I am presently trying to determine what sort of answer I and my colleagues could give to the Sub-Committee's pressing question what to do about NERC/NSC.

  First, I'd like to emphasise again that my sources are adamant that one could not find out the answer to the questions of what exactly is wrong with the system and whether it could be fixed in three weeks. We are firmly of the opinion that Tony Collins is underestimating that entirely.

  Second, costing a pre-audit. I'm working on three people, air fares of GBP1500, 3+ weeks expenses at HMG's GBP100/day, and fees of GBP80/hour for 100 hours. Plus 20 per cent and I get to nearer GBP50K than your first guesstimate of GBP100K. John McDermid confirmed that kind of calculation, and suggested expect GBP50K but set aside GBP100K for contingencies (e.g., if it should turn out that one hones right in on a major problem fortuitously, it might be appropriate to roll the sleeves up and get right in there. One could do that if the financial flexibility were there, and it might save time overall).

  [Nancy Leveson charges up to USD5K per day, but said she would be prepared to consider a fraction of that, were she to be asked.]

  I would like to emphasise here, as I did to my colleagues, that I am not presuming anything about what the Sub-Committee will wish to do concerning an audit or indeed any other aspect of the NERC/NSC question. I am simply trying to help the Sub-Committee with a cost estimate of a pre-audit such as I suggested. Mr Stringer's point that there might nevertheless be scheduling constraints is still valid, of course.

  Concerning any involvement by myself in a pre-audit, it could be considered improprietous for me to profit form the Sub-Committee's response to my own advice. Were the Sub-Committee to request my involvement in a pre-audit (and I am not hereby supposing that they will), I should have seriously to consider performing it pro bono (expenses only, compensation if leave of absence were required). That would save HMG at most GBP8K, so the price of disinterest would be less significant for HMG. But nevertheless, I think disinterest is a requirement.

  Second, alternatives to auditing. The only two possibilities that I see are (a) carrying on as we are, hoping that sometime L-M will get it right; (b) pulling the plug. My last reply to Mrs Dunwoody stated what I think about option (a). Concerning (b), there must be a viable alternative. That would be either asking Hughes or asking Thomson (for I believe these are the only system integrators with experience on this scale) whether they could put together something for NERC/NSC. Professor Leveson consulted for the Auditor General of Canada on CAATS and other projects, and she has a very negative view of Hughes' accomplishments on CAATS, as well as of NAV Canada's assessment of those accomplishments. No one I asked has been briefed on, or has any direct experience with, Thomson systems.

  Thus one option is to simply ask Thomson, as well as Hughes and LM, what they could do for NERC/NCS in, say, 3 years. That is, to admit failure, let LM take the rap and let the contractor define the requirements. There should be significant contractual penalties for failure. LM will say: we'll finish our NERC system. Then you say: bid it as your new system then, and we'll evaluate the bid at your expense (i.e., full technical audit). One should evaluate the Hughes bid against the experience of CAATs. That should also be equivalent to a technical audit. What to do with Thomson? Depends what they bid. They are not yet known to have failed (with emphasis on the word "known"). Is it known why they did not bid on NERC or NSC?

  Again, this is technical unknown territory. No one has yet succeeded in building a system anything as ambitious as NERC, despite two attempts (AAS and CAATS). It seems therefore as if there's a good argument for junking it and asking contractors what they can prove they can do, against significant penalties for overbidding.

  Again, this is only a suggestion. There can be a thousand reasons why this isn't a good idea. But it might be worth at least discussing it with people.

Peter Ladkin

15 March 1998