OpenVSP Airshow is Live!

[Rob McDonald]

Sure to be a crowd pleaser, I am thrilled to announce the new OpenVSP Airshow!

The Airshow is the place to go to show off your OpenVSP models.  It is an entirely re-written replacement for the now defunct OpenVSP Hangar.

Many thanks to Daniel Enriquez, who took this project on without telling anyone and later emerged with a nearly finished product.

All of the v3 models on the Hangar have been ported over to the Airshow.

With more than a decade of Hangar operation behind us, we've dropped a few features that never really panned out.

Development of the Airshow will continue.  Please make any suggestions on the Google Group.

[deenri...@gmail.com]

Thanks Rob! I am excited to hear how the community receives this update. 
I have worked to make the site very responsive and fast. 

I plan on adding more features soon. Of particular note is the automated X3D generation during new model upload. Currently users will have to manually create and upload this file in addition to the vsp3 model. 

Looking forward to all the feedback!

- Daniel Enriquez

[Brandon Litherland]

It's been brought up that the Search isn't using multi-word arguments.  For example "NASA UAM" doesn't find the two UAM reference vehicles.

[deenri...@gmail.com]

Pushed a fix. muli-word search should now work.

[Tim Swait]

Thanks for doing this! However, I've just tried to upload a model, I filled in all the boxes, but when I click the submit  button, this briefly flicks up and then it reloads the blank form I've filled in every box (including selecting meters for units).

[deenri...@gmail.com]

thanks for reporting this. I believe the form needs you to click on of the unit options in order to work correctly. I will look into a fix for this. 

[Tim Swait]

I had clicked on meters, but when I click 'Submit' then what's shown in the screenshot appears for about 1 second but before you can get the mouse there to click anything it has reloaded the page with a blank form. This is using Firefox on Ubuntu.

[deenri...@gmail.com]

Yup, I see the bug. The form is not working. Working on a fix. Will update here when done.

[deenri...@gmail.com]

The upload form should now be working!

[Tim Swait]

Thank you, yes it is. I've uploaded one of my hang gliders to it:
https://airshow.openvsp.org/vsp/hd2mrEwuoqgU8J5x9yIW
However if you scroll down on the model page then it drifts in weird way (checked on Firefox and Chromium, both do it) making it very hard to click on the Download button as it keeps chasing away from under your mouse! It's probably due to the length of text I put into the model description (which was maybe a bit excessive, if you can't fix the scrolling then maybe limit the length of text that can be uploaded by the form).

Also the rendering in the view window has a strange groove across the wing root that's not there in the vsp3 model. I'm not sure if that's to do with the x3d export process or the website rendering.

[Rob McDonald]

The link to this model does not work for me -- perhaps they are not permalinks.

The X3D model does not render for me at this time for this model.

The black line has to do with the X3D export and render -- it would have looked the same way in the old Hangar.  

It is theoretically possible to modify OpenVSP to improve the rendering, but it has not been a priority since X3D export was introduced in about 2012...

Rob

[deenri...@gmail.com]

It was not a specific model issue. It was an issue with how I was loading the x3d.js viewer scripts. Should be fixed!

[deenri...@gmail.com]

I mean to say, the blank model issue is fixed. The black line rendering is an openvsp issue.

[Tim Swait]

Nice! thank you, that works much better now. FYI, the links aren't permalinks, which does make it hard to share a model externally, you can't directly link to it by copying the address in the browser bar (as I did in the previous post).

[deenri...@gmail.com]

I fixed the permalink issue! Models should now be shareable. As a side benefit, the changes will allow all models to be indexed by Google. 

[Tim Swait]

Thanks again for your work on this, it looks really good!

[willj...@gmail.com]

I've spotted that the filters are probably not operating as expected. They seem to be filtering "alphabetically", even for numbers. So for example 10 would come before 2 because 10 starts with a 1.

[deenri...@gmail.com]

Thanks for the feedback!

[willj...@gmail.com]

Another observation if I may - it looks like perhaps the normals are reversed on wings in the preview 3d model. If not that then something else odd with the shading on wing surfaces is occuring.

[Rob McDonald]

The *.x3d files rendered by the website output each component as a single wireframe mesh.  This corresponds to the wireframe shown on-screen in OpenVSP.

A normal vector is calculated at each node point that is used in the rendering process.

At sharp corners (wingtips), the normal vector should be multi-valued (one normal to the wingtip, another to the wing surface).  Since the render doesn't know about the sharp corner, it creates a visual artifact.

In OpenVSP's native rendering, we split the mesh wherever sharp corners occur before calculating the normal vectors and sending it off to the graphics card for rendering.  That way, we are actually visualizing two surfaces (with separate normal vectors) that happen to meet.  This results in better rendering.  That said, there are still some cases where OpenVSP's normal vectors get ugly and our rendering has some artifacts.

Probably the best fix for this would be to modify the *.x3d export capability in OpenVSP to use the split meshes instead of the singular meshes.  I can give you some pointers if you want to try it yourself.

Otherwise, this is extremely low priority and probably won't change.

Rob

--
You received this message because you are subscribed to the Google Groups "OpenVSP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openvsp+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/openvsp/cfa6fa94-07f4-41ab-8c3e-6f004eaea431n%40googlegroups.com.

[willj...@gmail.com]

Rob, appreciate its a low priority thing. I just thought I'd mention it as I hadnt seen it referenced above.