OpenEuphoria

Re: Upcoming releases

• Posted by Ralf Nieuwenhuijsen <nieuwen at POP.XS4ALL.NL> Jun 11, 1997
• 1197 views

Wow, it's gonna be lightning-fast...


-=-=-=-=-=-=-= Before read past here, know that is gonna be long...
-=-=-=-=-=-=-= It's about possible new improvements to Euphoria..
-=-=-=-=-=-=-= If you don't care or don't want to read long messages
-=-=-=-=-=-=-=          PLEASE DONT READ IT!!!

-=-=-=-=-=-=-= (Is this the way, to solve my long message tick?)

        About the interpreter () routine, can't you just make an include
statement inserted like this:

        if  [some condiftion] then
                                include [somefile].e
        else
                                include [otherfile].e
        end if

        It would end all trouble, best is to have a include routine which
you can give a code-sequence containing the code.

        An other suggestion is atoms with a unlimited number of bits...
        So you can have quicker acces to a big package of data cause it
doesn't need the flexible sequence structure. It saves time and
memory and you can then add a lot of nifty tricks...
        Like getting some value of each 8 bits and then of each 16 bits with
no converting time at all!!!
        Or make a table that simply resizes it's dimensions with almost no
cpu!!

        And the NULL value, whenever some routine (intern or not) is called
with a NULL as a argument, it is ignored...
        Then when you poke for example, the pixels with the NULL value would
not be poked, this could make it all a lot easier....

        Also have some more preprocces commands, like enums and partners
(different var-name, same memory addres) and select case (like in
qbasic) and many others...
        Look at the preproccesor of David Cunny, it genuis, at least include
'with each .. in ... do'  constructs, they are great!!
        Having more clearly code will have a great impact on Euphoria..

        Like i said, OOP could easily be done by a preproccesor, why not
make Euphoria more OOP, so you can have clones of .E and collections
(sequences with search 'keys')

        Maybe a very quick inline compiled when loaded machine independent
but very low-low-close to the system- language can be included.
Perhaps a routine declared as 'system' or something.. ..it could
easily be possible and if the low-low code is then compiled for the
current machine, machine-specific code optimizations can be done
which can make Euphoria faster than any other language compared on
not typical default systems....

Ralf Nieuwenhuijsen
nieuwen at xs4all.nl