|
5 | 5 | @setfilename The_Steelbore_Standard.info |
6 | 6 | @documentencoding UTF-8 |
7 | 7 | @documentlanguage en |
8 | | -@settitle The Steelbore Standard 1.29 |
| 8 | +@settitle The Steelbore Standard 1.30 |
9 | 9 | @c %**end of header |
10 | 10 |
|
11 | | -@set VERSION 1.29 |
12 | | -@set UPDATED 2026-06-23 |
| 11 | +@set VERSION 1.30 |
| 12 | +@set UPDATED 2026-06-24 |
13 | 13 | @set SUBDOMAIN https://Standard.SpacecraftSoftware.org/ |
14 | 14 |
|
15 | 15 | @copying |
@@ -106,6 +106,16 @@ are separate concerns. |
106 | 106 | @section Changelog |
107 | 107 | @itemize |
108 | 108 | @item |
| 109 | +@strong{v1.30 (2026-06-24):} @strong{§3.2} reframed --- modern hardware |
| 110 | +universally provides multi-core/multi-thread capability; harnessing that |
| 111 | +concurrency is the primary performance lever. Concurrency is an |
| 112 | +@strong{architecture-level concern}, considered from the ground up |
| 113 | +throughout design (not bolted on at implementation). |
| 114 | +Adoption/abandonment conditions explicit: embrace concurrency where it |
| 115 | +advances performance; abandon it where it degrades performance (overhead, |
| 116 | +contention, serial workloads) or compromises Priority 1 (Stability). |
| 117 | +Compliance-checklist bullet revised. |
| 118 | +@item |
109 | 119 | @strong{v1.29 (2026-06-23):} Switch source of truth from GFM Markdown to |
110 | 120 | Texinfo: @file{The_Steelbore_Standard.texi} is now the canonical source; |
111 | 121 | @file{The_Steelbore_Standard.md} and @file{The_Steelbore_Standard.html} are |
@@ -640,19 +650,24 @@ not asserted by inspection alone. |
640 | 650 |
|
641 | 651 | @node Priority 2 |
642 | 652 | @section §3.2 — Priority 2: Performance |
643 | | -Performance is the foremost priority after stability. The default means |
644 | | -of achieving it is @strong{multi-core, multi-thread concurrency} --- |
645 | | -parallelism is the expected baseline, not an afterthought --- |
646 | | -@emph{unless} concurrency would materially degrade performance |
647 | | -(synchronization overhead, lock contention, or inherently serial / small |
648 | | -workloads outweighing the gains), in which case a simpler or serial |
649 | | -approach must be chosen and the trade-off documented. |
| 653 | +Performance is the foremost priority after stability. Modern hardware |
| 654 | +universally provides @strong{multi-core, multi-thread} capability; |
| 655 | +harnessing that concurrency is the primary means of achieving |
| 656 | +performance. Concurrency is not an afterthought --- it must be |
| 657 | +@strong{considered from the ground up}, throughout architecture design: |
| 658 | +data ownership, thread boundaries, synchronization points, and |
| 659 | +parallelism opportunities must be identified during design, not |
| 660 | +discovered during optimization. |
| 661 | + |
| 662 | +Concurrency is adopted where it genuinely advances performance. It is |
| 663 | +@strong{abandoned} where it degrades performance (synchronization |
| 664 | +overhead, lock contention, or inherently serial / small workloads) or |
| 665 | +where it would compromise Priority 1 (Stability). When a serial or |
| 666 | +simpler approach outperforms or is safer, it must be chosen and the |
| 667 | +trade-off documented. |
650 | 668 |
|
651 | 669 | @itemize |
652 | 670 | @item |
653 | | -Concurrency must be @strong{designed-in from the start}, never bolted on |
654 | | -retroactively. |
655 | | -@item |
656 | 671 | Release builds should use CPU-optimized flags --- @code{-march=native}, |
657 | 672 | LTO, PGO --- @strong{where the toolchain and target support them |
658 | 673 | reliably.} Any such flag known to break or destabilize builds on a given |
@@ -2100,9 +2115,10 @@ explicitly renamed |
2100 | 2115 | ☐ @strong{§3.1} Stability: memory safety (Rust, or ASLR+CFI documented); |
2101 | 2116 | robust error handling, fault tolerance, and test-verified |
2102 | 2117 | @item |
2103 | | -☐ @strong{§3.2} Performance: multi-core/multi-thread concurrency by |
2104 | | -default (or serial trade-off documented); concurrency designed-in; |
2105 | | -benchmarking before/after |
| 2118 | +☐ @strong{§3.2} Performance: concurrency considered throughout |
| 2119 | +architecture design; adopted where it advances performance, abandoned |
| 2120 | +where it degrades performance or compromises Stability; serial |
| 2121 | +trade-off documented; benchmarking before/after |
2106 | 2122 | @item |
2107 | 2123 | ☐ @strong{§3.3} Hardened security; PQC readiness addressed |
2108 | 2124 | @item |
|
0 commit comments