Customizing the layout (or look and feel) of PlantUML diagrams is tricky. There is a dense and comprehensive documentation, but there is still a gap between what we want and what we get. There are tricks and hacks to overcome default behavior or missing features of PlantUML. Here I am sharing a fun session with ChatGPT that have successfully assisted me in a quite creative way! The solution seems not to be in the documentation, and the “hack” can certainly be systematically applied. Or did I miss something with PlantUML?

Session and first tries

I had this PlantUML model and class diagram:

@startuml
class Model {
}

class Poll {
    +name: String
}

class Question {
    +id: String
    +text: String
}

class Option {
    +id: String
    +text: String
}

Model --> "1..* polls" Poll
Poll --> "1..* questions" Question
Question --> "1..* options" Option
@enduml

The layout was not as I wanted.

So, I asked the right question to ChatGPT, with a perfect English:

How to change the layout and have a horizontal?

ChatGPT recommended to use left to right direction. Indeed, it’s something I’ve used in the past and actually written in the documentation.

Unfortunately, the cardinality and roles’ names are not really readable. The association should be longer and well, we need more space.

How to enlarge space from left to right?

The recommendation about packagePadding parameter was not helpful.

Creativity and tricks: a signal

It has no visible effect.. any other idea?

And here it’s beginning to be interested:

@startuml
left to right direction

skinparam classBackgroundColor Transparent
skinparam classBorderColor Transparent
skinparam shadowing false

class Model {
}

class Spacer1 << (I,transparent) >> {
}

class Poll {
    +name: String
}

class Spacer2 << (I,transparent) >> {
}

class Question {
    +id: String
    +text: String
}

class Spacer3 << (I,transparent) >> {
}

class Option {
    +id: String
    +text: String
}

Model -[hidden]-> Spacer1
Spacer1 -[hidden]-> Poll
Poll -[hidden]-> Spacer2
Spacer2 -[hidden]-> Question
Question -[hidden]-> Spacer3
Spacer3 -[hidden]-> Option

Model --> "1..* polls" Poll
Poll --> "1..* questions" Question
Question --> "1..* options" Option
@enduml

It’s not working, but the idea is brilliant: faking elements to enlarge spaces and forcing the layout to move the cardinalities/roles’ names… and then hiding these elements!

Complaining

A magical recipe in prompt engineering is of course to complain

not working

and here is the new proposition with “notes” to fake modeling elements.

@startuml
left to right direction

class Model {
}

note " " as N1

class Poll {
    +name: String
}

note " " as N2

class Question {
    +id: String
    +text: String
}

note " " as N3

class Option {
    +id: String
    +text: String
}

Model -[hidden]-> N1
N1 -[hidden]-> Poll
Poll -[hidden]-> N2
N2 -[hidden]-> Question
Question -[hidden]-> N3
N3 -[hidden]-> Option

Model --> "1..* polls" Poll
Poll --> "1..* questions" Question
Question --> "1..* options" Option
@enduml

Very clever ChatGPT! A remaining issue is that “notes” are visible… And from what I have learned in prior interactions (see above) it’s certainly possible to hide elements.

ChatGPT + Human to hack

can't you use skinParams now to make notes invisible?

and here is the final result

@startuml
left to right direction

skinparam noteBackgroundColor Transparent
skinparam noteBorderColor Transparent
skinparam noteFontColor Transparent
skinparam shadowing false

class Model {
}

note " " as N1

class Poll {
    +name: String
}

note " " as N2

class Question {
    +id: String
    +text: String
}

note " " as N3

class Option {
    +id: String
    +text: String
}

Model -[hidden]-> N1
N1 -[hidden]-> Poll
Poll -[hidden]-> N2
N2 -[hidden]-> Question
Question -[hidden]-> N3
N3 -[hidden]-> Option

Model --> "1..* polls" Poll
Poll --> "1..* questions" Question
Question --> "1..* options" Option
@enduml

Absolutely perfect!

Conclusion

ChatGPT was a clever assistant to customize the layout of a PlantUML diagram. Mastering the documentation (aka RTFM) was not in my capacity… Teaming up with ChatGPT was a way to (1) speed up or remember some knowledge of PlantUML; (2) get inspiration of some possible strategies. At the end, I used what I’ve learned about PlantUML to eventually adapt the inspired and creative proposition of ChatGPT. And retrospectively I have impression the proposed “tricks” of faking model elements and then hiding them is a very effective strategy for PlantUML. What are your best practices for PlantUML layouts? Or am I just an ignorant of how to properly use PlantUML?