scientific / exponents in float fields

Imported from previous forum

Hi all,

whats the standard practice with regards to allowing scientific number representations in FIX messages ?

My old in-house FIX engine could encode/decode them, however I had an issue today, using QuickFIX, where it could not decode and threw an exception. (I’ve done a patch to my QF code and also raised the issue with the QF forum).

But I thought I ask a general question here to see what the consensus of other FIX users/vendors is.

The FIX (4.2) spec says a float field is ‘Sequence of digits with optional decimal point and sign character (ASCII characters “-”, “0” - “9” and “.”)’

Is a field such as ‘44=+1.234000e-02’ bad and to be avoided ?

(its normal with very large number i see the issue)

If both counter parties can encode/decode then is there any issue ?

(For the record, QF is C++, and I’m using debian/Linux, in case its a C language or locale issue).

regards
Nick