[Opengtoolkit-svn] SF.net SVN: opengtoolkit: [796] trunk/candidates/Average 1D Array

[<jk...@us...>]

Revision: 796
          http://opengtoolkit.svn.sourceforge.net/opengtoolkit/?rev=796&view=rev
Author:   jkring
Date:     2008-03-25 09:43:40 -0700 (Tue, 25 Mar 2008)

Log Message:
-----------
Added Meetings/2008-03-25 (Functional Spec Discussion)

Added Paths:
-----------
    trunk/candidates/Average 1D Array/Meetings/
    trunk/candidates/Average 1D Array/Meetings/2008-03-25 (Functional Spec Discussion)/
    trunk/candidates/Average 1D Array/Meetings/2008-03-25 (Functional Spec Discussion)/2008-03-25 (Functional Spec Discussion).txt

Added: trunk/candidates/Average 1D Array/Meetings/2008-03-25 (Functional Spec Discussion)/2008-03-25 (Functional Spec Discussion).txt
===================================================================
--- trunk/candidates/Average 1D Array/Meetings/2008-03-25 (Functional Spec Discussion)/2008-03-25 (Functional Spec Discussion).txt	                        (rev 0)
+++ trunk/candidates/Average 1D Array/Meetings/2008-03-25 (Functional Spec Discussion)/2008-03-25 (Functional Spec Discussion).txt	2008-03-25 16:43:40 UTC (rev 796)
@@ -0,0 +1,60 @@
+-= Decisions =-
+
+* VI belongs in the Numeric Palette/Library (as opposed to the Array Palette/Library)
+
+* Will use the "Numeric Primitive-like" Icon (triangle with X-bar)
+
+* Show polymorphic selector, by default, to make it clear which member is being used and that there are other poly options.  Users can easily hide this to save diagram realestate
+
+* 2D array in members should have icon with brackets around the X-bar to indicate that the output is an array of averages
+
+* Poly member labels and menus should use Title Case
+
+* Poly member labels should be "of Rows" and "of Columns"
+
+* 2D in - 1D out members should have an "arithmetic mean of rows" and "arithmetic mean of columns"
+
+* Use 3x3 icon to facilitate primitive-like icon/connector
+
+* VI Names: "Arithmetic Mean.vi" - "(2D Rows)", "(2D Columns)"
+
+* Code will be developed in 8.2.  We'll see how much push-back we get.
+
+
+* Drop NaN/Inf filtering because this is easy done.
+
+* No error handling: Output NaN when empty array in
+
+* Drop "rows and columns" (2D in, scalar out) since there is really no practical use case for it.  Also, there is no mathematical basis or experience for using it.
+
+* Drop integer, SGL, EXT types, so that we can use poly for selecting behavior.  There is not enough demand for these types.
+
+
+-= Other Notes =-
+
+* This is a perfect candidate for XNodes. However, since XNodes are not officially supported by NI, this is a bad path for OpenG to go down.  We want OpenG libraries to be rock-solid.
+
+* Users needing Fixed-point, EXT, 64-bit support users can be expected to have thier own analysis libraries.
+
+* Candidate page it should have:
+
+  ** indicator of what stage the candidate is in
+
+  ** history
+
+  ** information about who is working on it and what role they place
+
+* Numbered proposals are better than named proposals, because we might want to rename VIs (or the name might not even be know, at that point)
+
+
+-= Action Items =-
+
+1) Jim will send meeting notes to meeting participants to use for next steps (3/25)
+
+2) Justin should update the functional spec by early next week (EOD Monday, 3/31)
+
+3) Chris should update the candidate code (EOD Friday, 4/4)
+
+4) Justin will post to OpenG forum (Sat, 4/5)
+
+5) Jim will try to pull people into the discussion (Sat, 4/5)


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.