maurolepore
commented
1 week ago Topics (GitHub labels)
Broadly speaking the work ahead relates to the following topics, for which I may create GitHub labels:
- Authentication: How we implement it, what/how can we customize, if we pay, and who/how/when we pay, etc.
- Filters: What the use can filter for and how.
- Dictionary: Metadata about our tables. Who creates it, what/where/how the user will see it.
- Data: Tables resulting from the output of of software. What/where/how the user will see it.
- Plots: Visualization of our data.
- Downloads: What data the user can download and how.
- Text: Text on the app itself, e.g. welcome message, disclosures, labels for selectors.
- Links and embeddings: Links to methodology document, embeded videos or links to them.
- Aesthetics: theme, colours, fonts
- Layout: Pagination, placement of filter-controls, tables, buttons, etc.
Priorities
Today Anne and Tilman shared their priorities.
- Authentication: Implement user authentication in a way that is out of the box but also custom (potential conflict here).
- Filters: Pending review.
- Dictionary: WIP
- Data: WIP
- Plots: Drop for now.
- Downloads: No comment
- Text: WIP but I could facilitate their contributions
- Links and embeddings: WIP but I can facilitate their contributions
- Aesthetics: Anne seems to already have found what she likes. TODO: facilitate her contribution
- Layout: Pending
https://github.com/2DegreesInvesting/TiltDevProjectMGMT/issues/155#issuecomment-1997617776
Estimate
Revisiting my estimate on March 14
It seems we're on the most complex end of the range for almost every task, and that didn't include tasks like authentication which I didn't foresee at the time. So if we add authentication my new estimate is around 40 units of effort -- where 1 unit of effort approximates one full work days of exclusive focus.
In practice, in the current situation I estimate to complete 40 units of effort in about 6 months.
Here is why.
I typically spend a lot of time in other tasks -- like reviewing code, communicating with you and other developers, and managing projects. Let's say this brings my availability down to 1/2, and then again an extra 1/2 considering I work 3 out of the 5 weekdays. That's really a bit too much so lets adjust and say that my availability is not only 25% but 35% instead. A month has about 20 workdays, and 35% of that is 7 days. So in one month I only really have 7 days for the web tool -- meaning I would complete the estimated 40 days of full work in close to 6 months (excluding holidays and other unexpected delays).
Evidence
My velocity chart for the overall TiltTech project shows that each sprint I achieve around 7 units of effort on average. This is a bit surprising given that sprints where first 1 week long and later 2 weeks long. This may reflect that as tilt matured I have spent increasingly more time in tasks not directly related to issues assigned to me, e.g. reviewing code or communicating with managers.
My velocity chart for the tiltWebTool project shows a similar picture, althought as of today the dataset is too small to really be meaningul.
Taken together the evidence supports the idea that 7 units of effort weekly is a reasonable average and therefore an estimate of 6 months makes sense.
Resources
https://www.linkedin.com/pulse/estimate-your-velocity-insights-github-projects-mauro-lepore-9ri4e/