<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">Le 14/08/2017 à 14:50, Viktor
      Mileikovskyi a écrit :<br>
    </div>
    <blockquote
      cite="mid:1502712049.879715928.14p8z70q@frv42.fwdcdn.com"
      type="cite"><span class="xfm_42813660">
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">Dear Samuel!</span><br>
        </div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">Many thanks!<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">For me there is no problem to run SciLab as
            admin, use edit %s_i_s (i=1:4) and add<br>
          </span></div>
      </span></blockquote>
    <br>
    But it is a too specific status. Any user -- even not admin --
    should be able to extend the overload without modifying native
    Scilab's code, even without write access to Scilab installation
    directory.<br>
    <br>
    <blockquote
      cite="mid:1502712049.879715928.14p8z70q@frv42.fwdcdn.com"
      type="cite"><span class="xfm_42813660">
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">else<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">    TF=%f<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">or<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">else<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">    error("Comparison of complex numbers is
            undefined")<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">etc.<br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1"><br>
          </span></div>
        <div><span
            style="font-size:10pt;line-height:12pt;font-family:Arial;"
            class="xfmc1">The error is a good idea. It is more correct
            that in MatLab (comparing of the real part only). But %NAN
            is also possible, meaning that the operation can not provide
            any "logical number", 1 (%t) or 0 (%f).</span></div>
      </span></blockquote>
    <br>
    A boolean is expected as result. So we can't return %nan, that is of
    decimal type. In Scilab, unlike in Matlab (AFAIK), the boolean type
    is specific. <br>
    In addition, in boolean operations %nan does not propagate and is
    equivalent to %T:<br>
    <tt>--> [%f %t] & %nan</tt><tt><br>
    </tt><tt> ans  =</tt><tt><br>
    </tt><tt>  F T</tt><tt><br>
    </tt><tt><br>
    </tt><tt>--> [%f %t] | %nan</tt><tt><br>
    </tt><tt> ans  =</tt><tt><br>
    </tt><tt>  T T</tt><tt><br>
    </tt><br>
    This is why i will stick to issuing an error message. By the way,
    this is also the best way to make the user aware of the issue and
    informed on how overloading the comparison.<br>
    <br>
    Thanks for your input.<br>
    <br>
    Best regards<br>
    Samuel<br>
    <br>
  </body>
</html>