Day 4 - Federating STEM from machine to network - Caswell

Day 4 - Federating STEM from machine to network - Caswell

🎙 Tom Caswell 👥 1K 📅 July 18, 2026 ⏱ 50 min 👁 9 📄 expert opinion 🧭 2026-08-16
Available in: English (current) Français

Keywords

Blueskydata acquisitionEPICSTangoTiledorchestrationopen sourcesynchrotronelectron microscopy

Summary

Tom Caswell, a lead developer of the Bluesky ecosystem at NSLS-II, presents a lecture on the design, architecture, and deployment of Bluesky, a Python-based data acquisition framework. He begins by discussing the challenges of managing a large user facility like a synchrotron, where the complexity of instruments has grown from simple mechanical adjustments to thousands of networked devices. He emphasizes the need for reliable yet adaptable control systems, and introduces Bluesky as an orchestration layer that sits on top of lower-level control systems like EPICS or Tango. Bluesky provides a set of abstractions that allow scientists to write reusable experiment plans independent of specific hardware. He also introduces Tiled, a data access layer that serves data directly to clients in various formats. Caswell highlights the importance of open-source collaboration, noting that Bluesky was developed in the open from day one and is now used at multiple facilities worldwide. He discusses the technical goals of Bluesky, including streaming data, support for asynchronous data streams, and scalability from small labs to large facilities. He also touches on the sociological goals, such as avoiding NIH syndrome and maintaining well-defined component boundaries. Finally, he suggests how Bluesky could be adapted for electron microscopy, noting ongoing efforts to integrate with Tango and the potential for incremental adoption.

212 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into the design philosophy and practical considerations behind a widely adopted open-source data acquisition system. Caswell’s argumentation is clear and logical, grounded in his direct experience. He effectively explains the need for abstractions and the benefits of open collaboration. The discussion of scaling and the example of driving two beamlines simultaneously with AI demonstrates the system’s capabilities. However, the talk is largely a presentation of the speaker’s own work, so it lacks critical evaluation of alternatives or potential drawbacks.

93 words

Title / Content Match

The title is somewhat vague but accurately reflects the focus on federating STEM (Scanning Transmission Electron Microscopy) and the broader context of connecting machines to networks via the Bluesky ecosystem.

Quality & Reliability

8/10

The speaker is a primary author of the Bluesky ecosystem, providing first-hand expertise. The talk is technical and detailed, but it is a presentation of their own work, so there is potential bias. No external sources are cited, but the information is consistent with known public documentation.

Key Moments

Cited Sources

Concurring Sources

Contribution & Novelties

The talk provides an insider’s perspective on the design and evolution of the Bluesky ecosystem, which is a significant contribution to the scientific software community. It explains the rationale behind key design decisions, such as the separation of concerns and the emphasis on open collaboration. The discussion of adapting Bluesky to electron microscopy is timely and relevant.

Pour aller plus loin :

102 words

Radar Profile

The radar profile shows high scores in information quantity, quality, and technical level, with a slightly lower but still strong score in global reliability. This indicates a technically rich and informative talk, though the reliability is slightly tempered by the lack of external citations and the speaker's inherent bias as a developer.

Reliability 8/10

💬 No comments were provided for analysis.