The imp of the perverse

[Pierre Sahores]

Hello Alain,

A realy cool input you posted there :-)
> 
> Hello Pierre and y'all,
> 
> > Perhaps is it ways to code abstactions in C, Lisp
> > or Smalltalk. Probably is it many more natural to
> > do this in using an xTalk. It's just the reason i
> > use, instead of any other tool each time it's
> > possible, an xTalks to build my apps. Are there
> > any others, here, doing same for the same reasons ?
> 
> I refer to this as being in my "zone". Many others
> refer to it as "flow". Whatever you call it, it's an
> awesome experience, that can only be obtained after a
> long quest for knowledge and continuous mastery that
> takes years to achieve. You think of a problem and,
> voila, the solution pops up into your mind as if by
> magic. Someone describes the symptoms, and you
> spontaneously *know* exactly what the problem is AND
> how to fix it, even when the cause-effect relationship
> is not evident at all and/or even if this problem has
> never been encountered before.

Exactly that and, for sure, one of the reasons we gets more satisfaction
in building softs than in playing Chess !
> 
> It's a proven fact that most programmers become
> notably attached to the first language they learned.
> Kind of like a chick (baby duck) which bonds with
> whatever thing is the FIRST that it sees. For many of
> us xTalkers, our 1st experience with programming was
> scripting HyperTalk. Only natural, therefore, that it
> seems so "natural" to us.
> 
> OTC, I don't believe that languages are all the same
> in terms of ease of learning, daily use, power, etc.
> Low-level languages are more powerful, but their ease
> of use & their steep learning curve are very
> prohibitive. They presume that software is to be
> "engineered" and that all needs/specs/etc must be
> established before coding. Real life is seldom that
> straightforward though. I believe, along with lots of
> new "Agile" developers, that this old unrealistic
> approach is obsolete. Moreover, I believe that we have
> reached a point where our micros provide us with
> enough power to dedicate some of this capacity to
> making the software/systems easier to craft. Easy
> enough for casual developers. With xTalk, for ex, we
> don't have to declare the type of our variables before
> execution of the handler per-se, or casting from one
> type to another during execution. We also don't have
> to worry about RAM management. And a lot of other
> low-level annoyances that distract us from the goal of
> the program/stack.

Thanks a lot for explaining there what makes the fondamental and corely
difference betwin the two ways to gohead in AI, computers
sciences/development. What you says seems me specifially true in about
the web/vpn applications and databases servers development : java, c# 
and asp.net, perl or, even, the so usefull php, on the first side, mc,
rev, rexx, the unixes shells, the <select, insert, update, delete> main
part of the ansi sql, omnis studio or labview on what i belive to be on
the road of future, aka on the right side...
> 
> Consider the ease with which we can deploy our stacks
> on all the relevant platforms (Win, Mac, Linux, UNIX,
> etc). Porting from one platform to another is hell
> otherwise. You have to be very familiar with all the
> ins and outs of both/all the platforms. Coming from a
> Mac background, I find the propect of fiddling aroung
> with win's DLLs & such is a daunting one to say the
> least. But there's no need to know or deal with *any*
> of it when using MC/RR; not even any rescripting or
> reconfiguring ... nada ... just deploy it as-is.
> Awesome!

Hypercard maked me a MacOS 7/8 fan. When i became a Rhapsody
beta-tester, i could'nt understand what Apple was searching for. The day
i took my initial Metacard license, i tested mc under macos, win32 and
linux and got, some days later, my first suse-linux aware x86 laptop ;-)
 
> For speed, code-hiding, functionalities not provided
> by the xTalk, the [casual] developer can develop small
> code snipets in Pascal, C or other languages. I am
> referring to XCMDs of course. Not a piece a cake for
> mom-and-pop, that's for sure, but they are still far
> simpler to make than full-fledged programs in the same
> languages. You only have to learn/add what you need,
> as you need it, and many are already available in the
> HC community.

Do you remember Alain, Scott, about the so usefull Rinaldi's "HowMany()"
XFCN ? It would be realy great to have it usable as a native function,
from within Metatalk !!!
> 
> Okay, that's it. I will leave it at that.
> 
> xTalks are the best for most of us!  ;-)
> 
> Alain Farmer
> 
So, Friends :-)
--
Bests, Pierre Sahores

Inspection académique de Seine-Saint-Denis.
Applications et bases de données WEB et VPN
Qualifier et produire l'avantage compétitif