I left leantown at about 11-00am yesterday morning and drove down to St Anthony�s. James and I went down to the Oxford Modern Art Gallery for lunch. Tonight there is a Hallow-queen event at the college and so he had been out to the Oxfam shop to get something to wear drag-wise. Two Kazak girls in the shop had helped him chose something - some ironies here - Borat in reverse with James talking to them in Russian.

Apparently last year the CIA mounted a recruitment drive at the College - but the fellows drew the line at the idea of calling the annual ball �Spies Like Us�.James said that half his course are from outside the Uk - and how hard and effectively they work in English.

I sketched my idea of a thriller about a disgraced diplomat who drives an XJ out of sentimental attavhment to his former life - and he seemed to think there might be some life in it.

We went to the Estonian artwork on the ground-floor of the Gallery which was made up of interviews with Moscovite inhabitants with some connection with Estonia - conducted in Russian with subtitles which was just the thing given James� progress with the language. I mentioned the Slovakian piece

Then I drove along the A40 to Witley to have coffee with Laurence. Ben was home and also two thirds of Deguillo. The bass-player - Flora�s boyfriend - has just been given a Fender acoustic guitar with a metal resonator and he was kind enough to let me try it out. It made me think of Simon Prager�s Martin D28. Flora went off to the Modern Art Gallery where she works because tonight is the opening party for the Michel Burren work.

As Laurence and I talked about project appraisal, a set of ideas began to fall into place - the failure to do appraisal because of contractual considerations at the end of projects - well documented in How Buildings Learn - and the returns which might be available if this constraint is overcome. We both know of projects where the annual benefits may be in excess of �100m. To secure benefits on this scale something of the order of 10-20% of the benefits has to be invested. The question is whether lessons about how to secure this scale of return - say five to tenfold multiple between costs and benefits - can be extracted and applied to smaller interventions.

Smaller process-based improvement projects tend to run out of steam. The reason seems to be that the interface friction reduces the enthusiasm of the team doing the intervention work after while and so momentum decays. The reason the big projects work is that they link together interventions and so the interface problem is mitigated - also an investment on that scale inevitably brings a degree of top-down interest.

I drove from Witney to Charlbury climbing out of the Windrush valley as the sun set to the left and the full moon rose to the left. I missed the road I wanted - towards Banbury and headed for Chipping Norton as the landscape grew grander. Between CN and Moreton in the Marsh, where I picked up the Fosse Way I passed signs for the Rollright Stones. I have only ever been there once and while I was in the circle a fleet of helicopters flew directly overhead - it was an uncanny experience. I have felt there was something odd about this area ever since - bounded by the Fosse Way, the A40, the M40 and the river Cherwell. Mark Pavey lives in the middle of this zone - or at least he used to.