[Lars H], 2008-08-08:
In comparison to many other languages, Tcl has a tendency to require that you [Write What You Mean], as opposed to just giving an indication of what operation is wanted. I like this, because it means Tcl won't suddenly decide upon doing something other than I what I intended, just because there happended to another operation which looked sensible in that particular case.

For example, the core language has [append] and [lappend] as distinct commands, because they do different things — append as substrings, or append as list elements. In other languages, especially [OO]ed languages, there might rather be a single "append" operation, and it depends on the thing to which one is appending what operation is actually performed. In other words, I can't write what I mean, but rather have to give an indication of what I mean, which the language then has to interpret.

[NEM] All meaning depends on interpretation, so you always have to know the semantics of
the language you are using. Your argument seems to be against [polymorphism] and other forms
of sophisticated language semantics. Generally, such mechanisms are seen as a good thing;
leading to better modularity and programming to interfaces rather than specific implementations
(e.g., I don't care what container I get passed, so long as it supports insert and remove operations,
and behaves like a set). Of course, such mechanisms are not always well designed or well applied, 
but your argument seems to be stronger: that they are never a good idea? Are you also against
abstract data types?

----
!!!!!!
%| [Category Advocacy] |%
!!!!!!