Skip to main content

Remaining work and ideas

For developers that will work on Tyle in the feature, these are features that we didn't have time to implement, and ideas that we had but didn't have time to act on during development.

This page does not include things that were already mentioned on the page describing the current state of the code base.

Replacing the authentication solution​

Tyle currently uses authentication/authorization based on Microsoft Identity. The code for this is found in the Mimirorg.Authentication project in the backend.

When Tyle was adapted to the IMF type definitions in IMF 0.2.0, most of the backend was restructured and a lot of the code was rewritten. This is not true for the auth code. The reason was that the development team had a long term plan to replace this auth solution with using Azure AD or something similar for logging on.

If future development teams want to continue using the current auth solution, we would recommend doing some refactoring of the code, as it is currently in a messy state. It also still uses some conventions that has been phased out in other parts of the code, for instance custom exceptions. Also, we would explore the possibility of merging the auth database with the Tyle database, so that for instance creators and contributors to types can be defined by foreign keys in the database.

The current search of the types database leaves a lot to be desired. The most obvious problem is that the client retrieves all the types from the backend, and performs pagination/searching/filtering on the frontend. This should be changed so that these operations are done on the server side, to improve performance and reduce network traffic and database calls.

The search results are also listed chronologically from latest to oldest, instead of being sorted by relevance. This makes the results from the search very unintuitive, and is something that sorely needs to be fixed.

Improving UX/UI​

The development team working on Tyle in 2023 did not have access to design/UX resources. This was a frustration to the team, and also lead to a user interface that is unfinished, and in need of a redesign.

Specifically, one would have to take a look at the type creation forms. There probably needs to be a lot more information available to the user in these forms, both explaining what the different form fields are, and also containing more information about the choices in dropdowns etc. For instance, when selecting the purpose of a block, the only information available is the label/name of the purpose, with no way to find more information. The forms in their current state are mostly useful to SMEs who are already familiar with the IMF type definitions, and able to look up external references to Common Library or PCA.

Some of the components of these forms are also unfinished. For instance, when adding attributes to blocks or terminals, there are boxes that you can use to set the minimum and maximum cardinality of these attributes, but there is no indication in the user interface that this is what these boxes are for.

Expanding the type approval view​

In the current user interface, users can only really see the name and the description of a type that is up for review. This is not enough information to actually do the review, so the users will have to navigate out of the approval interface, and go back to the search menu, finding the type there.

This is obviously not a good work flow, so this section should be expanded with the needed information.

Adding user and access control functionality​

The current Tyle client has the possibility to change roles given to users, but no way to strip a user of all roles, deleting users or rejecting newly registered users. Also, the user does not currently have a way to delete their own account. This is functionality that should be added.

The importance of user feedback​

So far in the Tyle development cycle, usage of the application has been purely for testing. When actual users start using the application, problems and needs will inevitably start popping up, making this list grow. It is important that a future development team pushes for users to start using the application as soon as possible, to get this invaluable feedback.