Sources

paint

Wall paint coverage

Every row below was read off the manufacturer's own product data sheet, not a retailer listing or a search summary. Coverage figures are per coat.

Overlap across all three: 350-400 sq ft/gal. Every product's stated range contains it. All three state their number at roughly the same 4 mils wet, so they are directly comparable rather than measured on different bases.

Behr: dropped, not pending. behr.com serves a Cloudflare block page to both WebFetch and a headless browser, so no Behr figure can be read from a primary source. Rather than leave a permanent gap, SPEC.md section 5 was amended on 2026-08-09 to use Sherwin-Williams as its reference example instead. Do not reintroduce a Behr number from a retailer listing or a search snippet.

Primer coverage

Note the age of this TDS (2011). It is the current document served for the product, but re-check before leaning on it in a second tool.

Coats — manufacturer wording, not our opinion

Both statements below are quoted because the two-coat default is a real cost driver and needs to trace to the manufacturer rather than to us.

Door and window deductions — convention, NOT a manufacturer spec

No paint manufacturer publishes a deduction figure. These are estimating conventions, and the guide must present them as such. The door default does trace to a real door size; the window default does not trace to anything but common practice, because window sizes are not standardized the way doors are.

Triangulation (SPEC section 5.2)

Worked example held constant: 12 ft x 14 ft room, 8 ft ceilings, 1 door, 2 windows, 2 coats. Gross wall area = 2 x (12 + 14) x 8 = 416 sq ft.

Finding, and it belongs in the guide: the three sources land within 7 sq ft of each other on net area (365-372, a 1.9% spread), so the door/window convention barely moves the answer. The coverage constant moves it entirely — this example sits right on the 2-vs-3 gallon boundary, and which side it falls on is decided by 350 vs 400 sq ft/gal alone. That is the honest headline: stop arguing about window deductions and read the number off your can.

Consequence for the tool: default coverage to 350 sq ft/gal (inside all three manufacturer ranges, and the side that avoids a second trip to the store), make it user-editable, and show the result at both 350 and 400 so the boundary is visible instead of hidden behind a single rounded number.

concrete

| 2026-08-12 | QUIKRETE Concrete Mix #1101 (SPEC-DATA, 03 31 00 Structural Concrete) | "A 40 lb (18.1 kg) bag yields approximately 0.30 ft3 (8.5 L)"; "A 60 lb (27.2 kg) bag yields approximately 0.45 ft3 (12.7 L)"; "An 80 lb (36.2 kg) bag yields approximately 0.60 ft3 (17 L)". Also states 0.375 ft3 for 50 lb and 0.675 ft3 for 90 lb. | quikrete.com ↗ |

Re-sourced 2026-08-12 to the manufacturer's own copy. The 2026-08-09 row above cites the identical SPEC-DATA sheet hosted on a quarry's WordPress site rather than on quikrete.com. The figures are the same and none of them changed, but a third-party mirror is a weaker citation than the maker's own document, and SPEC section 5.1 asks for the primary one. The old row stays because this file is append-only; the new row is the one in force.

Found by the monthly source check on its first run. kingquarry.com served that PDF to a laptop and returned nothing to a GitHub runner, which is what prompted looking for a better home for the document.

Derived and printed on the page: 27 cu ft per cu yd means 45 bags of 80 lb, 60 of 60 lb, or 90 of 40 lb per cubic yard. The 1 cu yd bag-vs-truck crossover is a handling judgment, labeled as such in the guide, not a spec.

fence

Reconciliation, and it changed the tool's default. Quikrete's two-bag example is 0.75 cu ft. The geometry disagrees: a 9 in hole 30 in deep is 1.10 cu ft, the 3 in post displaces 0.16, leaving 0.94 cu ft needed. Subtract 6 in of drainage gravel at the base so concrete fills only the top 24 in and the requirement falls to 0.76 cu ft, matching their example within 1%. So the stated figure assumes a gravel base. The calculator exposes gravel depth as a field defaulting to 6 in, and the guide shows this working rather than hiding the discrepancy. Verified live: 9 in / 30 in / 6 in gravel returns 0.76 cu ft per hole, exactly 2.0 bags per post.

deck stain

Default coverage 250 sq ft/gal, mid-band for rough wood, because a weathered deck is the realistic case. Note this tool defaults to one coat, unlike paint, and that comes straight from the manufacturer's own instruction.

drywall

Rates used: 0.009 gal compound, 0.35 ft tape and 1.25 screws per sq ft. The chart is linear across every row from 100 to 1,500 sq ft, so the division is clean. Published typo: the 1,300 sq ft row prints "17.7 Gal." where the series gives 11.7. Every other row follows the rate exactly. We use the rate, not that row, and the guide says so.

grass seed

Rates are derived from those coverages rather than from a rule of thumb: 9.09, 7.50 and 3.03 lb per 1,000 sq ft for a new lawn. The 3:1 finding: all three products' overseeding coverage is almost exactly three times their new-lawn coverage (5,280/1,760, 8,000/2,665, 1,000/330). That consistency is what justifies the single overseeding multiplier, and it is in the guide.

mulch

Bag sizes (1.5 / 2 / 3 cu ft) are as sold, not a spec, and are a user choice because all three share shelf space. 27 cu ft per cu yd is definitional. The 2 cu yd bags-vs-bulk crossover is a handling judgment, labeled as such.

gravel

No manufacturer spec exists for this tool and the guide says so outright. Volume is exact geometry. Tonnage depends on the density of the specific stone at its current moisture, which ranges roughly 1.3-1.7 tons per cu yd across common materials. The calculator defaults to 1.4 and exposes density as an editable field, with the guide telling readers to get their supplier's figure. Do not replace this with a "sourced" constant; there isn't one.

tile

Waste factors (10% straight lay, 15% diagonal or offset, 20% herringbone or complex rooms) are an estimating convention in common use, not a manufacturer spec, and the guide labels them that way. Tile geometry (144 sq in per sq ft) is definitional. Tiles per box is a user input because it varies per product.

wallpaper

Deliberately no hard-coded product constants. Roll width, roll length and pattern repeat vary enough between manufacturers that any default would be wrong more often than right, so all three are inputs defaulted to common American values (about 20.5 in wide, about 33 ft double roll) and labeled "read your roll". The 4 in trim allowance and the half-strip credit per opening are conventions, both explained in the guide.

Verified behavior: the strips-per-roll integer, not area, drives the roll count. On a 7 ft wall with a 33 ft roll, a 9 in repeat still yields 4 strips per roll, but a 25 in repeat drops it to 3 and pushes 24 strips from 6 rolls to 8. That worked example is in the guide and was checked against the live tool.

topsoil

Application depths are not a manufacturer spec. Nobody sells "soil" to one standard, so every figure below comes from a university extension program and is quoted as stated. Volume itself is exact geometry.

Penn State's own rate confirms the formula. A one-inch layer over 1,000 sq ft is 1,000 x (1/12) / 27 = 3.086 cu yd by geometry; they publish 3.1. Two inches gives 6.17 against their 6.2. An independent authority arriving at the same number from its own work is the strongest check this tool gets.

What was deliberately not used. That page also carries a Table 2 of cubic yardages, and it would not read reliably through an automated fetch: the column alignment came back inconsistent across two attempts and one row did not reconcile with the others. Only the prose rates above, which came back identical both times, are used. A figure we cannot read cleanly from the primary document does not get used (SPEC section 5.1).

Bag sizes (0.75 / 1 / 1.5 / 2 cu ft) are as sold, not a spec, and are a user choice because all four share shelf space. 27 cu ft per cu yd and 324 sq ft per cu yd at 1 inch are definitional. The 2 cu yd bags-versus-bulk crossover is a handling judgment, labeled as such, and is the same one the mulch tool uses.

No settling factor. Bagged soil compacts after watering and no primary source publishes a percentage worth using, so the tool does not apply one and the guide says why. Do not "fix" this by inventing a shrink allowance.

Triangulated 2026-08-12 against two independent calculators, both of which publish their formula rather than hiding it. Inch Calculator's soil calculator gives a worked example of 350 sq ft at 3 in as 87.5 cu ft and 3.24 cu yd (inchcalculator.com ↗); ours gives 87.5 and 3.24. CalculatorSoup's cubic yards calculator states the same chain, area x depth then divide by 27 (calculatorsoup.com ↗).

fence — board lumber (added 2026 08 12)

The original fence section covers post count and concrete. Rails and pickets were added 2026-08-12 and turn on dressed lumber widths, so those get their own rows. The same table is what makes a 4x4 post 3.5 in and a 6x6 5.5 in, which this tool already relied on without a citation.

Read from the document, not from a summary. The first automated fetch of that PDF returned binary and could not report the table at all. The figures above were extracted from the downloaded file with pdftotext -layout and read off Table 3 directly, which is what SPEC section 5.1 requires.

Not specs, and the guide says so: rails per section (2 or 3) and the gap between pickets. Neither is published anywhere as a standard. Both are inputs, defaulted to 3 rails and a 0 in gap, and both are labeled as conventions.

Verified behavior: the picket count turns on pitch, not on area. A 100 ft run of butted 1x6 needs 219 pickets at the 5.5 in dressed width against 200 if you use the nominal 6 in, a 19-board error. A 1/4 in gap drops it to 209. That worked example is in the guide and was checked against the live tool.

fertilizer

No manufacturer figure is used here, and that is deliberate. A fertilizer maker publishes a setting for its own spreader and its own granule; the underlying rate is published by university turf programs, so that is what this tool cites. The conversion itself is arithmetic, not a spec.

The default state is the source's own worked example. The tool opens at a 5,000 sq ft lawn, a 26% product and a 1.0 lb nitrogen target, which must print 3.8 lb per 1,000 sq ft. If it ever prints anything else, the tool has drifted from Penn State's published arithmetic and that is the first thing to check.

Triangulated 2026-08-12 against Purdue's Turfgrass Fertilizer Calculator (turf.purdue.edu ↗), which is built on the same division, and against Penn State's second page on the subject, "Turfgrass Management: Determining How Much Nitrogen or Active Ingredient You Are Applying". Two independent turf programs, one arithmetic.

Not specs, and the tool says so: the maintenance selector uses the midpoints of MSU's ranges (1.5 / 3 / 4.5 lb per year) because a range cannot be a dropdown value. The ranges themselves are quoted in the guide and the hint. Application timing is deliberately absent: it is regional and grass-specific, and the guide sends readers to their own state extension rather than inventing a national calendar.

room

No new figures. Every constant on this page is lifted unchanged from the paint and drywall tools so the three can never disagree: paint coverage from the Sherwin-Williams / Benjamin Moore / Valspar overlap band (see ## paint), and joint compound, tape and screw rates from the National Gypsum estimator (see ## drywall). Those rows carry the URLs and the dates they were read; they are not duplicated here, because a second copy of a citation is a second thing to keep in sync.

The one thing this tool adds is the asymmetry between the two trades, and it is a convention rather than a spec, so the guide labels it as one:

Drywall is cut to the framed rough opening, which is deliberately larger than the unit that goes into it. Paint stops at the finished trim, where a 3 x 5 ft window is 15 sq ft. Both figures match the standalone tools exactly, which is the point: this page is the two of them run off one measurement, not a third opinion.

Verified 2026-08-12 by driving all three pages, not by reading them. Four rooms were entered into /tools/room/, /tools/paint/ and /tools/drywall/ and the answers compared: 14x12x8 with 1 door and 2 windows (12 sheets, 3 gal), 20x10x9 with 2 and 3 (14 sheets, 3 gal), 10x10x8 on 4x12 sheets (6 sheets, 2 gal), and 12x12x10 at one coat and 400 sq ft/gal (11 sheets, 2 gal). All three tools agree on every one. The painted area runs 5 sq ft per window above the boarded area throughout, which is the asymmetry above showing up in the output. Re-run that cross-check if any of the three tools' constants change.

paver

Base thickness is deliberately not a constant here, and the source is the reason why. That same specification leaves it as a bracketed blank (Section: [ ]-Aggregate Base) because it is a project design value that depends on soil and traffic, not a published figure. So base depth is an editable field defaulting to 4 in with the guide telling readers what drives it. Do not replace it with a sourced-looking number; there isn't one. Same stance as gravel density.

Paver geometry is definitional: 144 sq in per sq ft, 27 cu ft per cu yd. A paver's face area is length x width / 144, so a 4 x 8 in paver is 0.2222 sq ft and covers 4.5 per sq ft.

Waste factors are a convention, labeled as such on the page, not a spec. 5% for a straight running bond, 10% for most jobs, 20% for herringbone or a diagonal lay. Inch Calculator states "at least 10% more pavers" and "at least 20% more" on a diagonal, which is the same family as the 10/15/20 convention the tile tool already uses.

Triangulated 2026-08-12 against Inch Calculator's paver calculator (inchcalculator.com ↗), which publishes the same chain: paver sq ft = length(in) x width(in) / 144, then pavers = patio sq ft / paver sq ft.

A discrepancy worth keeping, and it is in the guide. Their worked example rounds the paver area to 0.23 sq ft before dividing and gets 2,609 pavers for a 600 sq ft patio. Carrying the unrounded 0.2222 gives exactly 2,700. Rounding that intermediate costs 91 pavers, 3.4%, discovered at the end of the job. This tool divides by the unrounded figure. Same class of error as using a nominal 6 in picket instead of its real 5.5 in dressed width.

⚠ ICPI's own installation page is gone: icpi.org 301s to masonryandhardscapes.org after the merger into CMHA, and the deep link no longer resolves. The Pavestone specification above is a manufacturer's own copy of the same CSI section and reads cleanly, so it is what is cited.

bed

No new figures. Every constant is lifted unchanged from the topsoil and mulch tools so the three can never disagree: the raised-bed mix ratio from UMN Extension (see ## topsoil) and the 2 to 3 in mulch depth from UMass Extension and the University of Florida IFAS (see ## mulch). Those rows carry the URLs and the dates they were read and are not duplicated here, because a second copy of a citation is a second thing to keep in sync.

What this tool adds is that a bed is three purchases from one measurement, with two different depths over the same area: a fill depth for the soil and compost mix, and a separate mulch depth on top of it. Both are the reader's own entry; neither is invented here.

Bag sizes (0.75 / 1 / 1.5 / 2 cu ft for soil, 1.5 / 2 / 3 for mulch) are as sold, not specs. The 2 cu yd bags-versus-bulk crossover is a handling judgment, labeled as such, and is the same one topsoil and mulch use.

Verified 2026-08-12 by driving all three pages, not by reading them. Three beds were entered into /tools/bed/, /tools/topsoil/ and /tools/mulch/ and the bag counts compared: 8x4 at 10 in fill with 1/3 compost and 3 in mulch (18 topsoil, 9 compost, 4 mulch), 12x4 at 12 in with 1/2 compost and 2 in mulch (24, 24, 4), and 10x3 at 8 in with no compost and 4 in mulch on 1.5 and 3 cu ft bags (14, 0, 4). All three tools agree on every one, including the no-compost case. Re-run that cross-check if any of the three tools' constants change.