In a web application, am I right to say inputs to the boundary class will be HTTPRequest which encapsulates user inputs in GET/POST arrays?
Then output will be a HTTPResponse or HTML string?
I ask this because so far what I’ve seen a boundary class accepts are things like
doSomething(input1: String, input2: String ...)
But in a web app, I dont suppose inputs are passed in as string? A more accurate implementation seems more like below to me?
If steps 1 – 1.2, 2 is same for all use cases, I still include it?
What will you think a sequence diagram for edit profile might look like? Or am I thinking too much for a school assignment?
Or do I go a “don’t think too much route”?
So in summary, what will the boundary class accept in the context of a web application (good to compare briefly differences between a boundary class of a web vs desktop app too)

In the analyses model I would not be bothered with implementation details. So just state the actual data/fields/properties that needs to be transferred. The software architecture document will give guidance on how that data is send to implementation classes. In this model you are still verifying that the ‘WHAT’ is clearly stated. When the design-model comes in play the HOW will emerge from that.
Suppose you would take into account implementation details at this stage how will your model change if you use HTTP GET with queryparameters? HTTP POST? RIA with JSON calls?
I think your model wil change and not for the better.