|
PureBytes Links
Trading Reference Links
|
Seems to me the TS flaw is in INTportion, etc. that are not designed to take
into account #s as they exist. Would not be hard to incorporate a rounding
element that would make it go away.
FWIW, TS asked me to provide code to show them what was going on. :-)
Chris
----- Original Message -----
From: "Alex Matulich" <alex@xxxxxxxxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Monday, June 07, 2004 4:14 PM
Subject: Re: Re[2]: TS7 EL Bug?
> >Then again, in hindsight, TS made a wise decision to thumb their
> >nose at [double precision], betting that those who wined about it
> >would make their own workarounds anyway, like linking with excel,
> >etc. In turn, TS would have software that is faster, albeit, not
> >quite accurate.
>
> ...which wasn't even a good excuse back in the late 1980s which
> I understand was the TS4 era. Most computers then had math
> coprocessors which were optimized for double precision math. In
> some cases single precision would actually slow the operations
> down because everything would get converted internally to double
> precision.
>
> The only possible excuse I can think of would have to do with the
> fact that storage capacity was much smaller then (720K diskettes,
> and a 30 megabyte hard disk was considered large). But it's still
> not an excuse for TS because TS seems to store all its data in
> integer format.
>
> -A
>
|