Date: 23 Nov 1999

   To:   LCB    
   From: Norman McCubbin (ATLAS Computing Co-ordinator)

   Concerning: OODBMS/RD45                          
 ------------------------------------------------------------------

  We (ATLAS) have been asked by the LCB Chairman to give our views on 
  future options for RD45 and OODBMS work. To be specific, we have 
  formulated the request in the form of two questions which we believe
  are faithful to the spirit of what was asked:

    1) What should RD45 still do w.r.t Objectivity?

    2) Given that there must be an element of doubt about Objectivity as 
       the long-term solution for LHC, what steps should be taken to deal 
       with this? 

  Before giving our answers to these questions, it is appropriate to set 
  out very briefly the present ATLAS position:
    - we believe that access to data (of all kinds) using OODBMS technology 
        could provide highly beneficial, and, in some cases, almost 
        essential functionality. 
        Our current choice is to use Objectivity as the underlying 
        OODBMS technology.
    - our present planning is that our use of Objectivity in areas such as
        simulation studies, test-beam data, detector description, etc
        will increase markedly during 2000;
    - from what is known today, we do not anticipate any significant problems 
        in beginning to use Objy in this way. It is, however, a requirement on 
        the emerging ATLAS software that the high-level applications be independent
        of the underlying persistency software. 
    - we are eager to participate actively and constructively in 
        future discussions in this area (see below).


   Turning now to the two questions, our views are the following: 

    1) RD45 has recently completed a status report (CERN/LHCC 99-28) which sets out: 

         - what has been achieved with respect to milestones; 
         - a summary of experience from CERN-based and other experiments; 
         - an assessment of the general "Objectivity situation", including a 
            risk-assessment;
         - and a proposal for future work.

        This is a very valuable report, and we agree with the general thrust for proposed 
        future work which includes  continued support for Objy to meet the ongoing needs 
        of the LHC experiments and further assessment of options for the future. 
        We would emphasise in particular the importance of the development of production 
        facilities and procedures, learning from the experience of the USA experiments 
        which are just starting to deal with real data.

        We address under 2) whether the future work is best achieved by a straightfoward 
        continuation of RD45.

   2) We believe that the "element of doubt" about Objy should now be addressed as follows:

          - carry out a careful  re-assessment of the needs of the LHC experiments for 
              all the various kinds of data (event data, calibration data etc.)           
              It is essential that that this re-assessment involve "users" from the LHC 
              and other experiments, as well as dBase experts;

          - assess possible courses of action, including continuing with Objy, the use of other
              commercial products, and the development of an HEP product, either new or based 
              on existing products; and make recommendations. 

     To elaborate a little on the first point, it seems plausible that requirements 
     may differ according to the kinds of 
     data alluded to above, and that one possible course of action might be 
     using Objectivity for some purposes and not for others.  An assessment 
     of the viability of coherent mixed-storage-solution strategies
     and the issues involved in providing such coherence, and recommendations 
     regarding such strategies, would be welcome.
     
     The above points are not in conflict with the future work proposed by RD45 itself,     
     but, to ensure the broadest possible input into this crucial assessment phase,
     we believe that careful consideration should be given as to what the organisational 
     structure should be: e.g. continue RD45, or some revised body.   
     It is of course important that any proposed organisational changes 
     should imply minimal disruption of ongoing work.