Discussion Notes

Feb 26, 2001

Graph Structure in the Web

• Can be viewed as modeling a second-level of structure (bow-tie), beyond power-laws. This leads to a substantially different calculation of the diameter of the web.
• Also addresses an `undirected' way of analyzing a directed graph
• Classification of web pages into the four categories as suggested by this paper might be helpful in the design of better web crawling strategies.

Searching the World Wide Web

• A very careful choice of constraints (on queries), to ensure that results are significant and thus, can be interpreted to mean differences in coverage
• Non-parametric estimation (model-free)
• It will be interesting if such a study can reveal how best to create a meta-search engine (to ensure combining of coverages)