Walter,
I will put my comments in red after each of the numbers you gave us as examples
based on what I learnt from reading the spec. I added one more on the end, just
for fun…
Arpad
=====================================================================
From:
ibis-editorial-bounce@xxxxxxxxxxxxx<mailto:ibis-editorial-bounce@xxxxxxxxxxxxx>
[mailto:ibis-editorial-bounce@xxxxxxxxxxxxx] On Behalf Of Walter Katz
Sent: Tuesday, April 27, 2021 5:44 PM
To: michael.mirmak@xxxxxxxxx<mailto:michael.mirmak@xxxxxxxxx>;
ibis-editorial@xxxxxxxxxxxxx<mailto:ibis-editorial@xxxxxxxxxxxxx>
Subject: [ibis-editorial] Re: New agenda topic for this week
All,
I do not think that there is question on what the parser can do. UI is
equivalent to Float, except to convert to seconds one must multiply it by the
bit_time (actually symbol_time). Since UI is treat like Float (except for the
conversion to seconds) and for Float “1” is a legal representation of “1.0”,
then the parser logic should be to convert every value in the list (and
include) Default to a floating point number. The comparisons should always be
done with the floating point number. Thus the following will all be identical
floating point numbers:
1 This is an Integer or Float according to the spec. You pick.
1. This is undefined. The spec talks about “fractional or decimal
component”, but there is only a decimal point here without a “component”
1.0 This is a Float
1.00 This is still a Float
10e-1 This is an Integer
1e0 This is still an Integer
.1e1 This is still a Float because it contains a “fractional or decimal
component”, even though it evaluates to an integer, and regardless that it does
not have an integer part before the decimal point
1e-1 This would be a nonsense Integer, because it does not have a
factional or decimal component, but evaluates to a non-integer
Does the parser give a warning or error on any of the case described in the
issue?
Walter
Walter Katz
Work 508.647-7633
Cell 720.417-3762
[Description: Description: Visit MathWorks.com]