By Bill Hartzer · Published
The book says, in its front matter, that it would rather be publicly corrected than quietly out of date, and this is the page where that commitment is either honored or exposed. Every correction to the printed volume appears here with the printing, the page, and the date, whether or not it makes the next impression.
A book about how professionals should handle the record of their own work is in a poor position to be casual about the record of its own errors. So this page exists from the day of publication rather than from the day of the first correction, and it will say none recorded for exactly as long as that is true.
Corrections
None recorded. The first edition has not yet been published.
Superseded material
None recorded.
Reporting an error
Write to record@theexpertrecord.com. Please say which printing and which page. Corrections are posted here with the date, whether or not they make the next printing, and substantive ones are credited unless you would rather they were not.
How corrections are posted
Each entry states the printing, the page, the text as printed, the text as corrected, and the date the correction was posted. Where a correction is substantive rather than typographic it also states what led to it, because a reader deciding how much weight to give the surrounding passage is entitled to know whether the error was a slip or a mistake in the reasoning.
Corrections are not removed once the printing that contains the error goes out of circulation. Copies stay in circulation for years and a reader holding the third printing needs to be able to find the correction that was made in the fourth.
What counts as an erratum
A factual error, a misattributed quotation, a broken cross-reference, or a statement that has been overtaken by events in a way the book's own framing does not already cover. The book deliberately omits material that moves — platform behavior, the status of decisions, rulemaking — and none of that is an erratum. It is maintained elsewhere on this site, which is the whole design.
Two categories that are not errata come up often enough to name. A reader disagreeing with the argument is not an erratum; disagreement is what the argument is for, and it is welcome by email but it does not belong on this page. And a development in the field that the book anticipated in general terms without naming is not an erratum either — the book was written to describe the shape of a problem rather than its current instance, and a new instance appearing is the framework working, not failing.
Changes to the Standard itself
The eight clauses are versioned separately from the book and from this site. Version 1.0 is the text printed in the interleaf and published on the Standard page, and it will not be edited. If experience shows a clause needs to change, the changed text is issued as version 1.1 or 2.0 and both versions stay published, permanently, so that an engagement letter citing a version keeps meaning what it meant on the day it was signed.
Any new version is announced here as well as on the Standard page, with a note of what changed and why. A standard whose text moves without notice is not a standard, and the whole value of naming this one is that two parties can agree to it and know what they agreed to.
Revisions to the working papers
The working papers are templates rather than text, and they are expected to improve. Substantive revisions — a new provision in the engagement letter set, a category added to the verification checklist, a corrected specification — are noted here with the date, so that a practitioner working from an earlier copy can tell whether they are missing anything.
Cosmetic and formatting changes are not noted. Neither are additions to the runnable scripts, which are maintained continuously against current language versions and carry their own notes.
Why a book was written this way at all
Most practitioner books about a moving technology fail in one of two directions. They fix the state of the world at the moment of writing, and become historical documents inside two years. Or they stay so general that they describe nothing a reader can do on Monday morning. The way out of that is not better prediction — it is deciding, sentence by sentence, which statements belong to the durable layer and which belong to the dated one, and then physically separating them.
So the book carries only the first kind: the eight clauses, the task taxonomy, the instrument-and-source distinction, the verification pass, the log, the deposition answer, and the arguments that support them. None of that turns on which systems exist or on what any tribunal decided last quarter. This site carries the second kind, where a date can be attached to it honestly and a correction can be published the same afternoon.
The cost of that split is that a reader has to look in two places, and the book says so in its front matter rather than pretending otherwise. The benefit is that neither half has to apologize for the other — the book does not go stale, and the site does not have to be right forever.