Re: OpenEU
On Mon, Aug 18, 2003 at 10:18:28AM +0000, Deric Wechter wrote:
This is just one man's opinion of course, but I find it a bit disturbing
that an Open Source Euphoria is being pursued.
I can emphasize and relate to the frustration of having a product you
love not getting the attention you think it deserves as I've been in
similar positions before with other products and have considered the
idea of copying them. The idea is seductive when something you care
about is not living up to what you perceive as its' potential.
However, all legal issues aside I personally think it's extremely
unethical. Euphoria clearly exists because of Robert Craig. The fact
that you're not satisfied with how he handles his product isn't a valid
reason to attempt to hijack it.
I've already responded to this in another email. In short: we're not
stealing it, we're forking it so we can make it better. Maybe then RC will
get it into his head now Euphoria can be improved and actually do it.
If you must take this drastic step, why not create a new language which
includes Euphoria-like features? All programming languages, including
Euphoria borrow concepts from previous ones however they don't amount to
exact replicas of each other.
OpenEuphoria will by no means be an exact replica of any language. If you
read past emails in the openeu list you will see that it is a very different
language indeed... it will duplicate enough to run all existing Euphoria
code (so we don't lose all those open source Euphoria contributions, especially
the ones like Win32lib which are considered by many to be an essential part
of the language, even if it isnt part of the language).
Almost all Euphoria projects are open source. Many open source movements have
born and died around Euphoria.
Also, this is not the first Euphoria clone. If you read past emails you'll find
out about peu, gnuphoria, eu.ex, Eu++, etc. So what is being done here isn't
even anything new.
While Mr. Craig is in this for profit, his prices are very reasonable.
It's certainly possible to pay a lot more for less. As is often
mentioned here, it's also an extremely stable product.
No one wants to see RC lose money. The point is, he already has lost money
because he didnt more pay attention to his users.
As Irv stated, perhaps the threat of OE will change RDS's views ... in
which case OE wouldnt be needed anymore.
I know one could argue that all of the major established languages are
not solely in the hands of their creators but most of them were
developed under different circumstances and with a few exceptions they
tend to be a lot older than Euphoria is.
On the other hand, the fact that users clearly are this frustrated with
his nurturing of the product should prompt Mr. Craig to consider working
with the users to implement reasonable improvements.
And yet, the fact is that this is not so. (At least, not to a large enough
extend to make most people happy.)
While some suggestions such as GOTO and possibly Pass by Reference may
not fall in line with his vision of what the language should be,
certainly many reasonable improvements have also been suggested, such as
assigning values to variables in the statements which you declare them
in. Improvements such as that one just make too much sense any way you
slice them for them not to be implemented.
I'm with the Pass by Value camp on the Pass by Reference discussion. It
it rarely _necessary_ to _need_ to resort to using global variables.
Convenience issues aside, almost always there's a Pass by Value
solution.
Since you can get a Pass by Reference effect with global variables,
there is a workaround if you prefer to do things that way. It may not
be pretty or best coding practice, but many don't consider Passing by
Reference to be best coding practice either.
Regardless where you stand on the issues of the language, nothing shall be
done. RDS isn't going to add in new features just because you said they are
a good thing to implement, even if they really are.
As far as strict data types go, it would be ideal if it were made a
seamless option. Sequences as we know them are indeed a double-edged
sword. The beauty and burden of sequences is how flexible they are.
The programmer is often in a position where he or she has to worry about
the type of data in a variable so the headache we're spared in
definition and scope is considerably offset by the headache of manually
ensuring the proper kind of data is being entered.
While not needing to manage various data types is an intended feature of
Euphoria, having the ability to optionally assign each atom in a
sequence a definite data type would only serve to enhance the language
while retaining the ability to use sequences as they are now.
Hypothetical Examples:
atom tatm = type_char
sequence tseq = {type_char, type_str, type_int, type_real}
if tatm = type_char then
...
end if
Even just having an optional int and real number type would make a big
difference. Character and String types could be used symbolicly while
still actually storing the data as sequences of ASCII codes.
That way, one could still ensure they're storing the type of data they
want to without fundamentally changing the way the language works.
Coders wouldn't suddenly start strict typing every sequence either.
It would be particularly useful for handling user input as Irv pointed
out.
Concurred, tho I'd prefer a different syntax. However, I doubt RDS will
add the feature you just suggested here to Euphoria either. Your chances are
much better with OpenEuphoria. Which is why it exists (or will exist) in the
first place.
-Deric Wechter
jbrown
TOPICA - Start your own email discussion group. FREE!
--
/"\ ASCII ribbon | http://www.geocities.com/jbrown1050/
\ / campain against | Linux User:190064
X HTML in e-mail and | Linux Machine:84163
/*\ news, and unneeded MIME | http://verify.stanford.edu/evote.html
|
Not Categorized, Please Help
|
|
