Week ending: 2019-10-18
Week notes are a way for us to keep people informed about progress on the project. Given the technical nature of the re-platforming work we will use them to explain technical choices that we are making, including the benefit and impact of these choices.
The goal for this week was to better understand the project’s technical constraints, and to prioritise the most important elements.
- Enable housing officers to use Manage a Tenancy offline when there are mobile coverage blackspots
- Enable Hackney to decommission Outsystems, saving £80,000 per year
- Enable Hackney to support, develop and deploy future improvements to Manage a Tenancy more quickly, at lower cost
The team now has a very clear steer on scope and priority, which we can turn into a more realistic roadmap for this phase of the project. We will focus initially on building offline processes rather than the hub and worktray; the benefit is that we can meet the needs of the Housing team first and foremost:
- It is the most rapid route to making the digital processes work offline
- It gives Hackney the greatest opportunity to practice building processes in React, putting the team in the best position to digitise further processes in future.
We also discovered that the work to replatform My Rent Account should be relatively straightforward; this is what we were expecting, but it was good to validate this assumption. This is work that the team will tackle at a later date, after the initial dxw-phase of the project is complete.
We learned more about the API that the Hub and worktray currently use, and we explored different ways in which we can ‘futureproof’ this aspect of the work. Importantly, we agreed on a way in which we can decouple the front-end work from the API; we can do this by having an object in the front end that can deal with changes to the API endpoints.
This was a useful discussion and should help Hackney in the longer term; we will continue exploring this once we start delivery on 30th October.
We learned about the status of the continuous integration pipeline project, and were able to feed into the planning of that work. This was very useful, and should remove potential blockers further down the line.
Other things we learned:
- The Google SSO project is at a relatively early stage for Hackney; this is now something that we can safely deprioritise from the Manage a Tenancy work for now
- The Outsystems contract is in place until September 2020
- The Qlik dashboards pull data either directly from Dynamics or from UH; this is no longer a dependency on us
- The work to replatform My Rent Account is relatively straightforward, and this can be picked up after the initial dxw phase of the project, and will benefit from the React groundwork that is laid in the Manage a Tenancy phase.
The trade-off that comes with prioritising offline processes is that the replatforming the hub will need to come at a later stage in the project. Although this can be done in time to decommission Outsystems, this does carry some degree of risk.
The work involved is largely front-end development; this is not currently a widely held skill in Hackney Council. A good mitigation would be for Hackney front end developers to get involved in replatforming at least one process, so that they can learn how to build things in React.
- We have a clearer and more realistic scope for the team to focus on during the first phase of this project; thank you to Matthew for giving us the steer that we needed.
- Thank you to David, Mirela and Shweta for being so flexible when we had to move meetings around because of illness.
We will :
- Update high-level roadmap that we created together, which will help us prioritize work and talk about the risks and dependencies more clearly.
- Amend the statement of work to reflect the revised scope and priorities.
- Carry out some technical prep before we start sprinting on 30th Oct; this includes things like:
- investigating how we might build React and HTML/CSS/JS components from the same source for Hackney’s component library
- looking into how we might automate testing our internal reflection of the API against the Swagger JSON