RE: [cgmo-webcgm] Flow of electricity

[Lofton Henderson <lofton@rockynet.com>]

Benoit et al --

I have been reading and pondering all of this thread.  I'll dump some 
cumulative thoughts in reply to Benoit's opinion piece...

At 10:50 AM 10/2/2007 -0400, Bezaire, Benoit wrote:
>Hi Stuart,
>
>That's a nice little animation. Clever!
>
>If I remember correctly, we do have a remove style... you have to do:
>setStyleProperty( prop, "inherit" );
>
>I have a felling I'm not about to make any friends with what I'm about
>to say; but I have to say it since I think it matters.
>
>Although the result Stuart generated is good looking (which is a great
>positive), I would not encourage users to do this. I would argue that it
>takes too long to author such an animation. Equivalent results can be
>achieve faster in other file formats.

Hmmm, that sounds to me like, "...instead producing some JavaScript, you 
should convert your graphics to another format and don't use WebCGM".  Some 
people (Molly/Larry at least, and maybe only them) seem to buck this advice.

Stuart's code is similar in effect to one of the Boeing examples (demo'd by 
M&L at Seattle F2F).  But its implementation is *much* simpler than how 
they are now achieving the same effect with CGM.  M&L have apparently 
rejected other formats like Flash.

I agree with what several people have said in this thread -- whatever the 
nature of any further potential animation support in WebCGM, we do not 
imagine many people hand-coding these things.  (What do you think this is, 
SVG?  ;-)  )

Even the current (CGM) Boeing animations, which are much more complex 
structurally, are produced by tools and experts, not by end users.


>If we want animations to be frequently used in CGM we need easier
>methods than what is shown in this example (I'm mostly referring to the
>flow aspect). The electrical path was broken down into multiple
>fragments, then each were assigned a specific name/id. I would argue
>that most revisions of this illustration breaks the animation.

Well, Stuart's animation is manipulating the visible styles and attributes 
of named objects.  Therefore, would it be any worse than XCF-based 
declarative animation (an external file that attaches animation 
declarations to named objects)?

>Additionally, I suspect that customers wiring diagrams are 20x more
>complex; thus, authoring animation must be much easier.

Point of information... The real M&L sample was indeed based on a much more 
complex diagram.  But they were still step-animating just a handful of 
circuit pieces (line segments) in any one animation sequence.


>Attached is an example of what I mean (in SVG). Displayed are two (out
>of many) ways of showing flow, both used at the same time. In this
>example, if the wire (say during a revision) goes from a rectangle to a
>some sort of polygon, or if the battery/switch are moved to a new
>location... the animation still work.

Two comments.  First, reading the SVG code, with smooth motion animation 
and smooth attribute animation it already exceeds the baseline requirements 
we have heard -- step animation -- from at least two sources so far.

Second, that level of immunity to editing ("goes from rectangle to 
polygon") is provided in your SVG example by indirect reference to a named 
path within an animateMotion element (which is an SVG-specific extension to 
SMIL 2.1, correct?).  Am I correct that this level of sophistication is 
beyond what you support in your private WebCGM animation subset for 
IsoDraw/IsoView?

I haven't thought it through completely, but I'm unsure that you could 
achieve that in WebCGM without embedding the animation elements directly 
within the WebCGM instance.


>Again, I'm not opposed to animation in CGM, but I very much favor a
>declarative approach than a DOM approach.
>
>Thoughts?

I don't really disagree with anything you have said.  I think our difficult 
task is to decide where we should set our sights, in the spectrum from ZERO 
to full-SVG/SMIL-like.  There are levels of possibilities:

a.) nothing more in 2.1:  people will continue to do DOM animation with 1.0 
and 2.0, as M&L and as Stuart.
b.) a little more help in 2.1 with a couple things that would simplify DOM 
animation
c.) a simple XCF-based set of animation declaration elements (in addition 
or instead of #b).
...
d.) like circle.svg -- intra-WebCGM embedded animation declaration 
elements, enabling much of the SMIL/SVG-like functional capability.

We can't go to #d in a quick 2.1, and perhaps never -- not only is there 
the risk that the vendors might not follow, but I'm not sure that I can 
envision this stuff intra-WebCGM, i.e., animation elements within the 
metafile itself (like circle.svg).

Whereas I can see "animate a picture" within the scope of WebCGM, I'm not 
yet sure I can see "animated picture".  (If you know what I mean.)

All for now,
-Lofton.


>-----Original Message-----
>From: Lofton Henderson [mailto:lofton@rockynet.com]
>Sent: Monday, October 01, 2007 6:10 PM
>To: Galt, Stuart A; cgmo-webcgm@lists.oasis-open.org
>Subject: Re: [cgmo-webcgm] Flow of electricity
>
>Stuart --
>
>Good stuff!  Again, we should be looking to put your examples (and
>anyone
>else's) of 2.0 capabilities somewhere on the Web site.  So, aside from
>technical (and 2.1 rqts questions), there is the Education/Outreach
>question:  when and where can we put these up and direct people to them?
>
>Technical questions...
>
>At 01:46 PM 9/28/2007 -0700, Galt, Stuart A wrote:
> >[...]
> >As I start making more complex images/applications I am finding myself
> >wishing I had a either getStyleProperty() or removeStyleProperty()
>
>Is the "get" an inquiry function?  Do you actually want both "get" and
>"remove", or did you really mean "either...or"?
>
>Do you propose that these be added to the 2.1 wish list?
>
> >method
> >(I realize that this is difficult to implement) and the ability to make
>
> >several DOM changes without waiting for a redraw between each one.
>
>I guess this has now been added to the 2.1 wish list.
>
>-Lofton.
>
>