The scope of the search, as we’re rolling it out in December 2018, includes the following locations:
The Corporate Web Presence and Marketing:
- https://www.ncl.ac.uk/
- https://microsites.ncl.ac.uk/
Academic Resources:
- https://roomfinder.ncl.ac.uk/
- https://blackboard.ncl.ac.uk/
- https://docking.ncl.ac.uk/
- https://my.ncl.ac.uk/staff/
- https://my.ncl.ac.uk/students/
- https://www.ncl.edu.my/itservice
Project Websites:
- https://research.ncl.ac.uk/
- https://teaching.ncl.ac.uk/
- https://conferences.ncl.ac.uk/
Personal Publishing:
- https://www.societies.ncl.ac.uk/
- https://www.staff.ncl.ac.uk/
- https://www.students.ncl.ac.uk/
- https://blogs.ncl.ac.uk/
This search is designed for external facing use, so does not index NUConnect or other elements of the Newcastle University Intranet (NU Connect runs on SharePoint, which already has SharePoint Enterprise Search capabilities built in, so we need to look at how Funnelback and SES work together and compliment each other, not duplicate or replace the other).
A large quantity of content is already filtered from the results as these pages contain some content that’s not immediately relevant to searches (RSS feeds for example) or only internally accessible.
This results in an index of around 80,000 pages.
Naturally, some parts of this estate are higher priority than others, but the algorithms built into the system should be able to address this without too much manual intervention..
That said, it may take time to tune the new system based on feedback received and requests to add, remove or prioritised parts lower the priority of something else. We’ll be working with stakeholders over the coming weeks to ensure that feedback is addressed with care.
There are a lot of factors to consider, in particular how we handle duplicate content – which can dilute its effect – a couple of examples:
- Many members of staff have multiple profiles across many websites.
- Often news is duplicated across multiple sites – and syndicated onto many lists.
- Blog or News Style sites generate a lot of pages of lists of their content, sometimes many times, but with different filters applied.
Future projects will be looking at integrations with other systems (Library Searches, NUConnect, ePrints, and other knowledge bases)
Please submit Feedback, Suggestions and Comments here.