Improve Classification of Codev Requests, Brainstorming, Bugs, etc.

The Codev Web has several SEARCH pages which assist with management of Requests for Enhancement, for example. I think it would assist the management of all these requests if the WebForm included a checkbox formfield which would allow the submitter to further categories the subject of the request/bug/idea. The checkbox list would probably include each of the modules in the TWiki code base, for example, and perhaps any embedded functionality like INCLUDE.

Here is a starter list just to evoke discussion.

and so forth. This list is by no means complete or even correct.

The idea is that when a participant submits a request, idea, bug, one or more of the checkboxes would be checked to indicate which module or topic is related. Note that I am stressing that these be related to a module or specific function that is (hopefully) somewhat functionally distinct from other functions.

Then, each on of the Names above would be a page with a SEARCH providing all the pages with the relavant information, pulling together all the comments into one place. On that page, someone may want to be somewhat of a moderator to summarize the situation with the discussion on those areas. A person making bug fixes or enhancements on that function could review all the information gathered there in order to quickly fix or enhance that code.

Currently, there are over 250 ideas in the Requests section alone, and organizing them by Revision Date does not assist with an understanding of the recent discussions. Collecting by the affected code or general topic should help the understanding and completion of these changes.

After sleeping on my idea above, I think that just creating summary pages which search on the content is OK, and perahps more powerful way to do it, although having the meta tag may assist with categorization. The problem of course is the accuracy of the tag may be wrong, given that we are trusting submitters to provide it. It is probably OK to consider this idea, but I am not going to push for it if we have the other summary pages, and indeed I may just be ignorant of those pages.

-- RaymondLutz - 12 Nov 2003

It turns out that a major feature of many KM platforms is automatic classification of content. I think TWiki could easily provide this, by leveraging engines like that under http://Popfile.sf.net

-- MartinCleaver - 12 Nov 2003

Yes, I see what you mean, although I'm not sure if popfile is the best starting point. It would be nice if we had a means to fully automate the organization of the content, probably by fully indexing all the words in every topic, and then performing some statistical analysis to provide the topics that are most highly related based on content and grouping those together. That is down-stream and will not address the near-term problem that may exist today of organizing all the concerns. Let's continue to think about it for the future.

-- RaymondLutz - 12 Nov 2003