Some preliminary programming considerations


1. Time to go up/awake: could be either time due triggered or by keywords' stack limitations.

2. The k's core to append to the search are defined in our Thesaurus database. 

3. The first query is done for the couple (k, s)

4. The outcome of this first query should be weighted by popularity metric or related. This will be one of the crucial ranking variables. 

5. We must proceed by  (pair, k1), (pair, k2), .... , (pair, kn), being k1, k2, ..... , kn successive core keywords. For each query we will have a rank, of course lower than with the first couple. 

6. Any search must be filtered on the spot once query is delivered. This applies from couple (k, s) to couple (k, s) with core keywords appended. 

7. Initially until we get enough experience, we should take into considerations two alternatives, to be chosen parametrically by the CE. One choosing the lowest popularity, and the other proceed to merge before and in both cases keep Top x, being x another parameter to be fixed by the CE. 

8. If the convergence is not good, CE could decide to proceed further (steps), namely: couple, k1, k2; couple, k1, k3, etc etc. 

9. Always, controlled from desktop, procurebot should deliver a process summary and Top x for each couple besides.. 

10. CE is enabled to add keywords to filters/masks at will.

11. Desktop Stepwise agents' operation and tune up must be enabled.