Skip to content

Lessons from the Post Phase 1 Discovery Sprint

Jump to bottom
Dave Methvin (USDS) edited this page Mar 29, 2019 · 1 revision

After the completion of Phase 1, the team performed a discovery sprint to determine what others in the government might be able to use a tool like US Forms System. Although there was some interest, no group stepped up to make a commitment to work with the team.

As a result, the USDS staff involved with USFS have ramped down their commitment for the moment in order to focus on higher priorities. If work were to resume, it would best be done by refreshing the discovery work done here and starting with a design that addressed those needs.

Obstacles

In talking with prospective library adopters, we encountered several recurring issues:

Timing: There is a sweet spot to hit when engaging with a team regarding the USFS. If they are too far into their own development, they have already committed to a solution and would need to revisit design decisions to use USFS. Conversely, if they are still in early planning stages their actual use of USFS may not come for a year or more.

Opinionated design: Although the library is now independent of the VA website, VA-influenced aspects of the USFS library still affect the way it looks and behaves. Those opinionated decisions do not always match the design preferences of the projects evaluating USFS. Although we consider these to be legitimate concerns, it would require significant USFS team effort to address many of them. Even then, there is no guarantee the USFS would be used any particular project if we were to undertake changes to provide the desired design flexibility.

Wanting a turnkey solution: In version 1.0, the USFS is a library that still requires experienced React developers to take across the finish line. It can make their job easier, but does not eliminate the need for developers, which makes it a less-than-perfect solution for teams wishing to build forms.

Wanting something different: Projects staffed with designers and front-end developers may prefer to create something unique to their own environment that fits their view of the problem solution, if their budget allows.

Underestimating effort: A significant amount of design and development effort went into USFS, and even now the VA team is continuing to find specific USFS issues that need to be addressed in layout, Section 508 compliance, and missing features. When teams decide they want to go their own way, they often underestimate the amount of work it will be to build such a system versus the comparative perceived discomfort of accepting USFS design decisions that they may not completely agree with.

New features

Users identified several things that might be useful in the new library:

  • Internationalization and localization
  • More and better widgets (e.g. date pickers)
  • Support for other tech stacks (e.g. Angular)
  • Support for other backends, or a back-end service

There are also ongoing fixes and enhancements being done by the VA team in their forms system that should be incorporated into new designs.

Notable technical debt

The team has identified several issues with the current codebase that may require significant refactoring work. It would be worthwhile to re-investigate whether these issues could be addressed more effectively by high-level design changes. Several of them already have tickets with useful discussion.

Home

Project

Releases

Developer Docs

Community

Research

Clone this wiki locally