Integer oddity

> On 07/12/2011 02:16 AM, Steffen Märcker wrote:
>> Hi,
>>
>> I've just observed that Integer zero/unity evaluates to
>> LargePositiveInteger zero/unity. This leads surprisingly to
>> Integer zero ~= 0.
>> Two questions here:
>> 1. Is it useful to get large integers here?
>> 2. Do we actually want LargePositiveInteger zero ~= 0 (or unity ~= 1)?
>
> This definitely seems like a bug.
>
> For
>
>     0 = Integer zero
>
> to answer false is certainly counterintuitive, probably just wrong.
> Given that "LargeInteger zero" answers 0, which is not a kind of
> LargeInteger, it seems most reasonable for "Integer zero" to do the same.
>
>
> Regards,
>
> -Martin
>
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>

> The issue runs a bit deeper than that.  There are a number of
> assumptions and rules that should apply everywhere, but that for one
> reason or another have evolved into a mostly stable compromise that is a
> little bit inconsistent here and there.  We could try fixing the
> individual defects right away, but we would be just moving the
> inconsistency somewhere else.  What we need is a comprehensive effort to
> thoroughly eliminate inconsistency.  Because of compatibility and other
> related work we might want to do, the right time to fix this family of
> problems is 8.0.  To that effect, now we have AR 63498: "Fix issues with
> zero, unity and generality".
>
> On 7/13/2011 2:39 PM, Martin McClure wrote:
>> On 07/12/2011 02:16 AM, Steffen Märcker wrote:
>>> Hi,
>>>
>>> I've just observed that Integer zero/unity evaluates to
>>> LargePositiveInteger zero/unity. This leads surprisingly to
>>> Integer zero ~= 0.
>>> Two questions here:
>>> 1. Is it useful to get large integers here?
>>> 2. Do we actually want LargePositiveInteger zero ~= 0 (or unity ~= 1)?
>>
>> This definitely seems like a bug.
>>
>> For
>>
>>     0 = Integer zero
>>
>> to answer false is certainly counterintuitive, probably just wrong.
>> Given that "LargeInteger zero" answers 0, which is not a kind of
>> LargeInteger, it seems most reasonable for "Integer zero" to do the  
>> same.
>>
>>
>> Regards,
>>
>> -Martin
>>
>> _______________________________________________
>> vwnc mailing list
>> [hidden email]
>> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>>
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

> Thanks for your explanation! Are there some known issues that prevent
> fixing it in my images until the AR gets addressed? Btw, when can we
> expect VW 8? ^^
>
> Regards, Steffen
>
>
> Am 13.07.2011, 23:48 Uhr, schrieb Andres Valloud<[hidden email]>:
>
>> The issue runs a bit deeper than that.  There are a number of
>> assumptions and rules that should apply everywhere, but that for one
>> reason or another have evolved into a mostly stable compromise that is a
>> little bit inconsistent here and there.  We could try fixing the
>> individual defects right away, but we would be just moving the
>> inconsistency somewhere else.  What we need is a comprehensive effort to
>> thoroughly eliminate inconsistency.  Because of compatibility and other
>> related work we might want to do, the right time to fix this family of
>> problems is 8.0.  To that effect, now we have AR 63498: "Fix issues with
>> zero, unity and generality".
>>
>> On 7/13/2011 2:39 PM, Martin McClure wrote:
>>> On 07/12/2011 02:16 AM, Steffen Märcker wrote:
>>>> Hi,
>>>>
>>>> I've just observed that Integer zero/unity evaluates to
>>>> LargePositiveInteger zero/unity. This leads surprisingly to
>>>> Integer zero ~= 0.
>>>> Two questions here:
>>>> 1. Is it useful to get large integers here?
>>>> 2. Do we actually want LargePositiveInteger zero ~= 0 (or unity ~= 1)?
>>>
>>> This definitely seems like a bug.
>>>
>>> For
>>>
>>>      0 = Integer zero
>>>
>>> to answer false is certainly counterintuitive, probably just wrong.
>>> Given that "LargeInteger zero" answers 0, which is not a kind of
>>> LargeInteger, it seems most reasonable for "Integer zero" to do the
>>> same.
>>>
>>>
>>> Regards,
>>>
>>> -Martin
>>>
>>> _______________________________________________
>>> vwnc mailing list
>>> [hidden email]
>>> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>>>
>> _______________________________________________
>> vwnc mailing list
>> [hidden email]
>> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>