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

Re: [Scheme-reports] *TELUS Detected Spam*Re: [r6rs-discuss] returning back to pattern matching

On 2010-12-23, at 15:36, Alex Shinn wrote:
>> How do you think it will genuinely place no burden on implementors?
> Because they can just use the reference implementation.

If I were an implementor, I would try very hard to use the reference implementation for such things, on the grounds that (a) doing so would avoid my introducing bugs in writing my own code, and (b) if the result has poor performance or is unreliable, that exposes bugs or deficiencies in my implementation. 

Given that thought, then I question whether such a library ought to be part of a Standard. Standards are supposed to specify constraints on independently-produced pieces of software. The loop macro, and many other such things, aren't that, they are instead part of a standard library that implementors are likely merely to include with their systems. 

It seems to me that there are two kinds of documents here. The WG2 document(s) will comprise a normative standard, while there might be an additional document that describes the WG2 Scheme Library. This would include not only the implementation, but also appropriate test code. This helps the implementor limit the scope of his or her work, though it does require some (small, I hope) ongoing maintenance work to fix bugs detected in the Library. Nothing, of course, stops an implementor from customizing the Library, but one would hope that's unnecessary. (Of course, implementations might include or offer many other libraries besides The Library.) 

I don't think this changes anything in what WG2 is doing, but organizing things this way has a considerable advantage in helping people understand issues of standards conformance. 

-- v
Scheme-reports mailing list