Just a quick update on progress following our most recent Third Party Software workshop - the fifth so far. The workshop was an opportunity to revisit last month's proposed plans on migration from the Government Gateway Transaction Engine and an ID Assurance prototype using HMRC's PAYE Desktop Viewer (PDV) application.
We're quite confident about the potential benefits to be gained from migrating existing transactions to a new service, the Digital Transaction Engine. The new transaction engine would be built on the Digital Tax Platform (our planned single online platform for all HMRC Lines of Business/tax regimes) and give us greater control over third party software interaction. It would also enable us to make improvements to the existing polling mechanisms for submissions, in the form of synchronous responses and potentially even 'call back' features where appropriate, which could reduce waiting times and inconvenience for submitters.
We presented some options for this migration to workshop attendees, which were split generally between replication of the existing document submission protocol or moving to RESTful APIs (a widely-used protocol that allows different pieces of software to talk to each other via the web). The former set of options would require less work on the part of software developers, but would require HMRC to replicate the GovTalk protocol on the Digital Tax Platform and would not offer any additional benefits until rework done by software developers to move to RESTful APIs. We compared some examples of the kind of code change required for each option and developers agreed the additional effort was worthwhile for the potential benefits, as well as for the opportunity to avoid later rework. We took the steer that a RESTful API approach was the favourable option and asked developers to factor those options into their planning for future software releases.
We then gave a walkthrough of PDV in its current form and presented detail on the changes we intended to make to its authentication process, which would allow us to demonstrate a working application of ID Assurance. This would initially be used as a working prototype, but the source code would be made available to software developers. This source could be used by developers to allow one of their own applications to request a token through the ID Assurance process and use the token to authorise the application. We presented a walkthrough of the solution we're looking to build for the prototype and software developers were comfortable with the approach we suggested, so we have some more confidence that we're on the right track for this prototype.
Next month we'll be looking at some questions on security and data sharing, as well as examining progress on the Digital Transaction Engine and PAYE Desktop Viewer strands of work.