Forum Moderators: open
from the WWW2003 Conference Proceedings as posted here [webmasterworld.com] by Msgraph.
mainly analysed from IBM's Intranet.
They put forward a number of interesting Axioms (some seem obvious).
Also:
The indegree and outdegree distributions for the intranet are remarkably similar to those reported for the Internet [8,14]. The connectivity properties, however, are significantly different.....On the intranet, a significantly larger fraction of pages have high values of PageRank (probability mass in the distribution alluded to), and a significantly smaller fraction of pages have mid-range values of PageRank. This suggests that PageRank might not be an effective discriminator among intranet pages--it can clearly distinguish between the very important pages and the others, but probably does not offer finer distinctions among the latter.
Could also be interesting for people considering to buy the Google Search Appliance [google.com] and wondering why things do not completely add up?
Thus, the search problem on the Internet is, in some sense, easier than on intranets--any reasonable subset of pages on the topic would be considered good answers for the tree sap query, but unless an intranet search engine hones in on the exact page for the queries of the kind mentioned, it would be considered poor.
Ain't that the truth. I haven't had many chances to search on strictly inhouse intranets, but judging the results on many major extranets I think there are loads of similarities. Performing vague searches can produce hundreds if not thousands of results that are extremely impossible to sort through. Only if you know a specific term(s) directly related to what you are looking for can you find what you need in the top 20. But I guess part of this leads to axiom #4.
I think intranet and extranet developers are at a good spot now in technology but there really isn't much else they can do unless those who actually design the pages, design them right. Sure they can go by a sort of PageRank or URL analysis but that can only go so far until site developer enter in more data to categorize the pages better.
Like Axiom #3 states: intranets are spam free. I think this allows for good use of meta information where a more detailed structured search can easily enter the picture. Again the problem is most internal site developers aren't experienced in the art of page optimization.
