Re: Standard Euphoria Library project

new topic     » goto parent     » topic index » view thread      » older message » newer message

Jason Gade wrote:

> Juergen Luethje wrote:
>
>> Christian Cuvier wrote:
>>
>> <snip>
>>
>>> The Standard Euphoria Library project on Sourceforge is dead.
>>> The OpenEuphoria group, which is the one you're referring to, is dead as
>>> well.
>>> Its homepage is at oedoc.free.fr actually.
>>>
>>>> Now I wouldn't want to organize such a thing because I am not an
>>>> experienced
>>>> developer.  But I would want to contribute ideas, documentation, or code.
>>>
>>> I was originally the doc manager for OpnEU. I wound up docig sequences
>>> in C++ when the group imploded. Now I'm admin of a souceforge defunct
>>> project. Wanna get in?
>>
>> I'd like to join the Standard Euphoria Library project. smile
>
> I'll look at the OpenEU docs for ideas.  I think many features
> requested in the language could be implemented in a new standard
> library.
>
> What needs to be done is look at what kinds of things users need, what
> kind of facilities exist in other languages

Yes, that is one of the meanings of the term "standard". There are
several functions -- such as abs(), ceil(), min(), max() -- that
exist in many other languages. Those functions should be contained in
our Standard Euphoria Library.

> (either in their library or as a standard part of the language) and
> figure out how the library should be structured.
>
> Find which parts of the library are already written and in the archive,
> and include them as needed. (strtok, win32lib or wxeuphoria, etc.)

IMHO win32lib or wxeuphoria or any other GUI library should *not* be
integrated in the Standard Euphoria Library. At least not yet, maybe in
version 10.0 or so. smile

IMHO the Standard Euphoria Library should be modular with a great
granularity, and it also should be 100% cross-platform for all
platforms supported by Euphoria.

We should make a plan and consider some things, but in order to avoid
what happened to the OpenEu project, we should not discuss too long, and
try to release rather soon a first simple version. Like all other people,
we'll need a feeling of success now and then.
Includeing a GUI library is rather complex, and will take much too long
time.
But important would be IMHO to *supplement* one or more GUI libraries.
E.g. AFAIK Win32lib does not only contain GUI routines, but also file
handling routines.
It would be very good, if our Standard Euphoria Library would contain
some standard (sic!) file handling routines, which can be (and hopefully
will be!) used by Win32lib, ARWEN, etc.
So another meaning of the term "standard" is: Suited to the needs of
other programs and libraries, and accepted by their authors, so that the
authers don't see a need for duplicating the code. Not each GUI library
needs its own file handling routines. Trying to avoid things like that
should IMHO be one goal of a Standard Euphoria Library project.

> I think one of the "gotchas" in doing it would be interdependencies
> between functional areas of the library.

I also think so. We should start with simple routines with only a few
interdependencies. The granularity of the library should be great.
For instance we could start with a 'math.e' file that contains some
general well defined functions -- mathematicians tend to define their
"stuff" very well smile -- such as abs(), ceil() etc., and a 'sets.e' file
( the existing 'sets.e'file in the archives is not very good ), and say
2 or 3 more .e files, and a good HTML documentation.

> One idea is to reference Win32lib in order to see how a good library is
> written and structured.

As far as I know Win32lib is one "monolithic" library. That is what I
certainly do not want a Standard Euphoria Library to be. IMHO it should
consist of several include files, which are as independent as possaible
from each other. A better model IMHO are the official RDS include files
that are contained in the Euphoria distribution.

> I'm willing but my experience is low.  I can't guarantee that I won't
> lose interest but I would like to try.  I think it will require much
> research and planning before any lines of code are written.

We should'nt wait too long until the first release. It is necessary to
establish the project. Currently, even someone else has taken the name
"Standard Euphoria Library" for his own project. I hope he will change
the name soon, and leave the name where it belongs.
But this shows, how important it is to actually *create* a project, and
not only to talk about it.

Regards,
   Juergen

new topic     » goto parent     » topic index » view thread      » older message » newer message

Search



Quick Links

User menu

Not signed in.

Misc Menu