Thanks for the kind words!

1. Use a ray between the camera and the anchor point to establish any obstacles/meshes that would render the anchored element invisible. Possible check box in properties for "Hidden behind meshes"

Doesn't Detect Visibility flag in Anchor's settings solve this problem?

2. Use css3 3d transforms to match visual perspective and skewing based on the anchors object x or y axis and the cameras possition to simulate things like moving to the side of the html element. Possible applications would be such as simulating html content on a 3d wall. Possible check box in properties for "Lock vector space"

Thanks for the suggestion. We'll think of something like this.

WebGL does not restrict us to use only one canvas per page.
It this short article it is shown how you can run two Blend4Web applications and even build simple interaction between them.

Link to Article

Project sources

WebGL не ограничивает нас лишь одним canvas на странице.
В этой небольшой статье показано как можно запустить сразу два приложения Blend4Web и даже построить несложное взаимодействие между ними.

Ссылка на статью

Исходники проекта

Вообще говоря, с этими багами мы боролись довольно давно, и насколько я знаю, они больше не должны проявляться. Если очистить браузерный кэш, это поведение воспроизводится?

Спасибо за репорт. Поправим и добавим тесты для подобных функций.

Сам метод должен выводить имена и соотв. им объекты, попадающие в область видимости основной камеры. Если убрать строку с GLOW из массива, то можно увидеть, как это должно работать.

Actually, there are no plans to support built-in mechanisms for AR in Blend4Web as for now and a primary reason for this is that there are already solutions out there which can be implemented with Blend4Web with a bit of effort. I'm speaking about awe.js for example.

We'll consider making some public examples with it in the near future.

Hi Brett,

The thing is that bootstrap modal resizes the canvas_container to 0x0 dimensions at the start. This leads to Blend4Web thinking that the canvas is completely invisible to the user because clientWidth and clientHeight properties equal zero. The engine tries to protect itself and resizes to 320x240. Also, this protects the canvas from bad camera aspect ratio values.

We'll reconsider this behavior to the next release.

As for now, I can recommend using something like this in the loaded callback:

Hi,

That's true about the mouse pointer. It is not possible to override the mouse position with Javascript.
But there is no need in this

Just use the pick_center method to get an object in the center of the screen.

По такому специфичному функционалу как якори возникало очень много вопросов. Мы решили, что пришло время написать статью, призванную ответить на большинство из них. Приятного чтения!

Ссылка на статью

Исходники проекта

There were tons of questions regarding anchors/annotations work. So we decided to write an article which aims to answer most of them. Enjoy reading!

Link to the article

Project source files