Skip to content

LongestPath.md

Latest commit

 

History

History
54 lines (46 loc) · 1.97 KB

LongestPath.md

File metadata and controls

54 lines (46 loc) · 1.97 KB

Longest Path (LP)

Similar to Average Path Length (APL), the Longest Path metric focuses on the size of the resource paths of an API. Instead of the average length, it represents the maximum length, i.e. the longest resource path. Again, the path length is defined by the number of slashes ("/"). Slashes at the end of the path are not counted.

The example shown below has three resource paths, namely /github/repositories/{username}, /github/repositories/{username}/{slug}, and /gitlab/repositories/{username}/{slug}.

/github/repositories/{username}:
    get:
        operationId: getRepository
        parameters:
            - name: username
              in: path
              required: true
              schema:
                type: string

/github/repositories/{username}/{slug}:
    get:
        operationId: getRepository
        parameters:
            - name: username
              in: path
              required: true
              schema:
                type: string
            - name: slug
              in: path
              required: true
              schema:
                type: string

/gitlab/repositories/{username}/{slug}:
    get:
        operationId: getRepository
        parameters:
            - name: username
              in: path
              required: true
              schema:
                type: string
            - name: slug
              in: path
              required: true
              schema:
                type: string

The first path is three elements long (/github, /repositories, and /{username}). The second and the third path are both four elements long. To calculate the LP value for this interface, we identify the maximum of these path lengths:

LP = MAX(3, 4, 4) = 4

Source

F. Haupt, F. Leymann, K. Vukojevic-Haupt. "API governance support through the structural analysis of REST APIs." In: Computer Science - Research and Development (July 2017). doi: 10.1007/s00450-017-0384-1 (cit. on pp. 21, 31, 32, 36, 37).