I just had a chat with a PHD-er from the psychology faculty of the Uni of Groningen Holland. He was interested in using Swiki to publish and organize their research projects on the net.
Cool
I showed him how to use and set up a swiki. After the demo he saw for their situation some limitations and gave some interesting suggestions: a. What if you could put in xml like tags so we can have a better search engine. Tags like '<author>' '<article><title>' would make it possible to search more effectively.
The Swiki search engine is pretty bad. I don't think anyone is working on making a better one though.
b. What if the edit-page has buttons to have an texteditor in a wysiwyg style. Psychology students are a bit afraid of bare html editors. I have seen this wysiwyg button option in a custom swiki during the demo of NetUnify at OOPSLA'00.
It's not bare HTML. Swiki has its own mark-up. You can think of it more like e-mail if you want. I'm not a big fan of wysiwyg, especially on the net, so I am not going to be creating this.
There are already so many swikis in use within universities, did one of them already implement these functionalities?
Nope.
Unfortunately, I'm only able to forward these suggestions and not to implement them, cause of priority list limitations.
Thanks for the suggestions though. You may want to join the PWS list, where we discuss Swikis. To subscribe, send a message to majordomo@cc.gatech.edu. In the body, put "subscribe pws".
Peace and Luck!
Je77
"Jochen F. Rick" wrote:
I showed him how to use and set up a swiki. After the demo he saw for their situation some limitations and gave some interesting suggestions: a. What if you could put in xml like tags so we can have a better search engine. Tags like '<author>' '<article><title>' would make it possible to search more effectively.
The Swiki search engine is pretty bad. I don't think anyone is working on making a better one though.
You might take a look at the search engine in SCAN. It uses a fulltext index and is able to search for text in specific fields of items. Items in SCAN are versioned, and if a word occured in an earlier version of an item but not in the current, that earlier version is presented. For searching, it's pretty fast. The indexing that's necessary when items are added is a bit slower, but still acceptable. One day I'll make it fast enough that indexing the whole contents of source files becomes feasible (which currently isn't...)
Look at http://squeak.heeg.de:8080/search and the 'Searching' help page accessible from there.
The whole SCAN source is downloadable from SCAN, too. Building a Swiki on top of the SCAN primitive structure shouldn't be too difficult, but it's not on my agenda.
Cheers, Hans-Martin
squeak-dev@lists.squeakfoundation.org