SPB Git

Repository Public License

competitive-analysis.md 10.9 KB

# Competitive analysis

# Scope and method

This is a qualitative comparison drawn from general public knowledge and public documentation of the tools named below. It is not a benchmark: no competitor was installed, timed, profiled or measured for this document, and no version number, price or performance figure is quoted for any of them. Feature descriptions of other tools are stated at the level of "commonly documented behaviour" and may not match the release a reader has installed. Where a statement about another tool could not be made with confidence, it is omitted or marked "varies".

Statements about ORIX Monitor, by contrast, are taken from its source code and can be checked there.

Tools considered:

  • Activity Monitor — the process and resource viewer shipped with macOS.
  • iStat Menus — a commercial menu-bar monitoring suite.
  • Stats — an open-source menu-bar system monitor.
  • Similar tools — command-line monitors (top, htop, btop) and Apple's privileged powermetrics tool, mentioned where the contrast is instructive.

# Positioning in one paragraph

Activity Monitor is the reference for what is running right now, with privileged access no third-party app has. Menu-bar suites optimise for glanceable live values, including hardware sensors. ORIX Monitor's design centre is different: every number carries its provenance, history is stored locally in tiers with explicit gaps, and anything the operating system does not expose through a supported interface is declared unavailable instead of being approximated or read through the SMC.

# Capability comparison

Legend: Yes = present; No = absent; Partial = present with a stated restriction; Varies = depends on the Mac model, configuration or release, per public documentation.

Capability Activity Monitor iStat Menus Stats ORIX Monitor
CPU total and per core Yes Yes Yes Yes
Memory breakdown and pressure Yes Yes Yes Yes (app / wired / compressed / cached / free, kernel pressure level)
Process list with CPU and memory Yes, all users Partial (top processes) Partial (top processes) Yes, own processes; other users' processes listed without figures
End a process Yes Varies Varies Yes, own processes only, after confirmation
Disk throughput Yes Yes Yes Yes (physical block devices, disk images excluded from totals)
Volume capacity Via Disk Utility / Finder Yes Yes Yes (logical volumes, with container BSD name)
Network throughput Yes Yes Yes Yes (hardware links summed; tunnels listed, not added)
Per-process network traffic Yes Varies Varies No — declared unavailable
GPU utilization Yes (GPU history) Yes Yes Yes, labelled "Best effort"
Per-process GPU usage Yes No / Varies No / Varies No — declared unavailable
Battery charge, cycles, health Via System Settings Yes Yes Yes; cycles and capacity labelled "Best effort"
Temperatures (CPU, GPU, SSD) No Yes (via SMC / sensor services) Yes (via SMC / sensor services) No — declared unavailable (battery temperature only, if published)
Fan speed and fan control No Yes Yes (reading; control varies) No — declared unavailable; no control
Thermal state (system's own level) No Varies Varies Yes
Menu-bar indicators Dock icon only Yes (core feature) Yes (core feature) Yes (one status item, five optional indicators)
Long-term stored history No (live window only) Yes (history graphs) Limited Yes (raw / minute / hour tiers in local SQLite)
Gaps in history shown explicitly n/a Varies Varies Yes (shaded, listed per metric in reports)
Threshold alerts No Yes Varies Yes (hold duration, cooldown, snooze)
Automated findings from history No No / Varies No Yes (seven detectors, each with its evidence)
Exportable reports Sample / diagnostics files Varies No Yes (PDF, Markdown, CSV, JSON; comparison of two JSON reports)
Source and method shown per metric No No No (source code is public) Yes ("Sources and methods" card on each page, repeated in reports)
Quality label per value No No No Yes (measured / derived / estimate / best effort / unavailable)
Requires a privileged helper Built into the OS For some features, per its documentation For some features, per its documentation No helper installed
Network access by the app n/a Update checks / weather features, per its documentation Update checks, per its documentation None, except the on-demand latency test
Source availability Closed (Apple) Closed Open source Proprietary (see LICENSE)

# Temperatures and fans: why ORIX Monitor shows none

Menu-bar monitors that display CPU, GPU or SSD temperatures and fan speeds obtain them from the System Management Controller (SMC) through an IOKit user client that Apple does not document, and, on Apple Silicon, from private HID sensor services. These interfaces work, and the tools that use them are widely trusted. ORIX Monitor deliberately does not use them, for four reasons that follow from its design rule (every displayed number must come from a supported interface, or be labelled otherwise):

  1. They are undocumented. SMC keys are four-character codes with no published meaning. Which key is "the CPU temperature" is established by community reverse engineering, differs between Mac models and chip generations, and has changed across hardware revisions.
  2. A wrong key yields a plausible wrong number. A mis-mapped sensor does not fail; it returns some other temperature. That is precisely the failure mode ORIX Monitor is built to avoid.
  3. They can break with any update without notice, since no compatibility is promised.
  4. A supported signal exists. ProcessInfo.thermalState is the system's own four-level assessment (nominal, fair, serious, critical) and is what macOS itself acts on when it throttles. ORIX Monitor shows, stores, charts and alerts on that.

Concretely, the Thermal & Sensors page shows UnavailableRow entries for CPU temperature, GPU temperature and fan speed with the text "Not available through supported system interfaces.", and SensorCapabilityResolver classifies them as unavailable. A UI test (testUnavailableSensorsAreDeclaredNotInvented) fails if any "rpm" value appears on that page.

The one temperature ORIX Monitor may show is the battery temperature, read with a public IOKit call from the AppleSmartBattery registry entry and labelled "Best effort". On the audited Mac (macOS 27.0) the key is not published, and the page says so. See the capability matrix.

Trade-off stated plainly: a user who wants degrees Celsius per sensor or fan RPM will not find them in ORIX Monitor and is better served by a tool that reads the SMC.

# Other differences in approach

# Process visibility

Activity Monitor runs with entitlements that let it read every process on the system. A third-party app without a root helper cannot: the kernel refuses resource statistics for processes of other users. Some tools install a privileged helper to close that gap. ORIX Monitor installs none; it lists those processes by identity, leaves CPU, memory, threads and start time as "—", counts them (ProcessSample.restricted) and explains why. It never shows zero for an unknown value.

# GPU utilization

macOS offers no supported, system-wide GPU utilization API. All third-party tools (and ORIX Monitor) read the PerformanceStatistics dictionary that the graphics driver publishes on its IOAccelerator registry entry. ORIX Monitor's difference is labelling: the value carries a "Best effort" badge everywhere it appears, an "About this figure" card explains where it comes from, and a single setting turns all undocumented-property readings off.

# Power

powermetrics reports per-cluster CPU power and package power on Apple Silicon but requires root. ORIX Monitor does not invoke it. It shows (a) the battery flow, voltage × current at the battery gauge, (b) the gauge's SystemLoad telemetry value when present, and (c) the kernel's per-process billed energy for the user's own processes. All three are labelled; (a) and (b) are "Best effort", and (c) is described in the interface as "an accounting figure, useful to compare processes, not a measurement at the wall". On a desktop Mac no system power figure is shown at all.

# History and gaps

Activity Monitor keeps only a short live window. Menu-bar suites typically keep history graphs. ORIX Monitor stores three tiers (every sample, one-minute and one-hour min/avg/max summaries) and treats absence of data as information: periods when the app was not running, the Mac was asleep or sampling was paused are detected (SeriesMath.gaps), shaded in charts, listed per metric in reports, and never bridged by interpolation. Findings detectors end an episode at a data hole rather than counting it. See storage.

# Privacy posture

ORIX Monitor has no account, analytics, advertising or update check in its code. Its only outbound traffic is the latency test, a TCP handshake to a host the user types, triggered by a button. It requests one optional permission (notifications) and does not request Location Services, which is why it does not show the Wi-Fi network name. Other tools make different choices (for example, automatic update checks), which is a reasonable convenience trade-off rather than a defect.

# Where ORIX Monitor is weaker

  • No hardware sensors (temperatures, fans, per-rail power, SSD wear).
  • No figures for other users' processes and no per-process network or GPU usage; Activity Monitor has all three.
  • Menu-bar customisation is modest: one status item with up to five indicators and optional mini bar charts, versus the highly configurable, multi-item menus of dedicated menu-bar suites.
  • No fan control, no weather, no world clocks, no Bluetooth-device battery levels and similar extras some suites bundle.
  • Apple Silicon only, macOS 15 or later. The release build ships an arm64 slice only; the app has never been run on an Intel Mac.
  • Validated on one Mac. See the audited environment in the capability matrix.
  • History exists only while the app runs. There is no background daemon.
  • No remote or multi-machine monitoring.

# Limitations of this analysis

  • Competitor capabilities are summarised from public knowledge at the time of writing and are not verified against specific releases.
  • No performance, energy or memory comparison is made. ORIX Monitor's own overhead methodology is in performance; it measures ORIX Monitor only.
  • The comparison table reflects default, documented behaviour; several tools are configurable beyond what a single cell can express.
Files main
Menu

Theme