The development of applications for virtual reality can be cumbersome, especially if the application needs to be tested frequently, as that may require frequent device switches by the developer. Self-sustaining programming systems, which allow code to be easily modified while it is running, may be a possible way to remedy this: By both testing and developing an application in VR, costly device switches are avoided, which leads to shorter feedback loops.

We present an implementation for a virtual reality live programming system which we call PIVR, with a development framework and user interface concept for coexisting applications. A block-based code editor allows the user to program within VR.

We analyzed applications that developers wrote for PIVR over the span of a semester, in which we found that it is a promising system for VR development. A preliminary user study of the block-based editor showed that it can be used to edit programs within VR and encourages users to engage in code exploration, as well as assist with the development of interactions, but still needs further work to achieve the same level of productivity as traditional development environments.

Using a computer for drawing has many advantages, including high expressivity, high precision, and the ability to build on the work of others. In my talk, I will show major milestones in the history of drawing programs. We will start with early CAD programs from the 1960s, and look at the first consumer-level art programs from the 1980s, leading to the lightweight drawing apps and websites of today. I will not only do a historic enumeration, but also categorize drawing programs to help you and me get a better understanding of the field. Preparing this talk is motivated by my current research on optimizing the tools in drawing programs for specialized use cases.

The temporally linear nature of videos poses a challenge in efficiently searching for and identifying specific segments of interest, such as instances when particular objects, individuals, or events are present. This challenge becomes more pronounced with longer videos and the occurrence of intricate actions within brief time intervals. There is a fundamental gap between the ability to formulate meaningful queries for text (e.g. natural language queries) and the ability to retrieve the parts of a video that correspond to such queries for videos. Filling this gap represents a key component for a video summarization system. Recently, unified vision-language models have been successfully used for approaching downstream tasks such as image labeling, text-to-image retrieval and generation, visual question answering, etc. In this talk, we explore the application of one such unified vision-language model, i.e.,  BLIP (Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation), for the task of Video Moment Retrieval (VMR), which involves finding moments from a single video described by a given natural language query. We present a conceptually simple approach that requires no task-specific training (i.e. zero-shot) and no additional modalities such as audio, while being competitive with current state-of-the-art. We discuss our approach, present some qualitative results, and address further steps for improvement.