nonVAO

NonVAO refers to drawing in OpenGL without using a Vertex Array Object to store vertex attribute state. In traditional or compatibility contexts, developers can rely on the legacy client-side state that predates VAOs, in contrast to the modern approach that centers around VAOs to encapsulate vertex input configuration.

In a nonVAO approach, per-attribute state is set with legacy functions such as glVertexPointer, glNormalPointer, glTexCoordPointer, and glEnableClientState. Vertex data may reside in client memory or in buffers bound to GL_ARRAY_BUFFER; the interpretation of the pointers changes depending on whether a buffer is bound. Since there is no VAO to capture this configuration, an application must reconfigure and rebind the vertex state whenever the data source, shader program, or drawing state changes.

Advantages of nonVAO include simplicity for small demos or tutorials and avoidance of creating VAOs. However, this method ties vertex state to the current context and can require more frequent state changes, making it harder to manage in larger applications. It is also incompatible with OpenGL core profiles, which require VAOs, and many drivers optimize VAO-based pipelines more effectively, potentially yielding better performance and organization.

Compatibility and context considerations: nonVAO usage persists mainly in compatibility profiles or legacy codebases and is enabled by extensions on older hardware. For modern rendering, developers typically adopt VAOs together with Vertex Buffer Objects and shader programs. The ARB_vertex_array_object extension historically provided VAO support in environments lacking core VAO support, and most contemporary implementations rely on VAOs by default.