Online Computing help systems

Review (October 2009)

Look at some WWW Review documents - in particular, the University's Web Review documents
Review help provision methods used here (RT, Wiki, paper, verbal, WWW, private notes)
Review scope and purpose of help system - educational material? non-CUED material? For help-desk staff rather than users?
Review current help system use
- analysis of web visits (log analysis of entry-points, duration of visits, links followed). Baseline stats.
- analysis of in-person visits. Experiments?
- analysis of existing material - who's it for?
Review WWW/non-WWW help provision elsewhere (Word, our Linux applications, Wizards, Expert Help Systems, etc) and online help provision elsewhere (Judge Business School using a web-based helpdesk package by LBE systems, running on an Access database)
Consider plans by others in Cambridge University (RT help desk and a few stats)
Do a user survey (ugrads, pgrads, academic staff, non-academic staff)
Do a info-provider survey (who adds, who doesn't - and why)
Consider integration of comp-staff material with acad-staff material - lecture videos; lecture notes
Review current search facilities (including how to deal with protected resources). Investigate how people search and the possibility of providing extra search facilities (see "Wolfram Alpha", "Ask.com", "A Study of Web Search Trends", "Lost in Cyberspace: Where to Go? What to Believe?", "More Effective Web Search Using Bigrams and Trigrams") .
jmrm notes that "Search in the context of the University's websites as whole is quite a hot topic at the moment and Autonomy's search product has received some attention as having technology better suited to an organisation's web site rather than Google-type solutions which rely on vast numbers of pages and users."

New System design and delivery

Control of visibility
- Wider access to individual queries? - see Lancaster University's help desk tips
- Access to help information for non-CUED people (security issues?)
More software
- Use a CMS? jmrm writes "day.com's content management looks quite good but [is] by no means cheap. It's based on free software, e.g. jackrabbit and sling, which the next version of CamTools will also be using so there would be a commonality of underlying data format, etc. Bits of the CMS are also, I think, public domain and there's a free download at http://dev.day.com/microsling/content/blogs/main.html along with some comments on Content Integration Standards. This might be useful technology for generating and maintaining help pages, indexing them, and keeping a consistent style. http://www.day.com generally has some useful material."
- Have our own CUED CamTools? Write a CamTools extension?
- More reporting tools? jmrm wrote "How do we ensure up-to-dateness and lack of duplication in our web pages. ... I at least would actually welcome a system like RT's that nags me if I forget provided that, like RT, it provides a mechanism for me to close the ticket ... Maybe we should simply have an RT queue for web pages needing attention with the ticket assigned at the outset to the author?"
- An augmented search facility?
- Create an adaptive system? Allow user-control of amount of detail shown?
Redesign of help system
- Retain overlap of www-h and www?
- If we retain file-based content, do we need the current file structure with a top-level mix of user and subject directories?
- Conversion of existing material to new design and new Univ house style (+ KwikitoHTML?)
- How will staff add material?
- User submission via wiki?
- Metadata?
- Staff training. User training
- Evaluation