[x3d-public] Inconsistencies in Requirements [was: Validation improvements for "USE" attribute]

Sun Jun 18 10:18:16 PDT 2017

I read the message below yesterday and I don't see how it answers the 
issue of USE statements. Perhaps it is my lack of understanding of 
something here. Can it the answer please be restated. To help things 
along I am summarizing my points.

1) The X3D specification requires nodes to have more than one parent in 
the case of DEF/USE.
2) For V4, there are statements that all X3D nodes will be in the DOM.
3) For V4, there are statements that it will be backward compatible (as 
much as possible)
4) The DOM specification prohibits nodes with more than one parent.

It is not possible for Web3D Consortium to change the DOM specification 
(though it is possible to petition for a change). It is possible to 
change the X3D specification (1). It is also possible to relax the 
statements in (2) and (3).

There are inconsistent requirements if DEF/USE means a reference to the 
DEF node and all nodes are in the DOM.
There are consistent requirements if DEF/USE means a copy of the node 
tree originating at the DEF node (drop #1&3)
There are consistent requirements if not all X3D nodes are in the DOM 
(drop #2)

I would be interested in hearing anyone else's ideas on how the 
inconsistencies in existing statements and specifications can be resolved.

Leonard Daly

P.S. regarding 'class' and USE. Going strictly by the X3D 
specifications, it makes no sense to allow 'class' in a node as the USE 
attribute indicates that what appears in this position in the scene 
graph is a reference to a node defined elsewhere. There is no provision 
for this DEF node to be anything other than a reference.

> Primary is getting to clarity on best possible USE definition for X3D 
> per se.  We discussed Thursday on spec teleconference.
>
> It would seem that allowing different 'class' attributes on USE in the 
> XML Encoding is over-generous and should be tightened to not be 
> allowed.  We were able to come up with examples that showed diverse 
> class+USE is problematic (e.g. cannot style a Material node to be a 
> class='somethingBlue' while styling a USE version of the same node to 
> be class='somethingRed').
>
> Next will be considering if 'class' attribute can be advanced to 
> abstract spec and hence all encodings; currently it is only 
> reserved/defined in XML Encoding.
>
> Regarding DOM tree and X3D tree, they do not have to be considered as 
> necessarily identical all of the time.  If event models are decoupled 
> and handing off in tandem, then synchronization would occur after each 
> respective update.  There are different ways for implementations to 
> accomplish this - all worth considering, with HTML5/DOM 
> Recommendations being authoritative regarding functionality.
>
> So, step by step.  If we can define semantics that work 
> consistently/coherently, and can be implemented, we then refine 
> iteratively for use in each X3D encoding.
>
>
>
> On 6/15/2017 10:58 AM, Leonard Daly wrote:
>> The original purpose (and still used in this manner) of the 'USE' 
>> attribute is to indicate that another node should also appear in 
>> place of the node declaring 'USE'. In fact the specification states 
>> (4.4.3 - 
>> http://www.web3d.org/documents/specifications/19775-1/V3.3/Part01/concepts.html#DEFL_USESemantics) 
>> that "the same node is inserted into the scene graph a second time, 
>> resulting in the node having multiple parents".
>>
>> This requirement is not allowed in DOM (see 
>> https://www.w3.org/TR/dom/#concept-node-tree for the standard, 
>> https://www.w3.org/wiki/Traversing_the_DOM#Nodes for the 
>> explanation). A DOM element is allowed to have at most one parent. It 
>> is possible to create a (deep) copy of the node and insert it into 
>> the tree. That gives a structure like:
>>
>> [Meant to be seen in fixed width font]
>>
>>    B - C - D
>> /
>> A
>>   \
>>     E -CC -DD
>>
>> Where A is the parent of this (sub-)tree, B is the node that start 
>> one branch (e.g., Transform). C is the 'DEF'ed node with a child of 
>> D. E is a separate child of A (e.g., a different Transform) CC is the 
>> 'USE' version of C. Since HTML does not allow multiple parents ('B' 
>> and 'E'), a copy of 'C' needs to be made. This needs to be a deep 
>> copy (all children) as no node can have more than one parent.
>>
>> The problem with a deep copy is that it is non-deterministic as the 
>> element is self-referential (it refers to it's parent, which refers 
>> to it...)
>>
>> It seems to me that there is a conflict in requirements between X3D's 
>> statement on DEF/USE and the requirement to put all X3D nodes in the 
>> DOM. There are several ways around this:
>> 1) Remove the multiple parent requirement from DEF/USE
>> 2) Remove the requirement of all nodes being in the DOM.
>>
>> Each has advantages and disadvantages. Which choice is made 
>> determines how X3D operates in the HTML/DOM environment.
>>
>>
>> Leonard Daly
>>
>>
>>
>>
>>
>>
>>
>>> Hi all,
>>>
>>> Recently we updated the DTD/schemas to make the “name” field of 
>>> nodes like MetadataBoolean, or FloatVertexAttribute a required 
>>> field. However, we then realised that when any of these nodes has 
>>> the “USE” attribute, the “name” field must not be set. Hence the 
>>> changes needed to be revisited.
>>>
>>> I started to look at the JSON schema for the MetadataBoolean node, 
>>> to try to implement this stricter validation requirement. With some 
>>> online assistance I found that I could easily make either one or the 
>>> other required, but not both. This would meet the original requirement.
>>>
>>> However, this raised a more general question in my mind. When any 
>>> node has the “USE” attribute set, what other fields/attributes are 
>>> permitted?
>>>
>>> As a test case, I concentrated on the MetadataBoolean node. I came 
>>> up with a JSON schema that might illustrate this better. I have 
>>> attached a snapshot image for this fragment of the JSON schema.
>>>
>>> The validation of the MetadataBoolean node begins with a “one of” 
>>> operator (shown immediately to the right of the MetadataBoolean box. 
>>> This operator requires that one, and only one, of the two subschemas 
>>> be valid. For the first subschema (the upper of the two) I simply 
>>> removed the “@USE” property, making the “@name” field required (to 
>>> meet the original requirement). For the second subschema (the lower 
>>> of the two) I made the “@USE” property required, and only added the 
>>> “IS” property. Note that both subschemas only permit those 
>>> properties listed (i.e. additional properties are disallowed).
>>>
>>> In principle, there should be no difficulty applying this validation 
>>> methodology within the JSON schema to all nodes.
>>>
>>> Some questions:
>>>
>>>  1. Is this validation methodology correctly aligned with the 
>>> standards?
>>>  2. Do we want to apply this methodology to all nodes?
>>>  3. Do we want to apply this methodology to other validation tools, 
>>> e.g. Schematron, and also consider whether the XML schema or the DTD 
>>> have sufficient expressive power too.
>>>
>>> All comments appreciated,
>>>
>>> All the best,
>>>
>>> Roy
>>>
>>>
>>>
>>> _______________________________________________
>>> x3d-public mailing list
>>> x3d-public at web3d.org
>>> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>>
>>
>> -- 
>> *Leonard Daly*
>> 3D Systems & Cloud Consultant
>> LA ACM SIGGRAPH Chair
>> President, Daly Realism - /Creating the Future/
>>
>>
>> _______________________________________________
>> x3d-public mailing list
>> x3d-public at web3d.org
>> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>>
>
>
> all the best, Don

-- 
*Leonard Daly*
3D Systems & Cloud Consultant
LA ACM SIGGRAPH Chair
President, Daly Realism - /Creating the Future/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20170618/76e05456/attachment.html>