OpenEuphoria

Re: Universal Namespace Clearinghouse

• Posted by daryl_vdb at HOTMAIL.COM Jun 23, 2001
• 377 views

I agree.  Having include files declare their own namespace is a bad idea.  
It should be left up to the file that is including it.

There is no real difference between:

win.openWindow(window, style)

and

win_openWindow(window, style)

if win32lib makes its own namespace for all its routines.  The first example 
is with a renamed namespace, the second example just has every procedure, 
function, constant and variable in win32lib renamed.  The second example can 
be done in the current version of Eu if someone would rename all the 
routines in win32lib.  But it doesn't solve anything.



>----- Original Message -----
>From: <president at insight-concepts.com>
>
>Subject: Re: Universal Namespace Clearinghouse
>
><You say this with tongue firmly planted in cheek, I trust?>
>
>I did not understand your comment............Can explain?
>
>Surely you are joking - I certainly was. A "clearinghouse"
>won't work. The idea just barely works in a large COBOL
>shop where people are well-paid to cooperate, and the compiler
>enforces those sorts of things. There's exactly zero chance of it working
>in the real world. Besides, as explained in my earlier post,
>and Travis alluded to also, having namespaces fixed into
>include files just delays the inevitable.
>
>Suppose I create an include file which is useful, and everybody
>wants to use it. I create it as namespace 'win', and send it to RDS.
>Maybe I shroud it first. It automatically conflicts with
>Win32Lib, and there isn't any way to change my code.
>
>No one is going to change Win32Lib's namespace,
>because then all programs that depend upon that will break.
>
>Even if you were willing and able to change the namespace of my
>library, how would  you know that doing so wouldn't
>break all other libraries and programs that might depend upon it?
>Including programs you wrote yesterday, before you modified
>the namespace?
>
>See the problem? We'd be right back where we started.
>The only workable solution is to name the space at the
>time it is imported, with a name that would be meaningful only within the
>scope of that one file. If other files also import the same library, they
>are free to rename it however they wish, and there will be no conflicts.
>
>Regards,
>Irv