P754, Standard for Floating Point Arithmetic, with an active 2019 standard, has reconstituted under Leonard Tsai’s leadership, with a revision PAR submitted for the June 5, 2024 NESCOM meeting. Officer candidates have volunteered and members have volunteered. (A large study group has been meeting regularly over the past year to lay the groundwork).
P1722, Standard for a Transport Protocol for Time-Sensitive Applications in Bridged Local Area Networks, also has a revision PAR to be considered at the June 5, 2024 NESCOM meeting
P3109, Arithmetic Formats for Machine Learning, PAR approved in September 2021, with a large diverse membership and actively meeting.
P3120, Quantum Computing Architectures: Quantum computing standards are under active discussion, although largely outside the MSC. Efforts are being made to determine which of the numerous standards across the IEEE are relevant and appropriate, and to unify efforts.
P1003.1 (POSIX, Austin Joint Working Group) had the P1003.1 (Portable Operating System Interface Base Specifications issue 8) standard was approved on May 21, 2024, and, as of May 29, 2024, was in the process of publication.
Current Standards Needs
The floating point arithmetic standard (IEEE 754), implemented partially or fully in virtually all desktop computers, some supercomputers, and some other computing equipment, is necessary to enable predictability, reliability, and portability of scientific computing, numerical simulation and models, as well as commercial software systems.
The IEEE 1722 set of standards specify protocols for interoperability of devices, such as audio and video devices, over IEEE 802-based networks (Ethernet and WI-FI). Standardization activity needs to continue as technology evolves and new devices and new applications emerge.
The arithmetic formats for machine learning standard, under initial development, has similar overall needs to the floating point arithmetic standard. However, a different standard is needed for machine learning, since machine learning applications require less accuracy than scientific and commercial applications, but have a bigger emphasis on efficiency, both in terms of execution time and computer memory requirements.
Quantum computing applications are in their infancy, although there are already a couple of commercial applications in use. Standards and recommended practice documents are needed for the interfaces between conventional computing circuitry and quantum computing apparati, as well as standardized metrics for the efficiency of quantum computers. These standards and recommended practices should enable transparency in the developing industry, as well as enable an orderly development and adoption of the technology.
The POSIX (Portable Operating System Interface) standards are needed to ensure operability between different Unix-like (Linux-like) operating systems. Such operating systems underlie much open-source software such as that in certain desktop computers and in many, if not most, mobile phones. Standardized interfaces are needed to permit efficient software development, to avoid having to custom-design application software for every system on the market.
Standards Stakeholders
Chip designers and manufacturers
Scientific software developers and users: Developers will have an interest in developing and using Floating Point Arithmetic (P754) and in following development in quantum computing (P3120), while users of scientific software will use such standards. These developers include people involved in modeling and simulation, such as in weather and climate forecasting. Some informed users will also be interested in participating in development.
Artificial intelligence software developers: These will be interested in developing and using the products of the arithmetic for machine learning (P3109) working group.
Commercial and operations research software and systems developers
Government laboratories and defense entities such as the national laboratories within the United States.
(Indirectly), a large, broad spectrum of end users.
