PTC and "SHOULD" vs "MUST"

David Herman dherman at
Wed Jun 1 17:55:22 PDT 2011

Agreed. Should be "must."


On Jun 1, 2011, at 3:42 PM, Brendan Eich wrote:

> Thanks for catching this, it goes back to (first rev). Cc'ing Mark. I bet it is just a thinko for MUST and we can fix it quickly.
> /be
> On Jun 1, 2011, at 3:36 PM, Peter Michaux wrote:
>> I notice that in the proper tail calls wiki page that a compliant
>> interpreter "SHOULD implement a call in proper tail position as a
>> PTC." In order for a programmer to use proper tail calls for
>> arbitrarily deep recursion, the programmer needs a guarantee that the
>> interpreter will use proper tail calls. This would mean the "SHOULD"
>> needs to be a "MUST". Otherwise what does the proper tail calls
>> proposal gain the programmer?
>> For example, R5RS has much stronger language compared with "SHOULD"
>> about Scheme using proper tail calls.
>> "Implementations of Scheme are required to be properly tail recursive."
>> With that language the programmer using Scheme can then depend on that feature.
>> Peter
>> _______________________________________________
>> es-discuss mailing list
>> es-discuss at
> _______________________________________________
> es-discuss mailing list
> es-discuss at

More information about the es-discuss mailing list