Hi Silvere,
The SceneGraph IO API in Java3D Fly Through does NOT use java.io.serialization,
it stores the data in it's own binary format which is designed for fast
loading of scene graphs. The API supports IO to disk (using RandomAccessFile)
and IO over a network (using Streams).
Java3D 1.3 will include a productised version of this utility as a utility not
part of core. Java3D 1.3 core classes will NOT implement java.io.serializable.
Rgds
Paul
----------------------------------------------------------
Paul Byrne Email : [EMAIL PROTECTED]
Sun Microsystems Phone : (650) 786 9926
Visualization Software Group Fax : (650) 786 5852
----------------------------------------------------------
>Mime-Version: 1.0
>X-Sender: [EMAIL PROTECTED] (Unverified)
>Date: Mon, 23 Apr 2001 19:57:51 +0200
>From: Silvere Martin-Michiellot <[EMAIL PROTECTED]>
>Subject: Re: [JAVA3D] Seralization of 3D objects
>To: [EMAIL PROTECTED]
>
>Hi,
>
>I believe that Scenegraph serialization will be optimized in the next
>releases of Java3D. Remember that the Fly through demo is only beta.
>Everyone needs a reliable efficient network transport of 3D objects.
>I belive people at Sun are making big efforts on their side.
>Nevertheless Justin states that there are much more efficient schemes
>for doing this. Justin please let us know.
>
>I'm currently working on a generic client and server for multiuser
>Java3D. Of course, I want to do minimal coding and benefit of the
>always updated Sun's API. This is why I do not wish to implement my
>own serialization but use Sun's. Justin, you seem to mention this is
>a bad solution. Can you help me understand why or why not ?
>
>
>Silvere Martin.
>
>
>P.S : for those who haven't been, try the J3D CVS code repository.
>Quality of code and usability is awesome.
>
>>Just to back up John's comment... I spent about 20 hours trying to figure
>>out why the texture coordinates of the 2nd texture unit were coming back
>>wrong after I wrote them out correctly. So a piece of geometry would save
>>fine and when loaded the detail texture would be all screwed up. I even had
>>it to the point where I printed out everything that I saved, and then
>>printed it out when it came back in... and darn if it didn't duplicate the
>>texture unit 1 into the texture unit 2 coordinates.
>>
>>Turns out that to "save memory defragmentation" I had created an array of
>>float and then used the same array to get and save the two sets of texture
>>coordinates... but writeObject(floatArray) thought that I was writing out
>>the same object so just put a reference to the first array, even though I
>>had changed the contents...
>>
>>Thats the types of serilization wierdnesses you can get.
>>
>>Dave Yazel
>>
>>-----Original Message-----
>>From: John Wright [mailto:[EMAIL PROTECTED]]
>>Sent: Thursday, April 19, 2001 10:34 AM
>>To: [EMAIL PROTECTED]
>>Subject: Re: [JAVA3D] Seralization of 3D objects
>>
>>
>>Serialization is seriously flawed in Java, the probability of writing
>>malfunctioning code using serialization is way too easy. There are two
>>"bug" reports that Sun states are not bugs yet dozens of comments saying
>>this is not the expected behavior (by the average programmer).
>>
>>Java really has missed the boat for easy data storage. I really miss
>>the good old Modula 2 days when I could just write "readRecord()" and
>>"writeRecord()" to read and write complex data structures.
>>
>>- John Wright
>>Starfire Research
>>
>>Justin Couch wrote:
>> >
>> > Firas MOHAMED wrote:
>> >
>> > > I am trying to save my 3D objects (GeometryInfo, for example) by using
>> > >
>> > > serialization ... But this does not work as these objects are not
>> > > serializable.
>> >
>> > Serialisation of Java classes has to be one of the worst approaches to
>> > data transport and (shared) replication. There are many much more
>> > efficient schemes for doing this. Not only is serialisation horribly
>> > slow, but the chances that it works on more than just your local machine
>> > are even worse. I suggest you look at some other scheme for geometry
>> > persistence or even if you are using them over RMI for shared
>> > representations, there are far, far better ways of doing this.
>> >
>> > --
>> > Justin Couch http://www.vlc.com.au/~justin/
>> > Freelance Java Consultant http://www.yumetech.com/
>> > Author, Java 3D FAQ Maintainer http://www.j3d.org/
>> > -------------------------------------------------------------------
>> > "Humanism is dead. Animals think, feel; so do machines now.
>> > Neither man nor woman is the measure of all things. Every organism
>> > processes data according to its domain, its environment; you, with
>> > all your brains, would be useless in a mouse's universe..."
>> > - Greg Bear, Slant
>> > -------------------------------------------------------------------
>> >
>> >
>>
>
>--
>Silvere MARTIN-MICHIELLOT
>
>
>Construit le futur d'internet sur http://www.digitalbiosphere.com.
>--
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
