TK 104732 Profile formonly should always use a bestfit type (NO FIT not allow

huangyhg

Profile formonly should always use a bestfit type (NO FIT not allowed).  Default to VECTOR fit.
<< Don Turcotte  --  09/18/08  11:49:46>>
Profile formonly should always use a bestfit type (NO FIT not allowed).  Default to VECTOR fit.
<<END>>

<< Jared Hess (Development Notes)  --  04/07/09  14:33:48>>
This has been addedtoreadme for v44 (or ignored if internal item).
<<END>>
<< Changes made by Yanhua Huang -- 03/26/09  14:44:15>>
Action:  CMM Group to Yanhua Huang
<<END>>
<< Don Turcotte  --  09/25/08  13:00:55>>
I have modified this to allow bestfit == NO_FIT for formonly, but default to VECTOR (non-prismatic features) or LEAST SQUARES (prismatic features).   This applies only to FCFs.  I have not changed legacy profile.
Files inserted to server
------------------------
V44B\DIMENS\FEATCTRLFRM.CPP
V44B\DIMENS\FEATCTRLFRMUTIL.CPP
V44B\MENU\DIMENSIO.CPP
V44B\MENU\FEATCTRLFRMDLG.CPP
V44B\MENU\FEATCTRLFRMOUTPUTDLG.CPP
<<END>>
<< Changes made by Don Turcotte -- 09/25/08  13:01:08>>
Action:  Don Turcotte to CMM Group, Status:  CODING to REVIEW
<<END>>
<< Don Turcotte  --  09/24/08  17:18:42>>
I have modified this so we keep NO_FIT as an option since the user may want to do a bestfit alignment external to the profile formonly FCF.    I am still basing the default bestfit algorithm on LEAST_SQUARES if the feature is prismatic, otherwise (for scans, etc.) VECTOR_FIT.  I noticed in testing that this default is conflicting with the usage of global_prof_dlg_bestfit.
<<END>>
<< Don Turcotte  --  09/19/08  17:10:26>>
I have added a message dialog to alert the user when F9 on an existing FCF dialog and the bestfit type is changed from NO FIT.  Also fixed some other related issues here.  Everything seems to be working correctly now.
However, based on the latest e-mail exchange which I have included below, I'm not sure we should be removing the NO FIT choice at all.  Maybe the better change is to move the Bestfit algorithm selection and Workplane to the first page of the dialog so the user will be sure to see it.
e-mail exchange:
"The problem here is that “formandlocation” will report 2 X max deviation when the UseISOCalculations registry setting is true.  Lothar wants the report to show MAX-MIN because he is really looking for “formonly”.   The UseISOCalculations registry setting must be set true to report profile “formandlocation” according to ISO1101.

Donald Turcotte     Donald.Turcotte@HexagonMetrology.com
Software Developer - pc?dmis cmm group
Wilcox Associates, Inc.
250 Circuit Drive
North Kingstown, RI 02852 USA
Office: 1-401-886-2577



--------------------------------------------------------------------------------
From: HOSPADARUK Bob [mailto:Bob.Hospadaruk@hexagonmetrology.com]
Sent: Friday, September 19, 2008 3:28 PM
To: SCHAEFER Lothar
Cc: Frank Herr; CREMA Armin; Don Turcotte
Subject: RE: Profile evaluation - PR #257022

I *think* I follow you…  my point is that if you use the data in a best fit alignment then you do not need to fit in the profile dimension.  The legacy “forma and location” profile dimension will evaluate the data without fitting.  Therefore, we don not need a “no fit” option in the “form only” mode.
Bob Hospadaruk   
Technical Sales Specialist
Hexagon Metrology Inc.
51170 Grand River Ave.
Wixom, MI  48393-3327

phone: 248-613-2476
fax:      248-449-7438
email:   bob.hospadaruk@hexagonmetrology.com
web:     www.hexagonmetrology.com
_____________________________________________
From: SCHAEFER Lothar
Sent: Friday, September 19, 2008 10:56 AM
To: HOSPADARUK Bob
Cc: HERR Frank; CREMA Armin
Subject: AW: Profile evaluation - PR #257022
In my example I have already fitted using all the scans. I think the difference in this case is the formula for calculation, independent wether you do a bestfit first or not.
Form Only             MAX-MIN
Form and Location  2*MAX
I think it has to be possible to do bestfit for any profile independent from calculating the evaluation. So I use the bestfit in the alignment section independent from the profile calculation. First do the correct alignment, using bestfit, then calculate the profile.
So the difference is the profile calculation. Right?
"

Files inserted to server
------------------------
V44B\INCLUDE\EDITRSTR.H
V44B\INCLUDE\FEATCTRLFRMDLG.H
V44B\INCLUDE\FEATCTRLFRMOUTPUTDLG.H
V44B\INCLUDE\strings.h
V44B\MENU\FEATCTRLFRMDLG.CPP
V44B\MENU\FEATCTRLFRMOUTPUTDLG.CPP
V44B\RESOURCE\LANGS\ENGLISH\Strings.RC
V44B\SOURCE\EDITRSTR.CPP

<<END>>
<< Don Turcotte  --  09/18/08  17:15:41>>
I have implemented this for FCF's only.  If the user creates a new FCF profile the fit type will default to VECTOR_FIT unless the feature is a primatic type in which case the fit type defaults to LEAST_SQUARES.  The NO_FIT selection has been removed from the list.  
If the user loads an existing program with profile FCF's, these FCF's could have NO_FIT selected. This is not changed unless the user edits the FCF with F9.  Then the fit type is forced to VECTOR_FIT or LEAST_SQUARES as described above.
Files inserted to server
------------------------
V44B\DIMENS\FEATCTRLFRMUTIL.CPP
V44B\INCLUDE\FEATCTRLFRMUTIL.H
V44B\MENU\FEATCTRLFRMOUTPUTDLG.CPP
<<END>>
<< Changes made by Don Turcotte -- 09/18/08  17:15:41>>
Status:  DESIGN to CODING
<<END>>

huangyhg

BOOL IsPrismaticFeature(int nfeatures,CPCDcommand ** features) // 104732
{
  int types[] = {F_CIRCLE, F_SPHERE, F_LINE, F_CONE, F_CYLINDER, F_PLANE, F_SLOT, F_ELLIPSE, F_POLYGON, OFFLINE };
  int ntypes = 0;
  while( types[ntypes] != OFFLINE )
    ++ntypes;

  return IsTypeFeature(ntypes,types,nfeatures,features);
}