|Theoretical Discussion On Linking|
| 1:35 am on May 4, 2003 (gmt 0)|
|discussion of how principles from research papers can be applied to the practical realities…how linking practices inter-relate with the various elements of on-page site optimization – Marcia - from Pumping it up! Let's Regroup and Talk [webmasterworld.com] |
|researching resources and posting them here. Like current papers from Colleges or private organizations. Discussing their applications / implications in regards to linking, specifically. - robertito62 |
|round up of some of the big papers discussing links/link text - brotherhood of LAN |
Start adding them here folks, when you find them and want to discuss them. We can also take these into their own discussion but I would like to at least get those discussions and references to papers listed here so we have a single place to refer people when they ask, ok?
| 8:02 pm on May 4, 2003 (gmt 0)|
Ok, I found this one, which I think is very interesting and wonder what others think.
Graph structure in the web [almaden.ibm.com]
Andrei Broder1, Ravi Kumar2, Farzin Maghoul1, Prabhakar Raghavan2, Sridhar Rajagopalan2, Raymie Stata3, Andrew Tomkins2, Janet Wiener3
1: AltaVista Company, San Mateo, CA.
2: IBM Almaden Research Center, San Jose, CA.
3: Compaq Systems Research Center, Palo Alto, CA.
|brotherhood of LAN|
| 8:43 pm on May 4, 2003 (gmt 0)|
I found this one the other day when reading about the "connectivity server", and was wondering if we had discussed it at WW before...
Good set of papers abouyt links included in that link.
>interesting and what others think
|Acknowledgment. We thank Keith Randall for his insights into our SCC algorithm and implementation. |
Searched for "links SCC algorithm" and got a related paper, which also mentioned dthe "connectivity server" in the first paragraph. Looks like we are in the right neighbourhood here :)......
|The Connectivity Server is a special-purpose database whose schema models the Web as a graph |
where URLs are nodes and hyperlinks are directed edges. The Link Database is the portion of the
Connectivity Server that provides fast access to the hyperlink data. It allows us to run arbitrary
graph algorithms over the Web graph in real time, usually by writing textbook implementations
of the algorithms. Applications of the Link Database include PageRank [BP98], which ranks the
importance of web pages based on their inlinks; Kleinberg’s HITS [K98], which refines web
query results by examining local subgraphs of the web; mirror detection, which uses links and
URLs to find web sites that mirror each other’s contents [BBDH99]; and studies of web structure
[BKM00]. While the Link Database is critical for each of these applications, we know of no
other database that can store as many links in as little space
Krishna Bharata, Andrei Brodera, Monika Henzingera,
Puneet Kumara, and Suresh Venkatasubramanianb
I think it's safe to say that today's hardware is capable of accommodating the growth of the web, so volume will never be a problem and that they'll be looking for more precision in these papers.
22 minutes vs. one month for google to calculate PR and update its index, understandable why they are looking for a patent to update SERP's "on the fly" (a recent WW thread mentioned this patent)
|how linking practices inter-relate with the various elements of on-page site optimization – Marcia |
Just as a general statement, from looking at these papers, the linking structure of your "theme" competitors would play as an important role as on page optimization. No doubt there are papers dedicated to both.
| 12:02 am on May 10, 2003 (gmt 0)|
Visit the Recent publications for a bit of heavy reading. Anyone interested in any of this?
| 1:38 pm on May 13, 2003 (gmt 0)|
I am interested but time is the issue - I can't read that fast! ;)