I know Giles Bowkett’s schtick is wrapping a thoughtful insight with a bunch of flamebait to generate controversy and thus attention to his projects (here, for Panda Strike), but in this post the hyperbole is limited to the opening image and paragraph. After that it’s all nice code examples comparing the alternate approaches to the problem.
That’s an extremely concise, readable, maintainable representation of the schema. The author glosses over this with “it’s nicer to read than JSON Schema”, which understates how hugely cumbersome it is to write even a simple example like this in JSON Schema.
I feel bad putting my schemata in code. But I still do it, because code is so much better than any of the alternatives.
I know Giles Bowkett’s schtick is wrapping a thoughtful insight with a bunch of flamebait to generate controversy and thus attention to his projects (here, for Panda Strike), but in this post the hyperbole is limited to the opening image and paragraph. After that it’s all nice code examples comparing the alternate approaches to the problem.
Yeah. And, for what it’s worth, this article is in fact discussing a concern I feel strongly about, as should pretty much any fan of strong types.
There’s one code example that’s crucially missing: what would this schema look like in JSON schema?
That’s an extremely concise, readable, maintainable representation of the schema. The author glosses over this with “it’s nicer to read than JSON Schema”, which understates how hugely cumbersome it is to write even a simple example like this in JSON Schema.
I feel bad putting my schemata in code. But I still do it, because code is so much better than any of the alternatives.