<html>
<head>
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Paul,<br>
<br>
There are no new objects being introduced into the analysis. I
don't understand your point at all. There will only be existing
objects frozen and then thawed out again:<br>
<ul>
<li>DMCReaction</li>
<li>DMCThrown</li>
<li>DTagger</li>
<li>DNeutralShower (subclass of DKinematicData)<br>
</li>
<li>DChargedParticleHypothesis (subclass of DKinematicData)</li>
</ul>
<p>Can you explain?<br>
</p>
<p>-Richard Jones<br>
</p>
<br>
<br>
<br>
On 7/2/2012 11:54 AM, Paul Mattione wrote:<br>
</div>
<blockquote cite="mid:EDE82FA8-0D27-4740-B0FE-84F299606F41@jlab.org" type="cite">
<pre wrap="">Hey Richard, how do you envision this REST format being used?
For example, right now everyone is writing analysis tools to look at data using the current reconstruction objects (e.g. DKinematicData). When we convert to using REST, will we have to completely change the internals of these tools to operate on these new objects instead? But wouldn't it be useful to be able to still do analysis on the original DKinematicData objects (e.g. generating monitoring histograms during track reconstruction to provide data quality monitoring/diagnostics)? You'd hate to have to maintain a parallel set of analysis tools.
Would the end user instead build DKinematicData objects out of the REST data and then proceed with the current tools?
- Paul
</pre>
</blockquote>
<br>
<br>
</body>
</html>