[x3d-public] good example for ECMAScripting and Protos?
Andreas Plesch
andreasplesch at gmail.com
Mon Aug 6 20:33:57 PDT 2018
SVG might be a model to follow as it is well integrated with html5.
SVG has defs:
https://svgwg.org/svg2-draft/struct.html#Head
and use
https://svgwg.org/svg2-draft/struct.html#UseElement
A use element clones a copy of the referenced element. This is different
from a x3d USE reference which _is_ the referenced element (which can have
multiple parents). Not sure when and how often the cloning happens.
I think the only purpose of defs is that included elements are not
immediately rendered. All elements with an id can be referenced by use.
Option (1) is also what x3dom does with its parallel tree where the actual
x3d scenegraph is represented. The dom is only an interface, and the dom
nodes in the x3d node the data.
That is why it is possible to use the parallel tree approach for X_ITE as
well. It is possible to treat dom nodes as data here as well and map them
on the fly and dynamically to the X_ITE scene graph.
Some more or less connected observations:
There is still demand for relatively simple, declarative 3d on the web.
A-Frame and Xseen are really the only other option in this respect.
For very performant, fancy 3d raw webgl, or three.js or similar engine will
be preferred since it allows limitless tweaking and minimal overhead. Even
glTF is considered basic, and has already a lot of vendor extensions. It is
used outside of the web also, as a replacement for fbx.
While it is really nice to have cross-platform compatible scripting, x3d
scripts and dom scripts are just too similar to justify introducing x3d
scripts in a web page environment. Then how could routing be integrated ?
Let's see.
from script: We can generate all types of x3d field values with a Dom
script using x3dom.fields. methods. Then we can use direct output style to
change values of target node fields. Or we can use strings and HTML
attributes to set a node field from a script.
to script: this is more tricky. The currently available pathway is the
onoutputchange event which is available for out fields of all nodes when
they are updated. So a 'route' consists of a listener to this event on the
from node which then makes the output value available to a function - the
script.
Another pathway from a x3d node to an external Dom script could be through
a gate node. It would be in effect a no action x3dscript node which defines
all userdefined fields. It could be routed to normally without code
changes. A Dom script could then redefine the gate node instance's
fieldChanged() method which is called by x3dom each time the node receives
input. This would be quite similar to the x3d script fieldname functions.
Then we can use direct output style or the .postMessage method to generate
out events which can be routed.
This seems quite complicated, so why not just use direct manipulation of
x3d nodes and outputchange dom event to receive output from x3d nodes ?
Andreas
On Mon, Aug 6, 2018 at 1:41 PM Leonard Daly <Leonard.Daly at realism.com>
wrote:
>
> John,
>
> In HTML a node can only appear once in the DOM. It is not possible for a
DOM node to DEF/USE in the sense that X3D requires it. It is not required
for the 3D scene graph to be in the DOM. That makes the 3D content, data
within the HTML page. From that perspective (and speaking of X3D here), X3D
would be embedded in the HTML page, but not part of the DOM. This is
somewhat equivalent to using a 'data:' URI for an image. The information
would be all there, but the individual nodes would not be accessible as DOM
nodes unless explicitly added by the code processing the 'data:' URI.
>
> As far as I have been able to tell (and I have been studying this issue
for 3+ years), the choice is
>
> 1) Include X3D code (tags, protos, scripts, etc.) as data within the HTML
page and do have it as nodes in the DOM
> 2) Change X3D so it is compatible with and follows the rules HTML5 making
X3D nodes full members of the DOM
> 3) Change HTML5 so it is compatible with X3D.
> 4) Decide that X3D and HTML5 exist is two separate spaces and should not
be integrated.
>
> (1) is certainly an option and (I think, perhaps Andreas should comment
here) this is similar to what X_ITE does
> (2) has been raised for several years and some people are absolutely
opposed to this
> (3) not going to happen
> (4) not an expressed desired of the Consortium.
>
> XSeen has taken the (2) approach. It supports the DEF/USE concept for
selected (proper) subset of node capabilities. It also focuses on assembly
/ integration of scene components rather than modeling objects. It is
expected that most models will be provided by glTF.
>
> Leonard Daly
>
>
>
> Leonard, I agree my approach may not work in all environments.
>
> Determining a valid or invalid mime type is extremely difficult if done
semantically, but extremely easy if done syntactically. My guess is that
many of the X3D tags are semantically invalid, unless someone has added
them to Apache. I do not know the status of this. I have to add mine
types to my server for every new extension I add. It is not a that big of
a deal.
>
> One may not need a script tag around X3D script to execute X3D script.
One may use eval in a script tag. One may make script inconsequential by
making the scripts only functions, and move the functions to a string which
is evaled. Essentially you are parsing twice.
>
> If we want script fields handled by X3DOM, we will have to find someway
to bring them into X3DOM. X_ITE has shown the way. We can take other
approaches than X3DOM, like straight DOM. I am beginning to think this is
preferable, but using DEF to find node which aren’t in the scenegraph has
kind of been, wtf?
>
> The question has become “Are DEFed SFNodes valid scenegraph objects if
they are in fields?” This may show a failing of X3DOM, which may need to be
corrected. I can probably find the scengraph object using USE, but I
really want to use DEF. Can someone show me how to do it in the HTML5
environment?
>
> John
> On Mon, Aug 6, 2018 at 12:24 AM Leonard Daly <Leonard.Daly at realism.com>
wrote:
>>
>> I am replying to an early message in this thread, but I hope to capture
comments from all messages in the thread. My comments only apply to the web
browser (HTML5) environment.
>>
>> First, any Script tag (in any case) will be handled by the web browser,
and the web browser will parse content before anything else. If you really
need to independently parse some tags, you will need to "read" it into a
JavaScript variable and handle it that way. If it's in the page file, it
will be parsed.
>>
>> There was a suggestion to use "text/x3dscript" or some other string. The
W3C has something to say about this at
https://www.w3.org/TR/html5/scripting-1.html#the-script-element. The value
of the type attribute must either be
>>
>> omitted
>> JavaScript MIME type
>> "module"
>> any other valid mime type
>>
>> So the use of an invalid MIME type is not technically valid HTML. This
may cause a problem with some browsers or validators; or present future
problems if that string is ever defined to have a particular meaning.
>>
>> X3DOM does not have X3D scripting. If you need scripting to correctly
process an X3DOM environment; then you need to do it in HTML5 Javascript.
X3DOM does not process the Script tag, that is done by HTML5. While X3DOM
parser could recognize that tag, it does not; hence, none of the X3DOM
methods are available to handle interactions on a Script node (a Script tag
parsed into DOM).
>>
>> The X3D Script / HTML5 Script tag name conflict is a long-standing and
known issue. To my knowledge no work has been done to break the conflict.
>>
>> Leonard Daly
>>
>>
>>
>>
>>
>> Here’s my current code for clearing the ECMAScript out of a X3D file so
that it doesn’t show on the screen:
>>
>>
>>
>>
>>
>> $(selector+" Script").contents().filter(function
() {
>>
>> return this.nodeType === 3 || this.nodeType
=== 4;
>>
>> }).remove();
>>
>>
>>
>> This retains the fields. I may change my code to work with fields
instead of parsing out the fields into properties (but directOutput is
nice). This would mean that X3DOM has to route to and from the script
fields (is this possible?).
>>
>>
>>
>> John
>>
>> Sent from Mail for Windows 10
>>
>>
>>
>> From: John Carlson
>> Sent: Saturday, August 4, 2018 4:56 AM
>> To: vmarchetti at kshell.com; X3D-Public; x3dom mlist; Andreas Plesch
>> Subject: RE: [x3d-public] good example for ECMAScripting and Protos?
>>
>>
>>
>> I have a problem with this file in X3DOM, because as far as I can tell,
the SFNode fields (the node, not the field) do not have getFieldValue for
point etc. as a function, for example (but one can double check me). In
other words, I don’t think this node is an X3DOM node.
>>
>>
>>
>> Thus I cannot take a reasonable length in the script. One would have to
parse the string.
>>
>>
>> That’s as clear as I can get. I think this might be because x3dom does
not process scripts correctly. We would need a script tag handler, and a
field handler inside that.
>>
>>
>>
>> Can you help Andreas? I’ve created a script tag before, but that code
is lost. It might be in my GitHub repository somewhere, not sure.
>>
>>
>>
>> We just need a script tag that has fields, but doesn’t execute its CDATA
section. I am pretty sure
>>
>>
>>
>> How is V4.0 handling this?
>>
>>
>>
>> Alternatively, I can change my nodeUtil code to look at the type and do
the right thing converting a string to the correct type. Suggestions are
welcome.
>>
>>
>>
>> Thanks for the great example, Vince,
>>
>>
>>
>> Should I adapt my code to deal with this failing in X3DOM, or should we
change X3DOM?
>>
>>
>>
>> John
>>
>>
>>
>> Sent from Mail for Windows 10
>>
>>
>>
>> From: vmarchetti at kshell.com
>> Sent: Wednesday, August 1, 2018 9:04 PM
>> To: John Carlson; X3D-Public
>> Subject: Re: [x3d-public] good example for ECMAScripting and Protos?
>>
>>
>>
>> See
>>
>>
http://www.kshell.com/pages/pointcloudvisualization/SphereDirectedPointSet.x3d
>>
>>
>>
>> It is a X3D file with with a Prototype + ecmascript definition of a
point cloud, with a vector attached to each point of the cloud.
>>
>> Example used is just points randomly distributed on sphere with directs
pointed radially outward.
>>
>>
>>
>> Potential uses would be to implement the scanning design pattern at
http://x3dgraphics.com/examples/X3dForAdvancedModeling/Scanning/X3dMeshDesignPatternIndex.html
, or
>>
>> visualizing fluid flow or electromagnetic field
>>
>>
>>
>> On Jul 31, 2018, at 10:30 PM, John Carlson <yottzumm at gmail.com> wrote:
>>
>>
>>
>> Is there a good example of ECMAScripting in X3D that I can use to test
my X3D JSON ECMAScript preprocessor? Preferably with Protos
>>
>>
>>
>>
>>
>> Thanks!
>>
>>
>>
>> John
>>
>>
>>
>> _______________________________________________
>> x3d-public mailing list
>> x3d-public at web3d.org
>> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> 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 Past Chair
>> President, Daly Realism - Creating the Future
>
>
> --
> Leonard Daly
> 3D Systems & Cloud Consultant
> LA ACM SIGGRAPH Past Chair
> President, Daly Realism - Creating the Future
--
Andreas Plesch
Waltham, MA 02453
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20180806/c743b710/attachment-0001.html>
More information about the x3d-public
mailing list