What we publish
JoistLedger publishes calculated results and the reasoning behind them. Every figure traces back to a published source:
- IRC Table R507.5 for joist spans, with all four of its assumptions stated on every page
- IRC R507.5.1 for the cantilever limit of one quarter of the actual span
- IRC R507.9 for ledger attachment, which is what actually fails when decks collapse
Where a figure is a convention or a rule of thumb rather than a measurement, the page says so in those words.
How the calculations are made
The arithmetic lives in code, not in a spreadsheet and not typed into a page. That matters for two reasons: a mistake in a formula is a mistake on every page it touches, and a formula in code can be tested against every combination of inputs it will ever be asked for.
Before a set of pages goes live, every combination it can produce is generated and checked — for impossible values, for missing numbers, for answers that contradict the text around them. Pages that fail do not publish.
What we will not publish
We do not design decks carrying hot tubs, roofs or unusual loads, and we do not replace a permit. Local codes amend the IRC and high-snow jurisdictions use higher live loads.
We also do not publish anything we cannot source. If the honest answer is "it depends on something we cannot see from here", the page says that instead of producing a number.
Corrections
If a number here is wrong, tell us at [email protected]. We would rather hear it from you than not hear it.
Corrections to a formula are fixed at the source, which means they propagate to every page that used it. Corrections to a source — a datasheet revision, a code update — are made with a note recording what changed.
Advertising
The site carries display advertising. Advertisers have no involvement in what is published, no advance sight of it and no ability to influence it. There are no sponsored placements, no paid links and no affiliate relationships.
AI and how this is written
Explanatory text on this site is drafted with AI assistance and edited before publication. The numbers are not: they are computed from the sources listed above by code that is tested, and no figure on this site is generated by a language model.
We think the distinction matters more than a blanket claim either way, so we state it plainly rather than leaving it to be guessed at.