValue" (they are members
> of a structure/union, aren't they?)  That style of naming might take
> longer to type, but it takes much less time to read and understand,
> especially for someone who is not a deep expert in that code.  :^)

The use of too longs names force large statements to be broken in
many, forcing to introduce intermediate variables and making things
even worse.  Then, larger programmatic structures do not to fit in a
window/page. As a result the programmer/reader must scroll back and
forth the program to capture  the larger scale programmatic  structure.

In other words: on a small scale structures are less visible because
drowned in "lexical sugar".  On a larger scale, structures that would
fit in a page must be remembered to get the program "general meaning".

Remembering the use of common identifiers is easy, figuring out the
structures of a program is difficult.  So large identifiers make easy
thing easier and complex things more difficult. Not a good thing.

Also, the trade-off choosen by Perl programmers is to use more complex
syntax structures than most languages to get visually distinctive
programmatic structures.

 Program are a form of knowledge.  It is all about smart compressions
 and shifting costs from repetitive tasks (writing and reading
 programs) to less common ones (learning a complex language, learning
 common identifiers...)

Traditional metrics don't apply. It may take longer to write or
understand one line of Perl program, but less to figure out the whole
program.

Parrot is programmed in C. So the general trade-offs are different but
the general principles are the same.



> 
> Donal.
> -- 
> Donal K. Fellows, Department of Computer Science, University of Manchester, UK.

-- 
St�phane Payrard -- s.payrard@@wanadoo.fr

Reply via email to