[plt-scheme] [redex] language keywords

Is this 'where' clause something that was introduced somewhat recently?
I remember from my first redex examples that the way to introduce a new identifier for instance was with ,(term-let ...).

The same question goes for the definition of metafunctions. For instance, the definition of lambda-v on the web:
http://redex.plt-scheme.org/lam-v.html
for the subst metafunction, uses ,(term-let ...) to introduce (x_new ...) 

It seems the preferred way to do that now would be with a where clause, right?

(I was surprised actually when seeing in the doc that term-let was low-level and not meant to use, while it was--and still is in some cases-- used in the simple introductory examples of plt redex)

Thanks,

-- Éric


On Jun 3, 2010, at 9:15 AM, Robby Findler wrote:

> Oh, great!
> 
> Robby
> 
> On Thu, Jun 3, 2010 at 8:12 AM, Eric Tanter <etanter at dcc.uchile.cl> wrote:
>> Thanks all,
>> 
>> I'm moving to 'where' everywhere, and it seems to work. It's even much more readable with where clauses than with the collections of (nested) term-lets that I had.
>> 
>> -- Éric
>> 
>> 
>> On Jun 3, 2010, at 2:36 AM, Robby Findler wrote:
>> 
>>> It would be good to avoid term-let, I think, if you can use things
>>> like 'where' in the reduction relation instead. But we'd probably need
>>> to see an example use if you don't see how to do the transformation.
>>> 
>>> Robby
>>> 
>>> On Wed, Jun 2, 2010 at 9:51 PM, Eric Tanter <etanter at dcc.uchile.cl> wrote:
>>>> Yes, I do use term-let and that's (at least up to now) exactly where the
>>>> problem manifests!
>>>> 
>>>> Should I be using something else?
>>>> 
>>>> -- Éric
>>>> 
>>>> On Jun 2, 2010, at 22:40, Casey Klein <clklein at eecs.northwestern.edu> wrote:
>>>> 
>>>>> On Wed, Jun 2, 2010 at 9:21 PM, Carl Eastlund <cce at ccs.neu.edu> wrote:
>>>>>> 
>>>>>> On Wed, Jun 2, 2010 at 10:15 PM, Casey Klein
>>>>>> <clklein at eecs.northwestern.edu> wrote:
>>>>>>> 
>>>>>>> On Wed, Jun 2, 2010 at 9:02 PM, Eric Tanter <etanter at dcc.uchile.cl>
>>>>>>> wrote:
>>>>>>>>>> 
>>>>>>>>>> Is there something as simple/direct as the list of keywords in
>>>>>>>>>> syntax-rules/syntax-case?
>>>>>>>>>> If not, any alternative suggestion?
>>>>>>>>>> 
>>>>>>>>>> Thanks!
>>>>>>>>>> 
>>>>>>>>>> -- Éric
>>>>>>>>> 
>>>>>>>>> Take a look at variable-prefix.  In my latest Redex work, variables
>>>>>>>>> all start with x, y, or z, and I just have to avoid any keywords that
>>>>>>>>> start with those.
>>>>>>>> 
>>>>>>>> Thanks Carl,
>>>>>>>> 
>>>>>>>> I'm using these as well, actually.
>>>>>>>> A problem I have is that, for readability, I have a non-terminal which
>>>>>>>> is define as (x >> x) where '>>' is a keyword. I know I could just use (x x)
>>>>>>>> for exactly the same, but it does enhance readability.
>>>>>>>> The problem is that >> is considered as a pattern variable...
>>>>>>>> 
>>>>>>>> with macros I could just say:
>>>>>>>> (syntax-rules (>>)
>>>>>>>>  ((x >> x) ...))
>>>>>>>> 
>>>>>>>> Any idea?
>>>>>>>> 
>>>>>>> 
>>>>>>> I don't follow. The variables in a Redex pattern are always named
>>>>>>> after non-terminals, so I don't see how >> could be treated as a
>>>>>>> pattern variable.
>>>>>>> 
>>>>>>> Do you mean that >> matches one of your six classes of variables?
>>>>>> 
>>>>>> Is it possible Eric is using one of the forms that match using
>>>>>> syntax-case instead of the Redex pattern matcher?  Do term-let and
>>>>>> where clauses still work that way?
>>>>> 
>>>>> Oh, good thinking! term-let uses syntax-case, but where clauses no longer
>>>>> do.
>>>>> 
>>>>>> (And if not, Eric, what version of PLT Scheme are you using?)
>>>>>> 
>>>> _________________________________________________
>>>>  For list-related administrative tasks:
>>>>  http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>>>> 
>> 
>>