[x3d-public] For JSON Schema 06, there's uri-reference; multidimensional string arrays.

[John Carlson]

For JSON schema draft 06, there’s uri-reference.  Should we upgrade to support relative URIs?

This is likely a big problem, and we should probably address it.
------------------------------------------------------------------------------------------------------------------------------------------------------
Another JSON schema issue is how to validate an even number of array values in MFString fields.  We should probably press for this with the JSON schema folks (I’ve already posted to google groups, but I think they’ll recommend using a different MFString encoding in JSON for those fields which have pairs.  Like [[“1”, “2”]].  This may be a big change if we apply it to more MF* types, and impacts WebGL implementations—I think we should keep 1 dimensional arrays for floats and doubles for typed arrays.  I don’t know what the performance impact on unpacking large multidimensional arrays is.  We should probably get on this quick before X3D JSON becomes a standard.

Looks like we made the wrong choice (in choosing single dimensional arrays) for validation on the multiple dimension array items, if JSON schema doesn’t support it, and we can’t get them to.  We probably need to decide this on a field by field basis instead of all MF* fields.  For example, I can see where multidimensional arrays would be good for MFStrings, and bad for WebGL in meshes.

Maybe we need an MFVec2String or MFVec2s type.

Anyone for starting over on converting MFVec* and similar types in X3D JSON?  Can we easily implement multiple dimensional typed arrays in webGL?

This is at least for GeoMetadata summary.  There may be other fields.

Do we need to be concerned about correctness over performance?

How is glTF validating multidimensional arrays, if they have them? Do they have a JSON schema?

I am sorry I pushed for single dimension arrays. It does save on transmit, I think.  I think we should pressure the JSON schema folks, if modulo sized arrays aren’t handled (we’ll find out, I hope).

John

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20171025/4e03a2b6/attachment-0001.html>