Re: Include language bindings? Re: Draft Standard Text, V02.1
Friends,
>
> Looking at IEEE 754-1985, one sees that it was a decade or more
> after the hardware standard was in use before there were language
> bindings. Can including language bindings help the process?
Yes (technically), and No (politically).
If we consider the current debate in the US over healthcare (and in YOUR country, whatever is your current big political football), we see two opposing political forces at play (or war, if there is any difference).
The proponents put in something for everyone, hoping to gain wider support. That tends to make bills comprehensive and loaded with things that have little correlation with the original intent of the bill.
There are some questions (abortion, for one) that cause people to vote "NO," no matter how much they like the rest of the bill. That tends to keep bills small to avoid upsetting anyone.
Once in a while, a truly comprehensive bill is passed, but most legislation takes small, incremental steps.
In P1788, the more (good stuff) we put in, the more useful to implementors it might be.
The more we put in, the greater are the chances of people seeing things they cannot support and voting "NO," no matter how much they like the rest of the bill. Oh, and the longer it takes to put it all together.
I would VERY MUCH like to see language bindings. I think my original informal proposal for a Basic Interval Arithmetic Subroutines library (BIAS) included rudimentary bindings for three languages. However, if I had to choose, I would rather see a hardware standard. Soon.
John, with the rest of us offering advice, is making MAJOR progress toward a hardware standard. I continue to hold that the best strategy moving toward getting T.C. MITS (The Celebrated Man In The Street -- S.I. Hyakowa) using intervals is as SIMPLE a hardware standard as will do the job. Then we can move on to enriched environments, language standards, etc.
George
>
> Baker
>
> On 3/15/2010 9:49 AM, Rudnei Cunha wrote:
>> Dear colleagues,
>>
>> I understand P1788 is a hardware standard and as such George's remarks
>> are correct.
>>
>> However, shouldn't we also be concerned with how these operations will
>> map onto high-level language constructs?
>>
>> I mean, should we have at the end of the work on the standard a proposal
>> for language bindings? Or should we leave it as a work to be carried out
>> immediately after the approval of the standard?