The current Document Harvester contains a subset of the FAST headings. The new requirement is for us to allow the whole set of FAST headings. This complicates the current design in a number of ways:
It means copying all the FAST headings into the W3ACT database, and keeping them up to date.
It means all the FAST headings are directly encoded in the page as a simple select list, which doesn't scale well.
There are two approaches:
Automate the process of keeping the W3ACT copy of FAST up to date, expose it via a JS API, and use a fancy autocomplete selection widget.
Use the FAST Assign service to use the FAST directly rather than copying it into W3ACT. And either create our own autocomplete widget or re-use theirs.
The current Document Harvester contains a subset of the FAST headings. The new requirement is for us to allow the whole set of FAST headings. This complicates the current design in a number of ways:
There are two approaches: