OpenEuphoria

there still is a case for CASE

• Posted by Mike <mdeland at NWINFO.NET> Mar 08, 1998
• 760 views

Irv wrote:
-------------------------------------------
The real reason for wanting a 'case' statement was to
simplify long if .. then .. elsif .. then .. else  .. end if
statements.
>>I'd like to clarify that a bit...simplify, in this case,
>>to me, is to simplify _understanding_ or
>>_readability_....not typing necessarily.
>>when I read a case statement like the one
>>below, I can figure out what it's doing much
>>faster and easier...they are far more instantly
>>intuitive than their counterpart if.then.elsif
>>(to me that is)

 (Not that they're all that bad, compared with other
languages,  thanks to the elsif!)
>>no. they are not bad at all. I keep wondering,
>>as Jiri intuned below, if we should 'bother'
>>trying to make this case statement...
>>{continued below}

--Pascal version:
   case key of
             esc    : fini := true;
               'P'    : dec(indx);
               'N',^M : inc(indx);
               'F'    : indx := 0;
               'E'    : if editmode then SO_DATA.MENU;
   end; {case key}

Anyone want to guess which is
(1) easier to understand
>>the real goal IMHO
(2) less typing
>>I could care less, frankly, about typing. I took a typing
>>class and saving typing is not my primary concern.
(3) less likely to have "unexpected consequences" (bugs)?
>>the other half of the real goal. if coding is intuitive,
>>and reading that subsequent code is intuitive,
>>bughunting becomes almost trivial

>>{continued}
>>however, when I see the above as compared to the
>>if.then.elsif that Jiri proposed, I realize that it
>>is indeed worth it to attempt to develop a coding
>>structure that mimics case.
Whatever procedure we come up with would have to be clearer
and shorter than the Euphoria if.. elsif.. code, or it wouldn't
be worth doing.
>>agreed. Pete's code is really good code IMHO.
>>it's fast at runtime, neat and clean....
>>alas, it's also a bit cumbersome...
>>gotta keep looking
--------------------------------------------------


Jiri wrote:
--------------------------------------------------
    if key = esc then fini = TRUE
    elsif key = 'P'  then indx = indx - 1
    elsif key = 'N' or key = ctlM then indx = indx + 1
    elsif key = 'F' then indx = 0
    elsif key = 'E' then
         if editmode then SO_DATA_MENU end if
    end if

Slightly more typing, yes, but you may even change
your mind about the rest.
>>not worried 'bout the typing personally....
>>yes, that is clean and neat code, but it
>>still takes me longer to decipher...
>>it's just not quite "instantly intuitive"
-----------------------------------------------------

>>another way to potentially implement the case
>>may be with a preprocessor.  I don't know
>>much about preprocessors (PP) as they relate
>>to Euph, but it's my understanding that
>>david cuny (i believe) has written one that
>>is quite good.  haven't had time to tinker
>>with it.
>>the PP could (i'm guessing) translate the
>>case code to if.then.elsif to a temp file
>>that euph runs instead of the main .ex file
>>this way, we don't have to over complicate
>>euph with features that aren't wanted by
>>the masses or with features that may
>>slow down run time....best of both worlds
>>perhaps?
>>Mike
(sorry for the reverse >> quoting, only way
I could include both irv and jiri's messages)



--------------------------------------
If it wasn't for typos,
I'd never get any coding done at all.
;)