Watch, Follow, &
Connect with Us

QualityCentral

Public Report
Report From: QualityCentral/Client Interface/Rating/Voting    [ Add a report in this area ]  
Report #:  1507   Status: Closed
Alternative #1:   Floating point votes + voting against
Project:  QualityCentral Build #:  1.0.0.78
Version:    1.0 Submitted By:   Robert Lee
Report Type:  Suggestion / Enhancement Request Date Reported:  6/13/2002 5:25:24 PM
Severity:    Serious / Highly visible problem Last Updated: 9/3/2003 11:08:56 AM
Platform:    All versions Internal Tracking #:  
Resolution: Won't Do (Resolution Comments) Resolved in Build: : None
Duplicate of:  None
Voting and Rating
Overall Rating: (4 Total Ratings)
1.00 out of 5
Total Votes: 2
Description
(Choose among the alternatives by VOTING for your preference)

part 1:  Make votes floating point with a total of 15 per project and a cap of 5 on any one report.  Additionally add the ability to case opposing votes (as negative) with a cap of 1.0 on any one report.  Note that favorable votes and opposing votes both count against (i.e. reduce your total).

Part 2:  change "rating" to "Quality of Report"


Note: current votes can be converted simply by multipling by 5 then applying the cap.  Add a one time use feature to "unrate all" so that those that have been using something other than Quality of Report as their metric can clear them all and start over.
Steps to Reproduce:
None
Workarounds
None
Attachment
None
Comments

Mike Orriss at 6/13/2002 9:54:57 PM -
There isn't a negative vote available to me, so I will have to make do with a comment.

I'm opposed to this suggestion for several reasons:

1) Results obtained will imply a greater level of accuracy than is possible. A bit like saying your speed is 30.45126 miles per hour

2) It adds further complexity to something I already consider to be somewhat over-engineered

3) The current system serves its purpose of providing away of alerting sysops to those reports worth transferring to Borland's internal tracking systems

4) there are other changes outstanding which would improve the system. I would hate to see time spent on this in preference

Robert Lee at 6/14/2002 5:52:59 AM -
The primary problem isn't over-engineering but rather under-specification.  Basically, everybody has been filling out "rating" based on their own rules.  This appears to be true even among TeamB and possibly even Borlanders.   I think that Deborah made the excellent observation that prior to the unlabeled 1-5 scale the ratings all "sounded" like they were about importance, not whether the report was accurate/reproduceable etc.  

As for your comment about illusion of accuracy.  I interpret it to mean that floating point is too finely grained.    Well yes of course it is.  However, it also offers a certain amount of built in scaling and there is no real downside.  

>3) The current system serves its purpose of providing away of alerting sysops to those >reports worth transferring to Borland's internal tracking systems

This is not true.  At least this is not the stated purpose of the project as outlined in John K's online documentation (http://community.borland.com/article/0,1410,27815,00.html)  The idea that the system will track the "importance" that users associate with reports is repeatedly stated.

Deborah Pate at 6/14/2002 12:54:09 AM -
There is only one thing that really matters about the quality of a report, and that it whether it allows an engineer to reproduce the problem or see exactly what feature is required. It is IMO quite absurd to have a five-point rating scheme for this.

It might also deter some people from submitting reports, which would be unfortunate.

Robert Lee at 6/14/2002 5:25:57 AM -
I don't really disagree with you about the 5 point rating.  The reason that I put it in this way is that those who have been using rating as report Quality would be able to carry over what they had already rated.  Secondly,  a 5-1 system handles differences of opinion better than radio buttons (i.e. you can average in a meaningful way).  Really "Quality of report" is too brief.  Something like  "This report contains all the information necessary?   Agree=5 Disagree=1"  But that is kind of wordy.

Kristofer Skaug at 6/24/2002 4:49:44 AM -
"Note that favorable votes and opposing votes both count against (i.e. reduce your total)."

The opposite would've been very interesting! <g> (gain infinitely many pro-votes by negatively voting on others)

Server Response from: ETNACODE01

Copyright© 1994 - 2013 Embarcadero Technologies, Inc. All rights reserved.