OpenEuphoria

2. Re: under_score

• Posted by Derek Parnell <dparnell at BIGPOND.NET.AU> Sep 17, 2000
• 389 views

----- Original Message -----
From: "Bernie" <xotron at PCOM.NET>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Sunday, September 17, 2000 11:38 AM
Subject: under_score
> --
>    How much trouble would it be to allow underscore characters
>    at the begining of a identifier, this would make interfacing
>    with "C" alot more readable and easier. There is no way to
>    define constants like :
>    constant __a_c_constant = #12345
>    when defining a "C" header for a library.
>    Sure you can change the name but then you are deviateing from
>    the library's documentation and you cause unneeded confusion for
>    the user of the Euphoria header.
>
>    Bernie

I agree Bernie. This artificial restriction seems a little too picky.

Most language translators reserve "non-numeric identifier characters" as the
first character in an identifier to make it easier to distinguish numeric
literals from identifiers. The parser examines characters left-to-right, and
if at the beginning of a word, it hits a numeric character, assumes that a
numeic literal is coming, otherwise it tries to see if it's a valid
identifier name.

In Euphoria's case, the underscore is a valid non-numeric identifier
character, so it seems perfectly reasonable to allow it as a first
character.

[Mildly related off-topic follows....]
I also work with a language (www.progress.com) that has a dash '-' as an
allowable indentifier character, and so to make it easier for its parser,
the language syntax insists that math operators be surrounded by spaces, so
it can distinguish between a minus sign and a dash inside an indentifier
name. I think that this was a bad choice of identifier characters. Perhaps
they did this so you didn't have to press the Shift key to get the
underscore character.

----
cheers,
Derek