FreePV

The control works well for me

Windows XPpro, service pack1, Internet Explorer 6.

Bill

Fulvio Senore wrote:

> Please let me know if the control is working well or not.
>
worked well here, w2k i6sp1 all updates
Yuv

Hello Fulvio,

on Thursday, October 20, 2005, 12:30:27 you wrote:

FS> Hello to everybody,

FS> I have been rather busy lately, but I have been able to build  a test 
FS> ActiveX control.
FS> The control does a very simple task: it copies a file from a remote URL
FS> to a local file in your computer. Nothing exceptional, but my first goal
FS> was creating a working control and being able to download data from the
FS> server.

Works fine, but the file is also kept in the Browser cache, so just
deleting the file, doesn't force a reload. Did you use
'InternetReadFile'? From the filesize I assume, that you have used
MFC? Just for your Info: You do not need to copy the OCX to the
system32 directory. Any directory will do the trick.

MfG, 
     Thomas.

Hello Thomas,

Thomas Rauscher ha scritto:

>Hello Fulvio,
>
>on Thursday, October 20, 2005, 12:30:27 you wrote:
>
>FS> Hello to everybody,
>
>FS> I have been rather busy lately, but I have been able to build  a test 
>FS> ActiveX control.
>FS> The control does a very simple task: it copies a file from a remote URL
>FS> to a local file in your computer. Nothing exceptional, but my first goal
>FS> was creating a working control and being able to download data from the
>FS> server.
>
>Works fine, but the file is also kept in the Browser cache, so just
>deleting the file, doesn't force a reload. Did you use
>'InternetReadFile'? From the filesize I assume, that you have used
>MFC? Just for your Info: You do not need to copy the OCX to the
>system32 directory. Any directory will do the trick.
>
>MfG, 
>     Thomas.
>
I used an AsyncMonikerFile MFC class. It is very useful because it 
automatically calls a callback function at intervals to give feedback. 
The file is loaded asyncronously so the control is not locked.
The class automatically uses the IE cache: I suppose that it calls 
InternetReadFile in some way.

I know that it is possible to copy the ocx anywhere, but then testers 
should write the path while calling regsvr32 and this is annoying and 
error-prone.

Fulvio

Yes, we invited him but he did not want to join.

It's a pity, but I can understand his reasons.

Fulvio Senore


Erik Krause ha scritto:

>Hello,
>
>as it seems DevalVR is growing better and better. There was a short 
>discussion about whether to invite Fiero. Did someone? 
>
>best regards
>--
>Erik Krause
>Offenburger Str. 33
>79108 Freiburg
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by:
>Power Architecture Resource Center: Free content, downloads, discussions,
>and more. http://solutions.newsforge.com/ibmarch.tmpl
>_______________________________________________
>Freepv-devel mailing list
>Freepv-devel@...
>https://lists.sourceforge.net/lists/listinfo/freepv-devel
>
>
>  
>

Am Friday, October 21, 2005 um 12:29 schrieb Fulvio Senore:

> Yes, we invited him but he did not want to join.
> 
> It's a pity, but I can understand his reasons.

Can you tell the reasons?

best regards
--
Erik Krause
Offenburger Str. 33
79108 Freiburg

He is selling a pay-per use software...
:-D




----- Original Message ----- 
From: "Erik Krause" <erik.krause@...>
To: <freepv-devel@...>
Sent: Friday, October 21, 2005 2:04 PM
Subject: [FreePV-dev] Re: Anyone invited Fiero Deval?


> Am Friday, October 21, 2005 um 12:29 schrieb Fulvio Senore:
> 
>> Yes, we invited him but he did not want to join.
>> 
>> It's a pity, but I can understand his reasons.
> 
> Can you tell the reasons?
> 
> best regards
> --
> Erik Krause
> Offenburger Str. 33
> 79108 Freiburg
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Freepv-devel mailing list
> Freepv-devel@...
> https://lists.sourceforge.net/lists/listinfo/freepv-devel

But I tried again.
Our goal should be to create a wide standard markup language.
The result is that you see the panorama whatever plugin you have 
installed...

----- Original Message ----- 
From: "Luca N Vascon" <vascon@...>
To: <freepv-devel@...>
Sent: Friday, October 21, 2005 3:01 PM
Subject: Re: [FreePV-dev] Re: Anyone invited Fiero Deval?


> He is selling a pay-per use software...
> :-D
>
>
>
>
> ----- Original Message ----- 
> From: "Erik Krause" <erik.krause@...>
> To: <freepv-devel@...>
> Sent: Friday, October 21, 2005 2:04 PM
> Subject: [FreePV-dev] Re: Anyone invited Fiero Deval?
>
>
>> Am Friday, October 21, 2005 um 12:29 schrieb Fulvio Senore:
>>
>>> Yes, we invited him but he did not want to join.
>>>
>>> It's a pity, but I can understand his reasons.
>>
>> Can you tell the reasons?
>>
>> best regards
>> --
>> Erik Krause
>> Offenburger Str. 33
>> 79108 Freiburg
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by:
>> Power Architecture Resource Center: Free content, downloads, discussions,
>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>> _______________________________________________
>> Freepv-devel mailing list
>> Freepv-devel@...
>> https://lists.sourceforge.net/lists/listinfo/freepv-devel
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Freepv-devel mailing list
> Freepv-devel@...
> https://lists.sourceforge.net/lists/listinfo/freepv-devel

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
Erik Krause ha scritto:
<blockquote cite="mid4358F580.20768.C7E165@..." type="cite">
  <pre wrap="">Am Friday, October 21, 2005 um 12:29 schrieb Fulvio Senore:

  </pre>
  <blockquote type="cite">
    <pre wrap="">Yes, we invited him but he did not want to join.

It's a pity, but I can understand his reasons.
    </pre>
  </blockquote>
  <pre wrap=""><!---->
Can you tell the reasons?
  </pre>
</blockquote>
He said that he has worked for 4 years to develop his software and that
he wants to continue developing it.<br>
He has not time for other projects.<br>
<br>
Filvio Senore<br>
</body>
</html>

Erik Krause wrote:

>Hello,
>
>as it seems DevalVR is growing better and better. There was a short 
>discussion about whether to invite Fiero. Did someone? 
>  
>
I suggested inviting him but did not want to give the impression of 
"taking control" ;-)

Maybe we should share a bit more what we are doing and keep each other 
updated?

After understanding (a bit late) that a small and loud minority was not 
ready to discuss licensing I focused on browser capabilities detection. 
I discussed on a one to one basis with a few people on this list and 
came to the conclusion that there are still much too different views 
about the goal of this project. I believe it would be appropriate to 
share our individual visions and goals and to work at forging a common 
vision - a mix of the individual visions and a compromise on 
invidividual asprations. I believe that if this project is to come out 
of the closet, it has to develop a strong team with a common vision and 
a set of shared goals and values. I would see the goals as a list of 
spelled out milestones/achievements that we want to reach together, in 
sequential order, with no time pressure. On some of them some of us will 
be instrumental and on some others other of us, like in a soccer team 
where the defense player is instrumental when the ball is near the own 
goal and the attack player is instrumental when the ball is near the 
adversary's goal. A commen vision and values drive these individual 
player to become a team and pass each other the ball when it is the 
right time. I believe freepv needs something similar if it is to 
transcend the indivdual and become a real tam project.

I know that for some technical minded people on this list, goals are 
things like:
- the best quality panorama 3d engine
- the most perfectioned plug in available on most platforms
- the smallest possible foot print
- learning and mastering the tools and skills needed to produce all of that

for commercially minded people on this list, goals are things like:
- getting out a common standard to describe a panorama so that they can 
focus on producing and selling the authoring tools for it

for marketing minded people on this list, goals are things like:
- getting rid of the obstacle between their content and their target 
audience
- reaching the widest possible audience / beat QuickTime and Java in 
terms of popularity
- making panoramas a common household concept

these people are driven by different values:
- technical excellence and knowledge discovery drives the t-shirts 
(technical guys)
- generating value and make money in the process drives the ties 
(commercially minded people)
- fame and success drive the turtlenecks (marketing minded people)

i'd like the individuals that have been lurking along here to come out 
and share their visions, their values, their goals.

My vision is of panoramas becoming something like a mouse. Tell a human 
being twenty years ago that he'd drag a mouse and they would think 
you're nuts. Now dragging a mouse is a concept as common as drinking 
water. To people that are confronted with panoramas for the first time 
the experience is not much different than for a teenage that drives a 
car for the first time. With the difference that all cars across brands 
and makes have standardized user interface, drive on all roads and are 
more or less stable, while panoramas require different complex 
technologies that interacts in strange ways and more often than not does 
not work as it should on their computers.

I think that establishing panoramas as a common concept can generate a 
lot of value for the public and for the community of panorama makers. 
There will be plenty of business opportunities. Ten years ago, if you'd 
say to a hotel that it needs a website it would think you are crazy. 
Today there is no business that does not have one. Panoramas are 
somewhere in between and in a few years there will be no hotel without a 
panorama. Opportunities abound for hadware maker, software makers and 
photographer. The viewer will not be a source of revenue but it will 
enable other sources of revenue.

My goal is for freepv to become more popular than any other competing 
panorama viewing technology. I understand that this can not happen from 
today to tomorrow, but I am confident that if tools like Skype have made 
it from 1 to 100 million users in a year or so, a well driven freepv 
project can as well. It is a long step until then and the first step is 
to develop a proof of concept, then to take it step by step until there 
is a marketable product. continue with product development and 
improvement while already working at the market, as opposed to wait too 
long to come out with the big bang super dooper perfect solution at 
release. Release often and release in small increments, while keeping an 
eye on backward compatibilty (or better: while planning for future 
features in advance).

That's me. Anbody else is ready to come out?
Yuv

>* panorama
>
>Subclasses: 
>
>-- cubic
>-- cylindrical
>-- equirectlinear
>  
>
In SPi-V, a panoramic scene can consist of many panoramic elements, each 
can be of different subclasses. This is IMHO a very powerful feature:
* tiles (mov) and cubic faces can be implemented using multiple 'flat' 
panoelement
* anything can be a hotspot
* blending from pano to pano
* cap image
* (animated) inserts

So I'ld propose a 'scene' class above what you call the panorama class 
(and rename that to something like 'panoramic element').

Two more subclasses for the 'panoramic element':
-- flat
-- fisheye (  ;-D )

>* controller
>
>Platform independent class to manage user interaction, like
>turn, tilt, ... goto next node,...
>contains [current] X [pan,tilt,fov]
>
>listens to browser events
>provide events
>  
>
Possible subclasses:
- keyboard (arrow keys + shift/control)
- pointer (click'n'drag + right click'n'drag)
- external (toucscreen, vrglasses, ...)

We could also implement scripted movements as a controller, but there 
might be a better way.

>* Render
>abstract baseclass. Platform dependent.
>Displays the image on the screen
>  
>
A render surface description would be nice... IE: render to different 
surface geometry configurations. Normal projection would be to a flat 
surface, but rendering to a curved surface would be interesting for dome 
projection, and for wide fov viewing (as we often do), rectilinear is 
not always the best projection.

'do

Hello Aldo,

on Monday, October 3, 2005, 11:16:39 you wrote:

>>* panorama
>>
>>Subclasses: 
>>
>>-- cubic
>>-- cylindrical
>>-- equirectlinear
>>  
>>
AH> In SPi-V, a panoramic scene can consist of many panoramic elements, each
AH> can be of different subclasses. This is IMHO a very powerful feature:
AH> * tiles (mov) and cubic faces can be implemented using multiple 'flat'
AH> panoelement
AH> * anything can be a hotspot
AH> * blending from pano to pano
AH> * cap image
AH> * (animated) inserts

AH> So I'ld propose a 'scene' class above what you call the panorama class
AH> (and rename that to something like 'panoramic element').

Very good idea! Maybe scene element?

AH> Two more subclasses for the 'panoramic element':
AH> -- flat
AH> -- fisheye (  ;-D )

Uhhhh... I hear iPix coming. Let say 'sphere' or 'quadric surface'
Maybe some day we can add Klein bottles ;-)

>>* controller
>>
>>Platform independent class to manage user interaction, like
>>turn, tilt, ... goto next node,...
>>contains [current] X [pan,tilt,fov]
>>
>>listens to browser events
>>provide events

AH> Possible subclasses:
AH> - keyboard (arrow keys + shift/control)
AH> - pointer (click'n'drag + right click'n'drag)
AH> - external (toucscreen, vrglasses, ...)

Good idea!

Also Joystick, 3D Trackers... oohhh I start dreaming....

AH> We could also implement scripted movements as a controller, but there 
AH> might be a better way.

I think, this is a property of the scene, so we should add this as
(non visual) scene element.

Also possible scene elements could bee auto rotator (=very simple path)
lens flairs, compass, ....

And if a scene element can interact with the controller we can embed
Flash as "Wired actions" but this is just to show the power of this concept,
not a real plan for the near future.

>>* Render
>>abstract baseclass. Platform dependent.
>>Displays the image on the screen
>>  
>>
AH> A render surface description would be nice... IE: render to different 
AH> surface geometry configurations. Normal projection would be to a flat 
AH> surface, but rendering to a curved surface would be interesting for dome
AH> projection, and for wide fov viewing (as we often do), rectilinear is 
AH> not always the best projection.

Good point. Also multiple screens with different viewing directions
should be possible, like a CAVE.

Hey I think, we are on the right track.

Do someone know a good and simple multi platform, vector - graphics
program, to draw this class structure. Because a word can only say less
than 10^-3 of an image.

MfG,
     Thomas.

Hello,

I am currently learning how to write an ActiveX control and a plugin in C++.

I have still a lot of things to learn, but here are the first thoughts 
about class structure.

We will need both an activex and a plugin version of the viewer: this 
will require some different code for each version. We will have to keep 
the different code at a minimum to avoid duplication.

Downloading of data from the server is done at the plugin/activex level 
(doing it in the viewer core would be much more complicated), so we will 
have to handle some communication between the viewer core and the 
plugin/activex code.

I would implement the controller as an abstract class that defines the 
required interface. It would be easier to create different controllers 
that inherit from that class in a second time.

As soon as I will be able to create a very simple test plugin it will be 
time to try porting it from Windows to the other platforms: this will 
require some volunteers. The task will be rather simple:
- install the Gecko SDK to be able to build the plugin
- write a few lines of native code to show something on the screen.
This will not have anything to do with the viewer, it will be a test to 
see if we are able to build a plugin on the various platforms. Of course 
the build environment will be used for the plugin too.

Fulvio


Thomas Rauscher ha scritto:

>Hello,
>
>as I said, I would like to work more the "Extreme Programming" way,
>not in a "Waterfall" model, so I would like to start with some
>design work. We all know that what we want to do, can be done (look at
>Quicktime!) so the open question for the plugins are just
>implementation details.
>
>What I want to start with is a discussion of the basic object model.
>If this is clear, I would suggest we can start implementing a
>standalone viewer prototype (because it's much easier to debug) and
>implant this code into the different platform specific plugins.
>
>In my opinion, the rough class structure should look like this:
>
>* panorama
>
>abstract base class, platform independent
>
>These objects should hold all
>the image data and viewing parameters of one panoramic image (node)
>for example [min,max,default] X [pan,tilt,fov]
>HDR, Hotspots,...
>
>Subclasses: 
>
>-- cubic
>6 cubefaces like QTVR
>
>-- cylindrical
>One image in cylindrical projection
>
>-- equirectlinear
>One image in equirectlinear projection
>....
>
>
>* controller
>
>Platform independent class to manage user interaction, like
>turn, tilt, ... goto next node,...
>contains [current] X [pan,tilt,fov]
>
>listens to browser events
>provide events
>
>* Render
>abstract baseclass. Platform dependent.
>Displays the image on the screen
>
>Subclasses: 
>-- Legacy
>simple draw pixel method
>
>-- Direct2D
>draw pixel method for older Computers/to less VRAM
>
>-- DirectX
>Win32
>
>-- OpenGLBase
>Win32, Linux, MacOS,....
>
>-- OpenGLEx
>Use Hardware accelerated HDR display 
>
>....
>
>* loader
>abstract baseclass, platform independent
>
>Decodes input files to panorama objects
>
>Subclasses: 
>-- Quicktime VR
>
>-- PTViewer
>
>-- SPi-V
>
>.....
>
>
>All these classes are packed into a small platform dependent
>plugin/stand alone player, which handles the low level input/output
>and distributes it to the classes, and listens to events of the
>controller/render class.
>
>The code structure/sharing will be more complex, because some things
>can be shared for the platform (Win32: activeX/Netscape/Standalone,
>...) and some things for the browser (netscape: Win32/MacOS/Linux,...)
>but for this I would need to know more about all this things.
>
>But I think these layers are very small, so we can start with some
>prototypes for each platform, and than look, whats simmilar/reuseable.
>
>What do you think about these ideas?
>
>MfG, 
>     Thomas.
>
>
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by:
>Power Architecture Resource Center: Free content, downloads, discussions,
>and more. http://solutions.newsforge.com/ibmarch.tmpl
>_______________________________________________
>Freepv-devel mailing list
>Freepv-devel@...
>https://lists.sourceforge.net/lists/listinfo/freepv-devel
>
>
>  
>

Thomas Rauscher wrote:

>>>* controller
>>>      
>>>
>AH> We could also implement scripted movements as a controller, but there 
>AH> might be a better way.
>
>I think, this is a property of the scene, so we should add this as
>(non visual) scene element.
>  
>

scripted movements and user controlled interactions look very similar to 
me. i'd propose making them subclasses of the same class. The only 
differences I see is who controls the movement (the script or the user?) 
and through what interface (directly or through a keyboard, mouse, 
joystick, whatever?). This will add an interface definition with its own 
code (i.e. to read the hardware and/or to validate the input).

Immervision has a camera and an autopath node in their xml description 
so that one panorama can have many cameras and (theoretically) many 
autopaths. Imagine switching between active interaction and passive 
viewing and while in passive viewing switching through different, time 
synchronized cameras...

>Do someone know a good and simple multi platform, vector - graphics
>program, to draw this class structure. Because a word can only say less
>than 10^-3 of an image.
>  
>
A generic vector graphics program I've been very happy with over the 
years is Canvas, from a Florida based company called Deneba that merged 
with ACD, maker of the world famous ACDSee image browser. I used it on 
the Mac back in 1989 and the files are still fully readable on the 
newest version, available for both PC and Mac. They are currently at 
v10. I've tried the demos but for my needs (very simple vector based 
graphics to be integrated into office presentations) the old v7 is enough.

Yuv

Luca N Vascon wrote:

> Interactions
>
> It should be able to interact someway with flash, html java applets 
> etc. to gather more interactivity...
> I think about compass, addictional informations etc.

Immervision has a Compass node in the GUI. this gives a quick solution 
for the programming-challenged publisher. The applet should make use of 
LiveConnect! and similar technology to publish the actual FOV, actual 
pan, tilt and rotation. Yes, the controller should be able to rotate, I 
wish to create a star wars like effect, with a jet plane taking a steep 
curve while an animated pano of moving clouds or starts is projected 
around me.

A compass integrated in the panorama could be projected it in 
perspective rather than flat on gui, like a real compass in the hand of 
the explorer walking through the panorama.

> I dream of a viewer able to alpha-fade 2 panoramas or more, able to 
> adjust light, to show "before and after", to  shoe overimpressed 
> informations, to act as a navigator of the whole site and all these 
> things....

... and I wish for a general alpha channel of the applet to have the 
freedom to show what is beneath (i.e. the webpage), like in Flash (and 
unlike Shockwave, Java, QuickTime).

Writing of Flash, has anybody already played with the new 3D functions 
embedded in Flash?

Object: <http://www.fenomas.com/flash8/>
Cylinder: <http://www.flashclub.nl/flash8/flashVR/>

Yuv

If it wouldn´t be too late or make the basic plugin too big:

Is it possible to have a motion movie class for transition between panoramas 
and for *localized events*_?

Regards
Jacek
 http://www.eu3d.com
----- Original Message ----- 
From: "Thomas Rauscher" <freepv@...>
To: <freepv-devel@...>
Sent: Sunday, October 02, 2005 7:21 PM
Subject: [FreePV-dev] Class structure


> Hello,
>
> as I said, I would like to work more the "Extreme Programming" way,
> not in a "Waterfall" model, so I would like to start with some
> design work. We all know that what we want to do, can be done (look at
> Quicktime!) so the open question for the plugins are just
> implementation details.
>
> What I want to start with is a discussion of the basic object model.
> If this is clear, I would suggest we can start implementing a
> standalone viewer prototype (because it's much easier to debug) and
> implant this code into the different platform specific plugins.
>
> In my opinion, the rough class structure should look like this:
>
> * panorama
>
> abstract base class, platform independent
>
> These objects should hold all
> the image data and viewing parameters of one panoramic image (node)
> for example [min,max,default] X [pan,tilt,fov]
> HDR, Hotspots,...
>
> Subclasses:
>
> -- cubic
> 6 cubefaces like QTVR
>
> -- cylindrical
> One image in cylindrical projection
>
> -- equirectlinear
> One image in equirectlinear projection
> ....
>
>
> * controller
>
> Platform independent class to manage user interaction, like
> turn, tilt, ... goto next node,...
> contains [current] X [pan,tilt,fov]
>
> listens to browser events
> provide events
>
> * Render
> abstract baseclass. Platform dependent.
> Displays the image on the screen
>
> Subclasses:
> -- Legacy
> simple draw pixel method
>
> -- Direct2D
> draw pixel method for older Computers/to less VRAM
>
> -- DirectX
> Win32
>
> -- OpenGLBase
> Win32, Linux, MacOS,....
>
> -- OpenGLEx
> Use Hardware accelerated HDR display
>
> ....
>
> * loader
> abstract baseclass, platform independent
>
> Decodes input files to panorama objects
>
> Subclasses:
> -- Quicktime VR
>
> -- PTViewer
>
> -- SPi-V
>
> .....
>
>
> All these classes are packed into a small platform dependent
> plugin/stand alone player, which handles the low level input/output
> and distributes it to the classes, and listens to events of the
> controller/render class.
>
> The code structure/sharing will be more complex, because some things
> can be shared for the platform (Win32: activeX/Netscape/Standalone,
> ...) and some things for the browser (netscape: Win32/MacOS/Linux,...)
> but for this I would need to know more about all this things.
>
> But I think these layers are very small, so we can start with some
> prototypes for each platform, and than look, whats simmilar/reuseable.
>
> What do you think about these ideas?
>
> MfG,
>     Thomas.
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Freepv-devel mailing list
> Freepv-devel@...
> https://lists.sourceforge.net/lists/listinfo/freepv-devel
>
>
>

Hello everybody,

as I said, I would like to work more the "Extreme Programming" way,
not in a "Waterfall" model, so I would like to start with some
design work. We all know that what we want to do, can be done (look at
Quicktime!) so the open question for the plugins are just
implementation details.

What I want to start with is a discussion of the basic object model.
If this is clear, I would suggest we can start implementing a
standalone viewer prototype (because it's much easier to debug) and
implant this code into the different platform specific plugins.

In my opinion, the rough class structure should look like this:

* panorama

abstract base class, platform independent

These objects should hold all
the image data and viewing parameters of one panoramic image (node)
for example [min,max,default] X [pan,tilt,fov]
HDR, Hotspots,...

Subclasses: 

-- cubic
6 cubefaces like QTVR

-- cylindrical
One image in cylindrical projection

-- equirectlinear
One image in equirectlinear projection
....


* controller

Platform independent class to manage user interaction, like
turn, tilt, ... goto next node,...
contains [current] X [pan,tilt,fov]

listens to browser events
provide events

* Render
abstract baseclass. Platform dependent.
Displays the image on the screen

Subclasses: 
-- Legacy
simple draw pixel method

-- Direct2D
draw pixel method for older Computers/to less VRAM

-- DirectX
Win32

-- OpenGLBase
Win32, Linux, MacOS,....

-- OpenGLEx
Use Hardware accelerated HDR display 

....

* loader
abstract baseclass, platform independent

Decodes input files to panorama objects

Subclasses: 
-- Quicktime VR

-- PTViewer

-- SPi-V

.....


All these classes are packed into a small platform dependent
plugin/stand alone player, which handles the low level input/output
and distributes it to the classes, and listens to events of the
controller/render class.

The code structure/sharing will be more complex, because some things
can be shared for the platform (Win32: activeX/Netscape/Standalone,
...) and some things for the browser (netscape: Win32/MacOS/Linux,...)
but for this I would need to know more about all this things.

But I think these layers are very small, so we can start with some
prototypes for each platform, and than look, whats simmilar/reuseable.

What do you think about these ideas?

MfG, 
     Thomas.

Hello Jacek,

may you explain with more detail what you are asking for?

Fulvio Senore

Lists@... ha scritto:

> I  am sorry, my former email did not reach the list:
>
>>
>> Is it possible to have a motion movie class for transition between 
>> panoramas and for *localized events* ?
>>
>> Regards
>> Jacek
>> http://www.eu3d.com 
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Freepv-devel mailing list
> Freepv-devel@...
> https://lists.sourceforge.net/lists/listinfo/freepv-devel
>
>