Introspected REST is a new model for evolvable networked services. It's an alternative to Roy's REST (and GraphQL).
In youtube there is a related talk I had given about a year ago called "Sliding away from Roy's REST model". At that time I hadn't figured out ALL the details about this new model but only the general idea, just mentioning that in case you bump on it.
Please take your time and take a look at it, it's the first time that I share it to the world and any feedback is welcome, here or in Github. It's around 50 pages of text so I understand if it takes you a while..
The reason I spend so much time in explaining REST is because we first need to understand what REST brings into table and what leaves out of the table, before proposing a new model. I hope I make these clear in the text. To be honest I think not many API designers really understand what REST is, unfortunately, that's why a lot of time is spent on that. People think that adding some hypermedia through links or even actions makes your API RESTful. And I am not telling that from the point of view "your API is not REST!" but to the other side of people, the ones who yield that. In fact, a truly REST API is quite complex. After all, REST is out there 20 years not but our APIs are still not truly evolvable and self-described so maybe the problem is somewhere else..
Maybe the text can be improved a bit (maybe explain a bit more the HATEOAS decoupled from REST), but I am very tired with it, mentally. Even proof-reading it takes a couple of hours. I think it's better to take a longer break now and to work on it again later with a clearer mind and after I gather some initial feedback, plus, other people can contribute as well. The idea is to lock the text in 1-2 years from now and publish it in introspected.rest domain.
Not sure what this text is or should, a publication, an article, a manuscript, a thesis or an antithesis. I name it a "manifesto" because I want, when API designers differentiate from Roy's REST and conventional API architectures, to have for them something to stand on, to feel that there is an actual community that supports them and embrace their decisions that break existing "best practices" and move towards Introspected REST.
1 comment
[ 3.0 ms ] story [ 10.6 ms ] threadIntrospected REST is a new model for evolvable networked services. It's an alternative to Roy's REST (and GraphQL).
In youtube there is a related talk I had given about a year ago called "Sliding away from Roy's REST model". At that time I hadn't figured out ALL the details about this new model but only the general idea, just mentioning that in case you bump on it.
Please take your time and take a look at it, it's the first time that I share it to the world and any feedback is welcome, here or in Github. It's around 50 pages of text so I understand if it takes you a while..
The reason I spend so much time in explaining REST is because we first need to understand what REST brings into table and what leaves out of the table, before proposing a new model. I hope I make these clear in the text. To be honest I think not many API designers really understand what REST is, unfortunately, that's why a lot of time is spent on that. People think that adding some hypermedia through links or even actions makes your API RESTful. And I am not telling that from the point of view "your API is not REST!" but to the other side of people, the ones who yield that. In fact, a truly REST API is quite complex. After all, REST is out there 20 years not but our APIs are still not truly evolvable and self-described so maybe the problem is somewhere else..
Maybe the text can be improved a bit (maybe explain a bit more the HATEOAS decoupled from REST), but I am very tired with it, mentally. Even proof-reading it takes a couple of hours. I think it's better to take a longer break now and to work on it again later with a clearer mind and after I gather some initial feedback, plus, other people can contribute as well. The idea is to lock the text in 1-2 years from now and publish it in introspected.rest domain.
Not sure what this text is or should, a publication, an article, a manuscript, a thesis or an antithesis. I name it a "manifesto" because I want, when API designers differentiate from Roy's REST and conventional API architectures, to have for them something to stand on, to feel that there is an actual community that supports them and embrace their decisions that break existing "best practices" and move towards Introspected REST.
/Filippos