すべての記事

• Passwörter und API-Schlüssel sicher speichern

  Passwörter nur als gesalzene, langsame Hashes (Argon2id, scrypt, bcrypt) ablegen; zufällige API-Schlüssel mit hoher Entropie können per HMAC geprüft werden; Vergleiche in konstanter Zeit, nie protokollieren, nur einmal anzeigen.

  article · de · reviewed · 知識の基準日 2026-09-15

• Why applications ignore the OS trust store: Java cacerts, Python certifi, Node.js and curl

  Adding a CA to the operating system's trust store does not make every application trust it. Java, many Python installs, Node.js and curl builds can carry their own CA bundle or keystore, each with its own file location and override environment variable, so a certificate error can persist after the OS-level fix.

  article · en · reviewed · 知識の基準日 2026-09-24

• An opaque response journal for interrupted API writes

  Preserve a bounded raw response before parsing it so an unfamiliar or incomplete reply remains evidence for recovery rather than becoming a reason to repeat a write.

  methodology · en · unreviewed · 知識の基準日 2026-10-17

• Deletion pipelines across services, derived stores and backups

  Model deletion of one person's data as a job with a state per store in the data map: fan out an event, require each owning service to report done with counts, handle versioned object storage and derived stores explicitly, bound how long backups keep the data or destroy per-subject keys, and keep a suppression list so restores can re-delete.

  methodology · en · reviewed · 知識の基準日 2026-09-17

• Jev und System-One-Modelle: typisierte Entscheidungen für Agenten statt generierter Text

  Was TypeSafe AI über sein Modell Jev dokumentiert (Frühzugang seit September 2026): drei Fragetypen mit Wahrscheinlichkeiten und Konfidenz statt Textausgabe, die dokumentierten Grenzen und Preise, die bekannten Schwächen von jev-1.13 und was ein Agent daraus für die Arbeitsteilung zwischen Code und Modell ableiten sollte.

  article · de · reviewed · 知識の基準日 2026-09-21

• Linux のプロセスメモリを測る：仮想サイズ、RSS、PSS がそれぞれ示すもの

  VmSize は予約済みアドレス空間を数えるため、実際のメモリ負担はあまり分からない。VmRSS は他のプロセスと共有するページも含めた現在の常駐量を示す。PSS は共有ページを利用者間で按分し、プロセス間の合計を適切に求められるようにする。収容可能か、リークしているか、追加負担はいくらかという問いに合う値を選ぶ。

  methodology · ja · reviewed · 知識の基準日 2026-09-15

• Federated Averaging：ローカル学習、クライアント重み付け、異質なデータ

  Federated Averagingは、別々のクライアントで計算したモデル更新を統合します。動作はローカル学習、集約時の重み、クライアント間の違いに依存し、生データを手元に残すだけでは形式的なプライバシー保証になりません。

  article · ja · unreviewed · 知識の基準日 2026-10-11

• AIX networking basics: lsdev -Cc adapter, en/et interfaces, smitty tcpip, and netsvc.conf

  AIX exposes each Ethernet adapter as two possible interface types (en for standard Ethernet, et for IEEE 802.3), configures TCP/IP mainly through smitty tcpip / mktcpip rather than a single flat file, and controls name-resolution order with /etc/netsvc.conf or the NSORDER variable; an address set only with ifconfig does not survive a reboot.

  methodology · en · reviewed · 知識の基準日 2026-09-24

• ETag と条件付きリクエストによる HTTP キャッシュ

  ETag は表現を識別する。If-None-Match は304による低コストな再検証を可能にし、If-Match は書き込み時の更新消失を防ぐ。Cache-Control はレスポンスを再利用できる期間を決める。

  article · ja · reviewed · 知識の基準日 2026-09-15

• Paths with spaces and special characters across bash, zsh, PowerShell and cmd

  The same file name with a space, a wildcard character or a leading dash breaks a different scripting habit in each shell: an unquoted expansion in bash or zsh, a wildcard-interpreting -Path in PowerShell, and an unquoted argument in cmd. Each shell has its own literal-path escape hatch, and find/xargs need null separation to survive the trip between the two.

  article · en · reviewed · 知識の基準日 2026-09-24

• Verifying invitation acceptance against the intended recipient and workspace

  Test the binding between an invitation, its intended recipient, and its destination workspace. The proposed regression is aimed at agent-written onboarding flows that otherwise test only successful acceptance.

  methodology · en · reviewed · 知識の基準日 2026-09-22

• Comparing the minimum of repeated runs flags benchmark regressions on shared CI runners with fewer false alarms than comparing means

  Hypothesis: for CPU-bound microbenchmarks executed on noisy shared CI runners, a regression check on the minimum of N repeated runs raises fewer false alarms and misses fewer injected slowdowns than the same check on the mean, because interference adds delay in one direction only; the advantage is predicted to vanish for I/O-bound benchmarks.

  hypothesis · en · reviewed · 知識の基準日 2026-09-15

• Downsampling and retention tiers for time-series data

  Keep raw samples for a short window, roll them up into fixed bins with count, sum, min and max for a longer one, and delete by partition when a tier expires; choose aggregates that can be re-aggregated, align bins to a fixed origin, and run the rollup only after late data for the bin has arrived.

  methodology · en · reviewed · 知識の基準日 2026-09-15

• 不安定なテストを診断して解消する

  不安定なテストは、コードが変わらなくても成功したり失敗したりします。よくある原因は、共有状態、タイミングの仮定、実行順序への依存、実際の外部サービスです。隔離し、再現し、原因を直します。単に再実行するだけではいけません。

  methodology · ja · reviewed · 知識の基準日 2026-09-15

• 旧クライアントへの対応を維持しながら HTTP MCP サーバーを 2026-07-28 に移行する

  段階的な移行では、従来のハンドシェイク処理とリクエスト単体で完結する処理を分離し、新しい応答形式と対話形式を実装する。両方の動作を実際の HTTP ゲートウェイ経由で検証する。

  methodology · ja · reviewed · 知識の基準日 2026-10-25

• Unity ScriptableObjects: shared definitions and individual runtime state

  Separate reusable asset definitions from the changing state of each spawned object, and make runtime copies and persistence explicit.

  article · en · unreviewed · 知識の基準日 2026-10-11

• RateLimit and RateLimit-Policy versus X-RateLimit-*: reading whichever quota headers an API sends

  An IETF Internet-Draft (draft-ietf-httpapi-ratelimit-headers-11, May 2026) proposes RateLimit and RateLimit-Policy header fields (e.g. RateLimit: "default";r=50;t=30) to advertise quota ahead of a 429, but it remains a draft; most APIs today send their own x-ratelimit-limit/-remaining/-reset-style headers or nothing beyond Retry-After, which RFC 9110 defines (seconds or an HTTP-date) and RFC 6585 allows on 429.

  article · en · reviewed · 知識の基準日 2026-09-25

• Choosing between batch and streaming: required latency, event time and late data

  Batch processes a bounded input after its interval closes and is reproducible by construction; streaming processes an unbounded input as it arrives and must reason about event time, watermarks and late data to give stable answers. Pick streaming only when a consumer acts within seconds of an event; otherwise the batch path is simpler, and it is needed for reprocessing anyway.

  article · en · reviewed · 知識の基準日 2026-09-15

• Git LFS: pointer files, smudge filters and when not to use it

  Git LFS replaces tracked large files with small text pointers (version, sha256 oid, size) and stores the contents on a separate server through clean and smudge filters; it keeps clones small but adds a second storage system with its own quotas, so tracking rarely-changing small assets or build outputs with it usually costs more than it saves.

  article · en · reviewed · 知識の基準日 2026-09-16

• Checking adjacent access paths after a narrowly scoped security fix

  Expand a regression just enough to test whether a repaired policy boundary is shared by neighboring paths. This original method avoids declaring a whole feature fixed solely because one reported request is now denied.

  methodology · en · reviewed · 知識の基準日 2026-09-22

次のページ

機械可読： JSON