COSMIC Sizing Forums Benchmarking Why is COSMIC only 10% of the ISBSG dataset

Viewing 2 reply threads
  • Author
    Posts
    • #3172
      Frank Vogelezang
      Keymaster

      In his article Sources of Software Benchmarks, Capers Jones gives the following summary on COSMIC:

      The COSMIC Functional Size Measurement method is the only such method developed by an international group of software metrics experts, designed to conform to fundamental software engineering principles, and to be applicable to business application, real-time and infrastructure software. It can be used in any layer of a software architecture and at any level of decomposition. Use of the COSMIC method for performance measurement, benchmarking and estimating is therefore expected to enable much higher degrees of precision than when using older software size measures.

      If we are that good, why are we – for years – still at around 10% of the ISBSG dataset? How can we improve that number and get more data into the ISBSG database.

    • #3580

      I think more promotion is needed, success stories, applications.
      Then, I believe that automation of the counting rules is needed for a more wide use.

      It would be very good if COSMIC could standardize such automatic counting (see what CISQ is doing with AFP).

      br
      Karl-Erik

      • #3634
        Alok Sharma
        Participant

        My 2 cents.
        Most of the data in ISBSG is based on FPA (unadjusted), which as good as Cfp for a given business functionality sizing. So for all practical purposes, we have the data.

        Having said this, the rate of project data submission using Cfp is low and that of FPA is declining. So we should encourage the data submission though more publicity, i agree.

        regards
        Alok

    • #4105
      Shashirekha Kandi
      Participant

      As Karl mentioned more promotion and success stories are needed.

      I think organizations are not submitting data to ISBSG due to confidentiality agreement with clients and too many levels of approvals within the organization. As Alok mentioned we should encourage data submission more publicly.

      Thanks,
      Shashirekha

Viewing 2 reply threads
  • You must be logged in to reply to this topic.