Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Selection Criteria for choosing a Discovery Method



All,
 
Thanks to Mike McCormack for committing to develop a matrix of tests that the two finalist "Discovery Methods" should undergo by the end of this month.  Based on the test results already presented, and the assessments of the experts at the interim meeting, it is likely that both methods will pass all the tests, assuming the tests are completed.  That being the case, we should establish an evaluation matrix to assist us in the selection of the best method.  I would like to propose that we attempt to list under criteria, not specific capabilities, but rather the benefits that are derived from those capabilities.  For example, elimination of a requirement for discrete capacitors or transformers might be listed under ability to cost reduce.  High input impedance might be listed under fail-safe protection from damaging a DTE. 
 
I would like to ask the proponents to send to the reflector their candidate criteria for the matrix.  Separately, they should be sending the characteristics of their preferred method that help in achieving a high score for the criteria listed.  Separately, there may be characteristics of your least preferred method that you believe will detract from its capability to achieve a high score for a particular criteria.  These characteristics should also be sent.
 
I volunteer to keep the matrix up to date by periodically (no less often than every two weeks) updating the matrix based on reflector input.  I will do this in as neutral a manner as I know how.  Based on the input I receive on the reflector, I commit to having the first "criteria selection matrix" posted on the reflector no later than Friday, September 29th.  Steve Carlson, I ask that we have this selection criteria matrix posted on our web site, along with the test matrix.
 
I would recommend that we do NOT attempt to quantify the importance of each of the criteria, or provide the weighting for each benefit and detractor.  The reason for this recommendation is that those that have a preferred criteria will have a tendancy to heaviliy weight those criteria where their method excels.  Instead, I ask that each of us try to develop our own quantification so that we can determine which method is superior.  Certainly, reflector discussion on the importance of the various criteria, and the specific characteristics listed, is in order.  We should have as an objective to come to the November plenary with our weightings already thought outl.  Then we should allow each person 1 opportunity to speak for a very limited time, and then vote.  I strongly encourage those with a strong interest in this topic to make their strongest arguements in favor of their approach on the reflector, so that we don't need to provide extensive time at the meeting to review this topic in detail.
 
Let me suggest a format for the matrix as follows:
 
  Method 1 Method 2
Criteria Pro Con Pro Con
 
Notes: 
 
 
Listed under Pro or Con would be specific attributes of the Methods.  Listed under notes would be no more than one sentence pro and one sentence con per attribute.  It would be an excersize for the reader to distill this data into numbers.  The first set of numbers would be for the weighting of the criteria.  The second set would be for the weighting of each of the Pros and Cons. 
 
Establishing this matrix should assist us in moving our work forward.  However, it is not meant to stifle further introduction of good ways to merge the benefits of these two proposals into something better than either of them.  I would strongly advice, however, that at this point we should not consider any substantially different proposals that introduce a new method, rather than merging features of these methods together, or improving these methods.
 
 
Best regards,
 
Robert D. Love
President, LAN Connect Consultants
7105 Leveret Circle
Raleigh, NC 27615
Phone: 919 848-6773
Fax: 720 222-0900
email: rdlove@xxxxxxxx