Is the difference here just various syntactical sugars or is there a reason to use one approach over the other? They all work, and to be a little more confusing what is the difference between this and evt.currentTarget?

the CSS

#reportDetails table tr:hover td,
#reportDetails table tr.hover td  {
    background: #aae4e2;
    color: #333333;
}

Sample html

<div id="reportDetails">
   <table>
     <tr>
       <td> something</td>
       <td> soemthing else</td>
     </tr>
     <tr>
        <td> something2</td>
        <td> soemthing else2</td>
     </tr>
   </table>
 </div>

dojo.behavior script

dojo.require("dojo.behavior");  
if (dojo.isIE <= 6) {
 dojo.behavior.add({
  '#reportDetails tr': {
   onmouseover: function(evt){ dojo.addClass(evt.currentTarget, "hover");},
   onmouseout: function(evt){dojo.removeClass(evt.currentTarget, "hover");
   }
  }
 });
}
dojo.behavior.apply();

dojo.query forEach script

    if (dojo.isIE <= 6) {
 dojo.addOnLoad(function() {
  dojo.query("tr", "reportDetails").forEach(function(node){ 
   node.onmouseover=function(){dojo.addClass(node,"hover");}
   node.onmouseout=function() {dojo.removeClass(node,"hover");}
   }
  });
 });
}

dojo.query ataching straight to the events

    if (dojo.isIE <= 6) {
 dojo.addOnLoad(function(){
  dojo.query("tr", "reportDetails")
  .onmouseover(function(evt){dojo.addClass(evt.currentTarget, "hover");})
  .onmouseout(function(evt){dojo.removeClass(evt.currentTarget, "hover");});
 });
}

I am assuming that evt.currentTarget and node could all be replaced with this and still work. I believe there is no real difference between 2 and 3 but the first one might actually use a different approach.