<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">2017-09-14 21:37 GMT+02:00 Kristian Evers <span dir="ltr"><<a href="mailto:kreve@sdfe.dk" target="_blank">kreve@sdfe.dk</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div lang="DA" link="#0563C1" vlink="#954F72">
<div class="m_-6227804635437603427WordSection1">
<p style="margin:0cm;margin-bottom:.0001pt"><span style="color:rgb(31,73,125);font-family:Calibri,sans-serif;font-size:11pt"> </span><br></p><p class="MsoNormal"><span lang="EN-US" style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Yes. It is not a path I want to go down, unless absolutely forced to. But it *<b>could</b>* solve the problem.<u></u><u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u></span></p></div></div></blockquote><div><br></div><div>Actually that was one of the cases anticipated by adding the "observation" registers in the PJ_OBS data type: You could have one step in the pipeline filling in stuff in the registers, passing them on to the next step, which would combine them with the coordinate stuff in whichever way it wants to.</div><div><br></div><div>This was also one of the reasons to introduce the switch that makes it possible to mute steps in either forward or backward execution of the pipeline (you want parameters filled in before the step using them, so need to mirror the operation in cases of backward execution).</div><div><br></div><div>So no pipeline-parenthesis needed - just put some slightly over-engineered elements from the early design phase into practical use :-)</div><div><br></div><div>That being said: I would certainly prefer one of the other solutions!</div><div><br></div><div>/Thomas</div></div></div></div>