[x3d-public] Update of Status: Meeting minutes for ECMAScript/TypeScript/JSON support. Modern shading languages for glTF?

[John Carlson]

A.4 action items in email.   We probably should track tasks in the JSON 
project in Mantis (TODO).

A.5 complete.   Used donall.sh from X3DJSONLD to generate JSON. Inform 
John when new JSON download zip is available.

A.6 is next work task.

Updates to Roy's hand-generated JSON schema is top discussion priority. 
John already has some changes.

Discussion of @geoSystem "N"orthern Hemisphere.   Squaw example still 
has many examples.  There are 90 problems in data with @geoSystem.   
Suggest we take "N" out with X3dToJson.xslt

We need to discuss @rotation subschema in X3D JSON schema.  Roy's schema 
has ranges, which creates 1800 errors when running through the X3D 
resources examples.  These appear in Everit, but haven't found them in 
Ajv yet.    See A.6.

On 3/30/21 9:08 PM, John Carlson wrote:
> Note, A.1 first half is complete.  Progressing with A.4.
>
> A.
>
> 1.  We will work towards testing the X3D JSON Schemas with draft-07 of 
> the JSON Schema (one of the widest implementations of JSON Schema, and 
> most recent one for Python).   This will be our primary short term 
> type-checking mechanism beyond X3DJSAIL-X3DJSONLD and x3d.py. (other 
> type-safe mechanisms appear to be burdensome for entire X3D standard, 
> investigations into profiles welcome).
>
> 2.  We will investigate JSON 5, but we see no reason to support 
> something which is not even on a standards track when our current 
> tools are focused on the JSON schema RFCs.  Is there a JSON 5 schema?
>
> 3.  XMLSpy is an important tool for JSON.
>
> 4.  We will test draft-07 based X3D JSON generated schemas and the 
> hand generated schema (draft-06?) against each other in the JavaScript 
> environment.
>
> 5.  We will support the Java everit JSON validator with draft-07 
> schema, and test it.  John will run everit code against the JSON 
> examples in the archives.
>
> 6.  We will attempt to resolve differences between JavaScript and Java 
> validators, make contacts with JSON schema validator source code 
> authors to ensure equivalent errors are reported.
>
> 7.  We will likely finish the X3D JSON encoding standard (but that 
> wasn't mentioned in meeting today).
>
> 8.  We will continue to support auto-generating the X3D JSON schemas, 
> versions 3.0-3.3, and 4.0.
>
> Note that we have packages for JSON -document to DOM document we can 
> test, including X3DJSONLD.js (found in X3DOM, X_ITE and X3DJSONLD), 
> X3DJSONLD.java (found in X3DJSAIL source code), and X3DJSONLD.cpp 
> (standalone, source code available in X3DJSONLD, binaries for Kubuntu 
> Groovy Gorilla available).   Further languages are upon request.  
> These modules take JSON and produce a DOM document.
>
> B. After that:
>
> 1.  We will support the X3D SAI update with feedback from the 
> ECMAScript/TypeScript community (Andreas, can we count on you to help 
> on X3D SAI standard feedback for JavaScript?  I'm thinking of your 
> X_ITE DOM code which I have not analyzed or even looked at).
>
> 2.  If there is demand, we will write/generate a typesafe ECMAScript 
> SAI based on X3D SAI and use similar patterns to other X3D language 
> bindings.
>
> 3.  If there is demand, we will modify the generator in item B.2 for 
> the ECMAScript SAI to create a TypeScript SAI (funding required, GSoC?).
>
> 4.  We will watch to see what 3D TypeScript libraries become important 
> and figure out if we can support them with X3D JSON import/export.  
> Likely of all things, this is the most important for X3D adoption in 
> ECMAScript/TypeScript.   Watch for more developments in and beyond 
> Three.JS, PLY and STL in the X3DJSONLD package.   Thus after JSON 
> standardization, this will likely become John's highest priority. 
> Similar to how Michalis has converted formats to X3D, I want to 
> support dispersion of X3D JSON to many 
> ECMAScript/JavaScript/TypeScript 3D libraries and more programming 
> languages beyond Java, JavaScript and Python.  We will use Don's 
> X3dToJSON.xslt to provide a bridge between Michalis' code and my code 
> (unless some other ECMAScript translation from X3D to JSON suddenly 
> appears).
>
> ============================================
>
> C. I am interesting in hearing about possible X3D JSON extensions for 
> newer shading languages coming out.  What shading languages are 
> supported in glTF?  Do they support raytracing?  I don't have a 
> raytracing card yet.   I am considering upgrading Mac Mini.
>
> D. We do not see on the fly modification of JSON necessary except for 
> Protos and Scripts  (consider source JSON immutable).   Those should 
> be supported by the X3D browsers, not X3DJSONLD.  X3DJSONLD has an 
> example Script preprocessor, but it's pretty much deprecated in favor 
> of X_ITE's example.
>
> Be excellent to each other!
>
> John
>