Rewriting webspeed ....

Been looking / poking around the webspeed code , and don't like what I see. There are shared tables, shared streams, shared variables etc etc. yuk.

I've got it in mind to start from scratch, by writing basic classes and building from there.

Previously, I thought that there was some voodoo / chicken bone magic in the webspeed code, but there isn't. It's quite a simple thing.

You have a procedure that responds to a WEB-NOTIFY event, a WEB-CONTEXT handle with methods to get data, and output to "WEB" to actually send the data back. That's it, really. The rest is supporting code.

There seems to be loads of areas where optimization can be done. For example, when creating the field/value pairs, there is code to optimise to lookuplists instead of a temp-table if the number of fields < 100. Now, first of all, in my testing I've created 1000 fields, and the difference is 13ms vs 12ms. What's the point of all that code to handle something that will never make a difference ?

Anyone care to join in ? Or should I make a start and present the basic version when it is ready in order to kick-start the process ?

It may make for some interesting discussions