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

Re: [Scheme-reports] Boolean hemlines

On Sat, Apr 14, 2012 at 11:40 AM, John Cowan <cowan@x> wrote:
> Alex Shinn scripsit:
>> There was no keyboard involved - it was a matter of skimming
>> through a log of results from different implementations and not
>> being able to recognize one.  As you said:
>>   [Cowan]: [Summary of which impls return #t for eq? for
>>   empty strings and for empty vectors, saying most impls
>>   return #f for both and listing only exceptions].
>>   [Shinn]: You missed Chibi, which returns #true for vectors and
>>   #false for strings and bytevectors.
>>   [Cowan]:  Right; it didn't jump out of the log.  I went back and
>>   scrutinized the log more carefully, and there are no more cases.
>> This is a perfect example of what I'm talking about -
>> long lists of alternating #t and #f are hard to read.
> What didn't jump out of the log was Chibi, because it doesn't print a
> banner on startup.  After that I changed my script so that it begins by
> outputting a great big comment block saying "RUN ALL SCHEMES" and then
> outputs a smaller comment block containing the name of the implementation
> before running each Scheme.  That way I am less likely to miss any
> Schemes altogether.  That had nothing to do with confusing #t and #f.

It's not a matter of confusing them - the problem
is that the difference doesn't jump out at you.  If
you had been scanning a log of all 0's and a 1
jumped out, you couldn't possibly have missed it,
and then would have looked up to see which
implementation generated it.


Scheme-reports mailing list