[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] [scheme-reports-wg1] Arthur Gleckler's rationales for 4th ballo votes

Am 31.08.2011 08:55, schrieb Alex Shinn:
>> Arguably, R7RS is not a revision of R6RS (or R5RS, for that matter) in
>> the sense of "one document created by augmenting and/or editing
>> another", but more of an reorganization in that the original single
>> report is now split into *two* reports. This introduces a whole new
>> notion of conformance levels and what consitutes a Scheme implementation
>> - which, I'd say, is a deep structural change. Calling this a "revision"
>> is, in my opinion, a stretch.
> It's much less a reorganization than R6RS was, which
> itself is divided into two separate reports.

I wouldn't say that R6RS consists of multiple reports, but of multiple 
documents forming a single logical report; the naming scheme of these 
documents support this view. The point is that you still have to support 
the functionality in all documents to be an R6RS Scheme, while in R7RS, 
we have a separation in multiple conformance levels which is expressed 
by the split into two reports. Therefore, I'd call the reorganization in 
R7RS much more substantial, because it is not only cosmetic.

So I'm not basing my argument against the "revision" term on some vague 
notion of edit distance, but on the fact that the upcoming reports 
present a substantial semantic restructuring in that they redefine what 
Scheme actually is - a "language tower" rather than a single language, 
if you will.


Scheme-reports mailing list