clipka wrote:

 > you accept as OO only what has a certain runtime structure

No, I call something OO if it has more "OO" to it than (say) stdio.h. :-) It 
needs objects, and data associated with objects, and stuff like that.

> If it /feels/ like OO (i.e. it supports your attempts at implementing an 
> OO design), it /is/ OO.

Tcl's an extensible language. People have written libraries that emulate 
some aspects of OO.  Tcl is no more OO than C is OO because someone invented 
cfront.  Just like the existence of Candygram doesn't make Python an 
actor-based language.

Tcl's OO works by creating a procedure for each object. Invoking that 
procedure causes it to use the first argument as a string to look up the 
method to run, and the name of the procedure itself as an index into a 
global hashtable to find the data associated with the object. I guess if you 
squint hard, you could say it supports OO.  It's not what I'd call an OO 
language, tho. It's a procedural imperative language with an optional and 
not widely used lump of library that rewrites procedural code to look sorta 
OO if you don't peer too close.

> That would (for most part) match "production" language in my set of 
> definitions.

Fair enough. That, then. :)  I was trying to distinguish "one-off 
experimental languages designed specifically to aboid what I was talkign 
about" from "real" languages.  (Altho, honestly, at this point I've 
forgotten what I'm talking about.)

> To be mainstream, something really needs to be... well, you know, in the 
> /main/ stream.

> You wouldn't call some clothes fashion to be mainstream just because 
> people sell the clothes instead of sewing them at home for their own 
> personal use, would you? Or call some musical style mainstream just 
> because you can buy that type of music on audio CDs?

No, but I'd call police uniforms "mainstream" even tho only policemen use 
them. :-)

-- 
   Darren New, San Diego CA, USA (PST)
   I ordered stamps from Zazzle that read "Place Stamp Here".

Post a reply to this message