- From: David W. Morris <dwm@shell.portal.com>
- Date: Mon, 8 Jan 1996 14:35:14 -0800 (PST)
- To: HTTP Caching Subgroup <http-caching@pa.dec.com>
On more reflection, I just realized what that what we are discussing here is classic inventory managment. I haven't tracked progress in the field since I wrote such applications 30 years ago but I assume there has been some. It seems to me that the only protocol issue would be to identify data which would be useful to modern inventory demand forecasting algorithms, design a way to exchange such data if required and we're done. In the olden days, forecasts were based on the sales experience of the seller. Initial orders might be as the result of a good sales job by the vendor but even without scientific inventory management, it seem likely that a retailer would retain or acquire additional inventory based on manufacturer reports that on the other side of the world, this product is well received. So lets agree that some very smart algorithms may be available to make caching more effective. The first order solution is certainly self contained within an individual cache. Perhaps we can thus agree do defer discussion until things reasonable for 1.1 have been dealt with. Dave Morris
Received on Monday, 8 January 1996 22:48:29 UTC