Workload Model for Computing @ Macquarie

This project implements the workload model for the Computing Department.

Input is a set of spreadsheets describing unit offerings and allocations, one for each year. Each spreadsheet should have sheets:

Units sheet

Unit Code - eg. COMP1000 Title - unit title Session - 'Session 1', 'Session 2' or 'Session 3' Enrollment - projected enrolment for this offering, integer New Unit - 1 if this is a new unit, blank otherwise, unit will attract a loading Co Taught - unit code of any co-taught unit (eg. COMP6100) Lecture Type - 'Mature', 'Refreshed', as per the workload model, or 'PACE' for PACE units

Activities sheet

Unit Code - matching those in Units sheet Title Session - 'Session 1', 'Session 2' or 'Session 3' Activity - 'Lecturer', 'Convener' Quantity - fraction allocated 0-1 for lecturer/convener, workload points allocated for marking New - additional load for staff new to the unit or new to teaching, workload points Staff - staff name, matching name in Staff sheet Notes - any notes on the allocation

Staff sheet

Name - eg. 'Cassidy, Steve' Adjunct - 'A' if the staff member is an adjunct, blank otherwise S1 - target load for S1 S2 - target load for S2

Running the Model

Requires that you have node.js installed. Download/clone this repository. To install dependencies run:

npm install

The project requires a configuration file to tell it where the spreadsheets are. Copy allocation-config-dist.json to allocation-config.json and edit the path names to point to your copies of the allocation spreadsheets.

To monitor the spreadsheets you also need to edit package.json to put the correct directory name in the "preprocess" line.

To build the web pages and serve them locally run:

npm start

This will serve pages on http://localhost:1234/. It will monitor for changes in the spreadsheets and the project code and should regenerate and refresh the page when you save the spreadsheet file.

Code Structure

The project is in two parts. src/js/preprocess.js reads the spreadsheets and generates a JSON version of the data with workload calculations added. The web interface is then generated via the Javascript incldued in index.html - this is bundled and served by parcel. parcel serve will run a local server and watch for changes. parcel build will build a single file version with all data etc embedded and written to dist/index.html.

The script publish.sh runs the build process and uploads the result to my (Steve's) web space.