binding a reference to a variable

[Andrew Koenig]

Tim> I did that a looong time ago, but the code didn't survive my
Tim> then-employer's financial meltdown.  Unless you bought their
Tim> assets at the fire sale, I'm not sure how you could know whether
Tim> your approach is completely different <wink>.

The implementation I was thinking of was actually done in Ada, and
gatewayed to Python...

Tim> It is, though.  The use of unevaluated expressions in SNOBOL4 is
Tim> crucial for advanced matching, and the "how do you do that with a
Tim> bare name in Python?"  problem seems unsolvable in a squeaky
Tim> clean way.  Like:

Tim>     ARB = NULL | LEN(1) *ARB

Yup.  But why not

        arb = null | (len(1) + indirect(lambda: arg))

?

>> Focusing just on the  (break("@") . id)  subexpression, how would
>> one implement it?  That's what led me on this path.

Tim> After I left that path <wink>, I found it was much easier to
Tim> model Icon's spelling of patterns, and with generators in 2.2
Tim> that can be a breeze.

Indeed.  But I want more Snobolish than Iconish patterns.

Tim> However, Icon pattern-matching has a more procedural flavor, and
Tim> SNOBOL4 pattern-matching a more declarative flavor, and each has
Tim> strengths and weaknesses following from that.

Agreed completely.

Tim> Before I left that path, it looked something like

Tim>     id = S.var('id')
Tim>     ... id.ass(S.break('@')) ...

Tim> were .ass() had an optional argument to choose between delayed
Tim> (.) and immediate ($) assignment, defaulting to delayed.  BTW,
Tim> operator overloading was much harder to predict in Python back
Tim> then, so I often settled for ordinary method-call notation.
Tim> After operator overloading because easy to predict in Python, I
Tim> learned to hate it from being forced to endure reading code that
Tim> used it <0.7 wink>.

Maybe there's another way.  If I knew, I'd say what it is :-)

Tim> An iteration before that I used string names as keys to an
Tim> explicit namespace dictionary (as others have suggesting), but
Tim> found that modeling "SNOBOL4 variable" as a class in its own
Tim> right felt at least a little better, despite needing explicit
Tim> "declarations", i.e. lines of the form

Tim>     var1, var2, ... = S.var(name1, name2, ...)

Tim> Remembering to do that once was less painful than needing to slop
Tim> in quotes everywhere.

Might well be.

Tim> Have fun!  It's an interesting challenge.

Indeed.  And a good learning vehicle.


-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark