A dashboard gauge is not a second thermometer. It is a mapping: a logger value is placed on a visual scale, often with color bands, so a visitor can read temperature, pressure, or wind at a glance. The mapping can be honest or it can imply a precision and a risk level the sensor never earned. Cumulus sites used analog-style gauges for that glance. This page is about the mapping rules. It is not a PHP version dump, not a thermometer-image generator, and not a live instrument panel.
The historical URL /phpversion/gauges.php lived next to /phpversion, which is about the publishing runtime. Sharing a directory does not make gauges into phpinfo. The gauges consumed station data—often via XML or tags—and drew needles. The original widgets are not rehosted.
Historical context
Early personal-weather dashboards borrowed the language of analog panels: a round temperature dial, a pressure dial marked in inches of mercury or hectopascals, a wind-speed arc. PHP with GD, Flash widgets, and later JavaScript libraries all did the same job. Operators liked them because a needle is faster to scan than a table. The scientific cost is that a needle hides the sample time, the unit footnote, and the siting note unless the page puts those back in text.
A temperature time-series shows change. A gauge shows one instant. Both are derived displays of observed samples. Neither is a National Weather Service warning device.
What a gauge actually maps
Three quantities dominated hobby dashboards:
Temperature. The needle position is a linear or piecewise map from the logger’s air-temperature sample to an angle. The scale ends are a choice. A dial that runs from −40 to 50 °C is a different instrument graphic from one that runs from 60 to 90 °F. The choice should cover the climate of the site with margin, not the full physical range of the sensor chip.
Pressure. Station pressure and sea-level-reduced pressure are different. A gauge that says “pressure” without saying which, and without a unit, is not readable. Hectopascals (millibars) and inches of mercury must be labeled. Color bands that paint “high” and “low” as fair and stormy are folklore unless you state that they are a visual convention, not a forecast.
Wind. Speed and direction are two instruments. A single needle on a speed arc does not report direction. Gust versus average must be named. A dial marked to 100 mph in a sheltered yard implies a range the site will never use and makes ordinary winds look like nothing.
Each gauge needs the same observation contract as a bulletin: variable name, value, unit, time, and station. The needle is optional. The text is not.
Scale choice is a scientific decision
Pick scale ends from the station’s climatology plus a documented margin, not from a widget’s default. A default that was drawn for a maritime climate will pin a desert summer needle against the stop. A default drawn for summer will hide winter. When the needle sits on the peg, the display has saturated: the underlying logger may still be in range, but the graphic has stopped being a mapping.
Tick spacing should match the resolution you are willing to claim. If the sensor is honest to a degree, do not cut ticks at tenths and then print a digital inset to two decimals. The inset will be believed over the spec sheet.
Logarithmic or compressed wind scales can make gales visible without wasting arc on calm air. If you compress, say so in the caption. A visitor who compares two sites with different wind scales is not comparing winds; they are comparing graphics.
Color bands do not add a sensor
Red, yellow, and green feel like alerts. On a temperature gauge they usually mean “cold / comfortable / hot” relative to an unstated human comfort band, not relative to WMO instrument limits. On a pressure gauge they often recycle a home-barometer plate. Those plates are not NWS products.
If you use color:
- define the band edges in the unit of the scale;
- state that bands are a display aid, not warnings;
- do not rely on color alone (a needle angle plus a numeric text value);
- do not paint “dangerous” on a backyard gust that has no official meaning.
Official watches, warnings, and advisories come from the National Weather Service and local emergency authorities. A red arc on a hobby gauge is not a warning service. TNET is not a warning service either.
Do not imply more precision than the sensor
WMO temperature practice is about exposure and a representative measurement, not about painting 23.4 °C on a 12-bit backyard probe in still air beside a roof. The gauge should:
- round to the precision you document;
- show the observation time next to the needle;
- avoid animation that interpolates between samples as if the air were that smooth;
- refuse to display a value when the logger is stale, rather than holding the last needle as “current.”
Stale data is the failure mode gauges hide best. A table with a timestamp looks old. A shiny needle looks live. Pair every gauge with a clock and a max-age. If the sample is older than the logger interval plus a small tolerance, the display should read as unavailable, not as a confident analog truth.
Gauges versus charts versus thermometer images
Use this URL for mapped analog scales of several variables. Use the temperature chart for history. Use a thermometer image page, if present on this host, for a single-variable pictorial. Use /phpversion for the runtime that drew them. Crossing those intents produces duplicate articles and unreadable archives.
XML and tag files that feed gauges are observations plus structure. Parsing them is a different job from drawing a needle; see /parse_xml.php.
Practical checklist
- Numeric value, unit, and timestamp sit in text beside each gauge.
- Scale ends fit the site; saturation is visible or avoided.
- Temperature, pressure type, and wind statistic are named.
- Color bands are defined and not sold as official alerts.
- Display precision ≤ claimed instrument precision.
- Stale samples show as stale.
- PHP/GD version is recorded privately, not on the gauge page.
Modern relevance
Glanceable instruments remain useful. They remain easy to over-read. A connection-research product has the same temptation: a single score or needle that hides whether the input was observed, modelled, or derived. TNET’s public grouping of those inputs is the evidence families behind a connection outlook. Station-era gauge pages belong with the Cumulus legacy hub.
This URL does not drive live needles. It states how a mapped instrument should behave if someone draws them again.
Sources
- World Meteorological Organization, Guide to Instruments and Methods of Observation (WMO-No. 8), temperature, pressure, and wind chapters: community.wmo.int
- National Weather Service, official forecasts and alerts: weather.gov
- PHP, GD image processing (historical drawing path): php.net/book.image
- TNET, The evidence families behind a connection outlook