The Forge Reference Project

 

Topic: Citation analysis
Started by: Walt Freitag
Started on: 11/23/2002
Board: Site Discussion


On 11/23/2002 at 8:39pm, Walt Freitag wrote:
Citation analysis

It occurs to me that listing Forge threads in descending order of how many times each thread was linked to by another thread would make it easier to find some of the boilerplate threads that are often useful to refer new members to... because they're likely to have been so cited before.

What would be necessary would be a program that would search the threads for tags, parse the URLs for whether they link to a Forge thread, and compile a count for each thread linked to.

The basic idea is to leverage the work already performed by posters who've gone to the trouble of linking to prior threads in their posts. Ron probably knows a lot more about the utility and possible pitfalls of doing this. (For example, would having fewer people looking through old threads for references cause some old threads of merit, which just by chance haven't been linked to before, to be prematurely forgotten?)

- Walt

Message 4350#42846

Previous & subsequent topics...
...started by Walt Freitag
...in which Walt Freitag participated
...in Site Discussion
...including keyword:

 (leave blank for none)
...from around 11/23/2002




On 11/24/2002 at 1:16am, Ron Edwards wrote:
RE: Citation analysis

Hi Walt,

This sort of thing is Clinton's bailiwick, actually. The good news: it happens that an algorithm something like what you're describing is being developed already, for a kind of thread-based glossary thing deal. Just how real this is at the moment, and how long it will be before it goes live, isn't something I know, but I do know it's happening.

Best,
Ron

Message 4350#42852

Previous & subsequent topics...
...started by Ron Edwards
...in which Ron Edwards participated
...in Site Discussion
...including keyword:

 (leave blank for none)
...from around 11/24/2002




On 11/24/2002 at 6:34am, Clinton R. Nixon wrote:
RE: Citation analysis

Ron Edwards wrote:
This sort of thing is Clinton's bailiwick, actually. The good news: it happens that an algorithm something like what you're describing is being developed already, for a kind of thread-based glossary thing deal. Just how real this is at the moment, and how long it will be before it goes live, isn't something I know, but I do know it's happening.


Right now, it's still in the planning stages. It's a project that's at the limit of my technical skills, to be honest, but I do plan to implement it - as part of a new unifed Forge system - in 3 to 6 months.

Message 4350#42859

Previous & subsequent topics...
...started by Clinton R. Nixon
...in which Clinton R. Nixon participated
...in Site Discussion
...including keyword:

 (leave blank for none)
...from around 11/24/2002




On 11/24/2002 at 5:03pm, Walt Freitag wrote:
RE: Citation analysis

That sounds great. I was thinking of it as a quick and dirty solution, but having it as part of a real indexing system would be even better.

My reference to Ron was in his capacity as a scientist. Citation statistics for papers have become an increasingly prevalent tool in research.

- Walt

Message 4350#42875

Previous & subsequent topics...
...started by Walt Freitag
...in which Walt Freitag participated
...in Site Discussion
...including keyword:

 (leave blank for none)
...from around 11/24/2002