Wish List - Wish List: Wish #274

<Member picture

0 Like 1 Dislike

Steven Snyder

Reintegrate ranking with search

There is a ticket about this, but it’s kind of a long-range goal contingent on some changes to the ranking algorithm. So, I’m moving it here to serve as a reminder and to get some feedback in the meantime.

Comments (2)

  1. Gerhard Klimeck

    We really need to have a ranking formula first that is meaningful.

    Reply Report abuse

    Please login to comment.

  2. George B. Adams III

    Executive Summary: Delete Rankings from nanoHUB.org

    It is inherently difficult to represent anything meaningful of a complex situation using a single number metric.

    The most serious, professional effort to rank the performance of computer systems is the SPEC effort (www.spec.org). Who among us has based the buying decision for their next computer based on what machine/system software combination has the highest SPEC CPU2006 CINT2006 (2290) Peak result, or should that be the Base result, or should it be SPEC MPI2007 for MPI-parallel workloads, or something else? No one? Of course, no one SPEC “ranking” will make sense to guide your purchase.

    Why should we believe that there will be a Ranking for nanoHUB content items that is useful to the nanoHUB community?

    Because finding a ranking formula that is meaningful for more than a narrowly defined situation is so hard, I think we should put Ranking on the back burner of development as well as remove it from the site.

    For further justification for removing Ranking from view on nanoHUB, consider our Review and Rating data. As noisy as that data is (good Reviews and Ratings with obviously boneheaded/malicious ones mixed in), unlike with Ratings, we don’t hear complaints about Reviews or Ratings as flawed in and of themselves. Only individual reviews or stars assignments are cited as flawed.

    Reply Report abuse

    Please login to comment.