What is the appropriate way to model API requests in sequence diagrams?
I'm going to create some sequence diagrams to model interactions in a web API.
There is a duplicate request / response pattern that I'm not sure how to simulate.
Let's say I have an API request to request attributes from an API endpoint. Would it be advisable to use a standard response template like this? Good style for assigning multiple return value variables ( encrypted_attributes, hashes
)?
Or custom queries / responses so whether to create an image that both request_attributest
and return_encrypted_attributes
are performed independent of each other entities?
source to share
Answers to what is appropriate or what is good style can only be opinion based. I don't have a strong opinion on this, but I see some solutions:
1) Sparx Systems Enterprise Architect can draw synchronous messages with an "implicit response message" as shown here: http://www.sparxsystems.com/resources/uml2_tutorial/uml2_sequencediagram.html
Only 1 line (not 2 lines). Simple diagram, easier to read.
2) Google: "UML Sequence Diagram Best Practices" returns multiple assets with pros / cons.
3) Usually, when a solution like this is needed, following the path of least resistance (only 1 mile), also known as the KISS principle , is a good guide .
source to share