<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Le 12/04/2016 13:36, Samuel Gougeon a
écrit :<br>
</div>
<blockquote cite="mid:570CDDB2.5050805@free.fr" type="cite">
<meta http-equiv="content-type" content="text/html;
charset=windows-1252">
<font face="Arial">Hello,<br>
<br>
I am wondering about the following answers:<br>
<br>
</font><tt>--> s = %nan;</tt><tt><br>
</tt><tt>--> or(s)</tt><tt><br>
</tt><tt> ans =</tt><tt><br>
</tt><tt> F</tt><tt><br>
</tt><tt><br>
</tt><tt>--> and(s)</tt><tt><br>
</tt><tt> ans =</tt><tt><br>
</tt><tt> T</tt><tt><br>
</tt><font face="Arial"><br>
With this last answer, we have a case such that although all
entries are false -- otherwise</font><tt> or() </tt><font
face="Arial">would answer </font><tt>%T</tt><font face="Arial">
--, </font><tt>and()</tt><font face="Arial"> applied to them is
True.<br>
<br>
Is there a reason for that, or is it a bug?<br>
<br>
This behavior was already present in Scilab 4.1.2 and is still
here in Scilab 6.0-b1<br>
</font></blockquote>
<font face="Arial">.<br>
The title of my post was and(%nan)==%T, while the strange thing is
rather that or(%nan)==%F.<br>
Indeed, s being a real with unknown value,<br>
prob(s==0)/prob(s~=0) = 0, in such a way that or(%nan) should
rather be %T<br>
</font><br>
</body>
</html>