Re: [httpapi] Discussion about JSON payloads and code generation

Christoph Kappestein <christoph.kappestein@gmail.com> Sun, 24 January 2021 09:46 UTC

MIME-Version: 1.0
References: <CALcRZn6-ojAAdJcMWFHef70Xp32O2iFatuw-YjGLKtr8VnbmYQ@mail.gmail.com> <DM6PR00MB0845824BE38945071F9774A5F0A69@DM6PR00MB0845.namprd00.prod.outlook.com> <CALcRZn6HdXE2WAn0vSod0XDY_yJd7jV7m5JnRWKWGDJaDaag-A@mail.gmail.com> <BL0PR00MB083637DAD075EC8E5342A0A9F0A59@BL0PR00MB0836.namprd00.prod.outlook.com> <CAEdRHi4q4d4RF1x8YJcJiWrMSij9FjK9pWjj7DMygqX38dZC1w@mail.gmail.com>
In-Reply-To: <CAEdRHi4q4d4RF1x8YJcJiWrMSij9FjK9pWjj7DMygqX38dZC1w@mail.gmail.com>
From: Christoph Kappestein <christoph.kappestein@gmail.com>
Date: Sun, 24 Jan 2021 10:46:02 +0100
Message-ID: <CALcRZn5QMQK3xzKHSYJAsbzvHN3SJ2xNJGZ2R1i60Tc543+xxQ@mail.gmail.com>
To: "httpapi@ietf.org" <httpapi@ietf.org>
Cc: Darrel Miller <Darrel.Miller=40microsoft.com@dmarc.ietf.org>
Content-Type: multipart/alternative; boundary="0000000000001738c005b9a24a2d"
Archived-At: <https://mailarchive.ietf.org/arch/msg/httpapi/eRdooimyR3DEWdyCxFxL_ICHZAA>
Subject: Re: [httpapi] Discussion about JSON payloads and code generation
Precedence: list

Hi,

Thanks for the feedback.

@Jason this is true and I totally understand your concern. I think the
difficult
part is to define a "core" which works for every use case i.e. code
generation,
validation, documentation, etc. We should ask what is the common ground of
all
these specs? In JSON Schema this is probably only the part how documents and
types are referenced (JSON Path). But thanks for the offer and the great
feedback which you have provided, I will also give you an answer at our
github
issue, if I find the time finally ;)

@Darrel the idea sounds great to define the representation of data types. At
TypeSchema we have defined the following primitive types:
- date (rfc3339 full-date)
- date-time (rfc3339 date-time)
- time (rfc3339 full-time)
- duration (ISO 8601)
- binary (base64 encoded)

We have found that most programming languages have internal object
representations for these kinds of types. If we would have an official list
with
a fitting identifier for a type, then we could also use this identifier at
the
"format" in JSON Schema/TypeSchema to describe the type.

I am wondering whether it would be also possible to describe complex types
in such a specification, i.e. that we describe how a list, map or
object/struct
type could look and is defined through keywords. So I would totally support
this
effort, since this is basically the reason why we have developed TypeSchema
and
we would be happy to build upon such a specification.

best regards
Christoph


Am Mo., 18. Jan. 2021 um 11:17 Uhr schrieb Asbjørn Ulsberg <
asbjorn@ulsberg.no>:

> søn. 17. jan. 2021 kl. 22:17 skrev Darrel Miller
> <Darrel.Miller=40microsoft.com@dmarc.ietf.org>:
>
> > In the spirit of "do one thing and do it well" we could have different
> > specifications for specific use cases, which would be a much more
> flexible
> > framework than forcing every schema to be a vocabulary. A generator
> could then
> > support only specific schema types.
>
> Ulysse Carion has already created a schema language specifically
> tailored for code generation with RFC 8927.
>
> https://www.rfc-editor.org/rfc/rfc8927.html
>
> --
> Asbjørn Ulsberg           -=|=-        asbjorn@ulsberg.no
> «He's a loathsome offensive brute, yet I can't look away»
>

[httpapi] Discussion about JSON payloads and code… Christoph Kappestein
Re: [httpapi] Discussion about JSON payloads and … Darrel Miller
Re: [httpapi] Discussion about JSON payloads and … Christoph Kappestein
Re: [httpapi] Discussion about JSON payloads and … Jason Desrosiers
Re: [httpapi] Discussion about JSON payloads and … Darrel Miller
Re: [httpapi] Discussion about JSON payloads and … Asbjørn Ulsberg
Re: [httpapi] Discussion about JSON payloads and … Christoph Kappestein
Re: [httpapi] Discussion about JSON payloads and … Darrel Miller