Eidola

Massimo Dentico m.dentico@galactica.it
Sat, 10 Feb 2001 22:10:31 +0100


Tril wrote:
> 
> I'm very interested in this due to the focus on independence from textual
> representation.  I haven't read anything except the brief description on
> Slashdot. But I hope to have a discussion later.
> 
> http://slashdot.org/article.pl?sid=01/02/08/2344210&mode=nocomment
> 
> --
> David Manifold <dem@tunes.org>
> This message is placed in the public domain.

The idea of independence from textual representation is very
Tunes-like but the semantic model seems (at a first look)
biased from C++ and Java: not clean, nor simple, nor orthogonal
features. It's not surprisingly because, for example, they judge
C and Java as "clean good" as Scheme; from Eidola FAQ:

  - http://www.eidola.org/faq.html#11

===================================================================
[...]

Why doesn't Eidola have important features X, Y and Z? 

Right now, even in the limited scope of the semantics, there are
some notable omissions (e.g. protected and static members). The
plan is to start simple, and make the language too skinny at first.
In the future, the language will become more robust, but additions
to the language will always be minimal and judicious. The Eidola
aesthetic is toward a small set of powerful constructs. Clean good
(C, Java, Scheme); monstrous and messy bad (C++, Perl, Ada)

[...]
===================================================================

Good intents but they understanding of programming languages
semantics seems a little "weak" (euphemism).

For a better inspiration I suggest to look at Borneo, soon in another
post of mine.



Tom Novelli wrote:
> 
> Seems like good idea but it's too bad that it's being modeled after Java and
> C++, as opposed to a nice simple language like LISP or FORTH or even machine
> code.

I agree completely.

> Then again, the could be another aspect of the database fad and the
> "visual" fad... like the Windows Registry and MS Access.
> 
> I can foresee a problem where multiple representations cause clutter: you
> need to keep track of the position, color, etc. of each "thing", for _every_
> representation.  Like those stupid MS Access queries, where you lose your
> text formatting by using the GUI, and vice versa.  I'm getting by just fine
> with textual code and a few diagrams on paper.
> 
>  - Tom Novelli

Here I disagree. A bad implementation is NOT a proof that an idea
is bad /per se/. In this case, for example, is sufficient to maintain
enough information to reconstruct the original text from an intermediate
representation (in case, you can resort to store the original text _as is_
and attach it as an annotation, with cross references, to the intermediate
representation).

About the importance, from the psychological/cognitive perspective,
of multiple representations see the reference in my post (Tunes general
mailing list) of Thu, 04 May 2000 21:42:25 +0200, "Meta-programming
important for Psychology of Programming". Here is the archived message:

  - http://lists.tunes.org/archives/tunes/2000-May/002789.html

Best regards.

-- 
Massimo Dentico