The Shape of a Regulated Document Is Data

Treating a filing's sections, numbering, and cross-reference grammar as versioned, governed rules rather than a per-document formatting exercise

This paper is provided for engineering, operational, and research purposes and does not constitute investment advice or a solicitation to invest in any Woodfine direct-hold solution. Statements marked "planned," "intended," "targeted," "may," or "expected" are forward-looking and subject to change. Full disclosures appear at the end of this paper.

Working Paper PDS-016 · v0.1.0 · CC BY 4.0

Our thesis is that the required shape of a regulated document should be recorded once, as governed data, and reused — not reproduced by hand every time a new document of that kind is needed. A prospectus has a shape. So does a set of statutory year-end financial statements, a subscription agreement, a trust deed, a schedule appended to any of them. Which sections must appear, in what order; how a heading is cased; whether clauses are numbered plainly or carry the word "Section" in front of them; how a clause inside the document refers to another clause, and how that differs from the way the same document refers to a rule in a statute. None of this is decoration. Much of it is constrained by law, by a regulator's filing requirements, or by drafting conventions old and settled enough that departing from them is itself a signal to a professional reader.

In ordinary practice that shape is reproduced by hand. A firm produces its next subscription agreement by opening its last one, or one it admired, and editing. The governing rules live partly in a style memorandum, partly in a senior practitioner's habits, and mostly in the previous document. We hold that this is the wrong place for them, and that the right place is the same place a design system already keeps the rule for what colour a warning is or how much space sits under a paragraph: a named entry in a file, with a value, a type, and a written description of when it applies.

This paper takes four things in turn. What a token actually is, from the beginning, because the argument does not work without it. What a regulated document's form is actually required to do, and by whom. How the platform's own print-and-document token set puts the two together, with one shared base and a named set of differences per document category. And last, honestly, what happens when a governed rule is broken anyway — including by us, in our own material — because a rule that cannot point at a real failure has not yet earned its keep.

1. The thesis

There are two ways to hold a rule about how a document must look.

The first is as prose in a style guide: "subscription agreements use flat numbered headings in bold title case, with a terminal period; do not use all capitals." That sentence is perfectly clear. It also has no mechanical relationship to any document. It governs by being read and remembered. When it is not read, or is read and misremembered, nothing announces the fact. The document that resulted looks confident and is wrong, and the only detection mechanism is a person who happens to know better reading it closely.

The second is as data: an entry with a name, a value, and a description, sitting in a file that a build tool reads. The rule is now addressable. It can be pointed at from more than one place. It can be changed in one place and take effect everywhere it applies. It can be compared against what a document actually did. It can be versioned, so that a document produced last year can be shown to have followed the rule as it stood last year rather than as it stands now — which matters a great deal when the question is asked by someone examining the document years after it was signed.

Our position is that the second way is correct for regulated documents specifically, and that the reason is not tidiness. It is that a regulated document's form carries consequence. A misnumbered clause is a broken cross-reference. A cross-reference that reads "Section 4" when the document's own divisions are called paragraphs is an ambiguity a counterparty's counsel will find. A heading scheme that changes halfway through a hundred-page instrument is the kind of thing that makes a careful reader wonder what else changed halfway through. Form is not separable from substance in a document whose whole purpose is to be precise.

The composition is the claim. Design systems have known for a decade how to hold a visual decision as data. Legal drafting has known for far longer what the conventions are. Financial reporting regulators have been pushing filings toward machine-readable structure for years. What is less common is treating the drafting conventions themselves — the heading form, the numbering scheme, the cross-reference grammar — as entries in the same governed file as the page margin and the rule weight, so that one mechanism answers both "how wide is the margin" and "does this family use the word Article."

2. The problem, in the reader's terms

Consider a law firm's precedent bank: the filing cabinet, physical or otherwise, holding the firm's best previous versions of each kind of document it produces. It is an enormously valuable thing. It is also, as a governance mechanism, exactly the weakness this paper is about.

A precedent bank holds documents, not rules. The rule that every internal cross-reference in a document must be capitalised the same way is not written anywhere in the bank; it is merely true of the good documents in it, and quietly untrue of one or two of them. When a junior drafter copies a document from the bank, they inherit whatever that particular document happened to do — the rules it followed, the rules it broke, and the accidents nobody noticed. Two documents copied from two different precedents diverge, and the divergence is invisible until someone reads both side by side, which nobody does.

The remedy every firm tries first is a style memorandum. It is the right instinct and it does not hold, for a reason that has nothing to do with discipline: the memorandum and the documents are two separate things maintained by two separate acts. The documents change constantly. The memorandum changes when somebody remembers. The gap between them widens on its own, without anyone deciding anything. This is the same failure that makes software documentation drift from software, and the remedy found there is structural rather than disciplinary — keep one source, and generate everything else from it, so that the copy which could drift does not exist as a separate thing to drift.

There is a second, sharper version of the problem, and it belongs to regulated documents in particular. A regulated document is read by people who are entitled to rely on it. An investor reading a prospectus, an auditor examining a set of statements, a securities regulator reviewing a continuous-disclosure filing — a filing an issuer is required to make on an ongoing basis after it becomes publicly accountable, rather than only once when it first raises money [ni-51-102] [bcsc-continuous-disclosure]. Each of them is reading for substance, and each of them uses form as a proxy for care. A document whose internal references are inconsistent invites the question of whether the diligence behind its numbers was any better. That inference may be unfair in a given case. It is not unreasonable as a general reading habit, and a firm that wants to be read generously has an interest in removing the grounds for it.

And the direction of regulatory travel makes the point for us. For listed issuers in the European Union, the annual financial report must now be filed in a structured electronic format in which the figures themselves carry machine-readable tags, not merely appear on a page [ixbrl-esef]. In the United States, the rules governing how certain regulated firms must preserve their records have for years specified properties of the record-keeping system — that records be preserved in a non-rewriteable form, indexed, and reproducible on demand — rather than merely instructing firms to keep good files [sec-17a-4-f]. The regulator's expectation has already moved from "produce a document that looks right" to "operate a mechanism whose output is right by construction." Our thesis applies the same move one layer further in, to the document's own structural grammar.

3. What a token is

This section establishes the vocabulary the rest of the paper needs, from the beginning. A reader already at home with design tokens may go to Section 4.

A decision written down as data

A token is a design decision recorded as data: a name, a value, and — in a mature system — a type and a written description of when to use it. That is the whole idea, and it is smaller than it sounds.

Take the most ordinary example. A company's primary colour is a particular blue. In the usual arrangement, that blue appears as a code — a short string of letters and numbers — typed directly into every stylesheet, template and slide master that needs it. The value is duplicated dozens of times, each copy silently independent of the others. When the blue changes, every copy must be found. Worse, nothing distinguishes the copies that meant "this is our primary colour" from the copies that happened to use the same blue for an unrelated reason, so a search-and-replace changes both.

Recording the decision as a token means the blue exists in exactly one place, under a name. Everything that needs it refers to the name. The name carries the intent: a stylesheet that says "the colour for things a person can act on" is making a different statement from one that says "the specific blue number." Change the one entry and every surface that meant the first thing follows; surfaces that meant something else are untouched, because they referred to a different name.

The description field is the part most easily overlooked and, for this paper's argument, the most important. In the platform's own published token set, a timing value is not merely a number of milliseconds; its description records what it is for — the quick beat of a button press or a focus outline fading. A spacing value records that it is the floor for the rhythm between paragraphs. The rule for using the value travels with the value. It is not in a separate guide that a person has to remember to open.

Three tiers, and why the middle one matters

The design-system field has converged on arranging tokens in three tiers, each answering a different question, and the platform's published set follows that convention rather than inventing one.

Primitive tokens answer "what values exist?" They are the raw inventory: a numbered scale of greys, a set of spacing steps, a handful of durations. A primitive has no opinion about where it is used. It is a swatch in a case, not an instruction.

Semantic tokens answer "what does this value mean here?" A semantic token points at a primitive and attaches a role — "the surface a page sits on," "the colour of text that is secondary to the main text," "the outline that shows which control has focus." This is the tier where meaning lives, and it is the tier that makes the whole arrangement portable: a different organisation adopting the system re-points the semantic names at its own primitives, and every document and screen follows, without a single component being rewritten.

Component tokens answer "what does this specific part use?" This button, this surface, nothing else. Scoping one more name to one component costs a line and buys the ability to change that one thing later without disturbing anything else.

The discipline that keeps the arrangement coherent is the direction of reference. Components refer to semantics; semantics refer to primitives; nothing refers the other way. A component that reaches past the middle tier to grab a raw value directly has re-embedded exactly the duplication the arrangement exists to remove.

One format, so the file is portable

For most of the last decade, every tool that handled tokens used its own file shape, which meant a token file was a private convention rather than something you could hand to someone else. That changed when the Design Tokens Community Group — a community group operating under the World Wide Web Consortium, the body that maintains the web's core standards — published the first stable version of its format, announced in October 2025 [dtcg-w3c]. The format is a plain, vendor-neutral text file: each entry declares its value, its type, and its description; entries nest into groups; and one entry can refer to another by name.

The practical effect is that a token file stops being a thing your tooling understands and becomes a thing anyone's tooling understands. That is what makes the next section possible. A document-structure rule recorded in this format is not a rule inside our software. It is a rule in a file, readable by a build tool, an auditor's script, or a person with a text editor.

4. What a regulated document's form is required to do

Before describing the mechanism, it is worth being concrete about what is actually being governed, because "document formatting" sounds like a matter of taste and this is not one.

The document families are real and specific

The platform's print-and-document token set covers a set of named document families, and they are not generic categories invented for tidiness. They are the real kinds of document the work produces.

A preliminary prospectus prepared under the Canadian securities rules for prospectus form. A statutory year-end financial statement set, prepared in the compilation register a Canadian accountant issues under a Notice to Reader — portrait, symmetric margins, pure black on white, because that is what the register looks like and departing from it signals something the document does not mean. A projection and proforma reporting layout, which is a deliberate sibling of the statutory set rather than the same thing: landscape, wide multi-period tables, a different density and a different tone, because a projection is not a statutory statement and should not be dressed as one. A subscription agreement in two variants that share a recipe and differ in their schedules. An engagement-letter suite with its memorandum of understanding and lettered schedules. Two Mexican trust-vehicle instruments — a trust deed and an offering document for a listed property trust, drafted in Spanish under civil-law notarial convention, governed by that country's income-tax and securities statutes and its securities regulator's registry. A print-safe corporate relationship diagram for the ownership chart such an offering document carries. And an interactive binder-navigation layer for assembled document sets.

Two of those families are worth dwelling on together, because their relationship is the whole argument in miniature. The Canadian prospectus family and the Mexican offering family do the same job in two jurisdictions. They are siblings and they are deliberately not the same: different language, different section taxonomy, no running header on one, noticeably narrower side margins on one because a fixed-width ownership diagram sets them, a different body type size matched to its companion trust deed. A single "prospectus template" covering both would be wrong in both. Five hundred separately hand-built documents would drift. What is wanted is a shared base and a named, deliberate, written list of the differences — which is precisely what a semantic tier is.

The shared base is genuinely shared

Underneath those families sits a primitive layer they all draw from, and the sharing is real rather than nominal.

The rule-weight ladder is four steps — a hairline, a light rule, a standard rule, and an emphasis rule — and it is identical across every legal and financial-statement family. A hairline draws the border of a key-terms table, the line under a running header, and a statutory subtotal. A standard rule draws a form cell, a signature line, and the rule above a grand total. Two families extend the ladder rather than redefining it: a prospectus data-table total row takes a doubled rule heavier than the ladder's top step, and an engagement-suite form note takes a heavier accent bar still. The extension is recorded as an extension. It is not a fifth private ladder.

The typography rule is one sentence and it is a real drafting protection: every document family pairs a serif reading face for body and heading text with a different sans face reserved for form-fill zones. A blank line a person is meant to write on should never be mistaken for printed text they are meant to rely on. That is not a preference. That is the difference between a subscriber filling in a blank and a subscriber believing a term was already agreed.

The page geometry is per family and real: bound margins wider on the binding edge, a reduced top margin on a cover page, a landscape page for the wide projection tables. These are production values measured from documents that were actually produced, not defaults carried in from a template.

5. How a document becomes a token contract

The structural values above — margins, rule weights, type sizes — are the easy half. They are measurements, and a design system has always known how to hold a measurement. The harder and more interesting half is the part that has no number.

One shared base and five named differences

Alongside the concrete per-document families, the token set carries a cross-cutting register for legal drafting conventions. Its structure is one shared base plus five per-category variants: a commercial agreement, a constitutional agreement, a schedule or exhibit, a letter, and a preliminary instrument such as a memorandum of understanding or a term sheet.

What it governs is heading form, numbering, and cross-reference grammar — the three things a precedent bank transmits by accident. Each cell in the register is marked with how firmly it binds: some rules never vary, some are the strong default drawn from the weight of filed precedent, and some are an accepted house alternative where genuine practice differs. That grading is itself the useful part. It distinguishes "this is how it is done" from "this is how we have chosen to do it," which a style memorandum almost never does.

The content is specific. A commercial agreement takes a flat numbered heading — the bare number, a bold title-case title, a closing period, left-aligned, not capitals. It does not put the word "Section" in front of the number; putting it there is the house alternative, not the default. Its internal cross-references are capitalised and its references to outside instruments are not, so that a reference to a clause in this agreement is visibly a different kind of thing from a reference to a rule in a securities instrument. A constitutional agreement — a shareholders' agreement, a partnership agreement, a constating document — uses a centred article line in capitals with a title-case sub-level beneath it, and the two levels are always distinguishable by case; that one is a rule that never varies. A schedule's own label is its heading, and whichever word the document chose — schedule, exhibit, or annex — it holds that word throughout, without exception. A letter matches its cross-reference word to whatever its divisions are actually called: sections if it has headed sections, paragraphs if it has bare numbered paragraphs.

Then there are the rules that apply across the whole family, and these are the ones a style memorandum tends to leave out because nobody thinks to write them down. Never a lone sub-clause with no sibling — a section with only one block stays bare under its heading rather than acquiring a decimal it does not need. Subdivide only where two or more parallel provisions actually exist; uniform depth throughout a document is not required and forcing it produces the orphan the previous rule prohibits. Keep the "of such-and-such instrument" qualifier on every external reference, so a clause numbered the same inside and outside the document cannot be confused. An unnumbered lead-in paragraph under a subdivided section is legitimate and is cited as the whole section. Do not formally define the words "Section" and "Article" in the definitions clause; capitalise them and carry a short construction clause instead. And one rule marked as binding absolutely across the entire family, which is notable precisely because it is so small: never two internal cross-reference pointers in different case within the same document.

The rule that exists because it was broken

One rule in that register deserves its own paragraph, because of how it got there.

The token set also carries a number-display rule for financial reporting output — decimal precision, whether to round or truncate, how to scale a per-unit figure against an aggregate. It is a sensible rule for a financial report, where consistent presentation of figures is the whole point of the register.

It was applied to a document that was not a financial report. It was a corporate resolution: an instrument whose operative content includes a figure the parties are agreeing upon. The rule did exactly what it was written to do, and in doing so it changed that figure — in every place the figure appeared. Nobody decided to change a number. The renderer changed it, correctly, according to a rule written for a different kind of document.

The answer was not a reminder to be careful. The answer was a new rule, binding across all five legal variants without exception: every operative figure in a legal instrument — a stated price, a consideration amount, a share count, any number the document resolves upon — renders exactly as it appears in the source, with no rounding, no truncation, and no rescaling applied by the machinery. The scope of the financial-reporting rule was narrowed in the same change, so that it can no longer reach a signable instrument at all. That correction is in the published token file, with its reasoning, where anyone can read it.

This is what it looks like when a rule is data rather than prose. The defect was a rule applied outside its scope; the fix was a change to the scope, recorded once, in the place the machinery actually reads. In a style memorandum the same fix is a sentence somebody must remember.

What is honestly not finished

Most of the values in the legal drafting register are placeholders. The structure is decided — the shared base, the five variants, the qualitative lock table — and the real measured values for the categories are being delivered a category at a time, commercial agreements first, from a reference implementation. The token file says so on its face. The concrete per-document families beside it — the subscription agreement, the prospectus, the engagement suite — already ship real production stylesheets; the cross-cutting drafting register does not yet.

6. When the rule itself is broken

A rule earns its keep when it can point at a real failure. Until then it is a plausible opinion, and a platform arguing for governed rules ought to say so about its own.

The failure described above is the strongest version of that test we can offer within this subject: a real rule, correctly written for its own domain, causing a real problem by being applied outside it. But there is a second and more uncomfortable version, and leaving it out would make this paper less honest than the token file it describes.

The platform's published rules include a vocabulary register — a list of terms retired from public writing, each with the reason and each with a stated plain-language substitute. One entry on that list is a word that carries a completely ordinary meaning in finance, describing growth that accumulates on itself, and a second, unrelated meaning in this platform's own technical usage. The rule does not ban the word outright. It requires that where the technical sense is meant, a plain-language translation precede it on first use, so that a reader arriving with the financial meaning in mind is not silently handed the other one. That is a good rule, written for exactly the reader this corpus is written for, and it is published as data alongside the colours and the margins.

Here is the honest limit. Publishing a rule as data is not the same as enforcing it. The platform's structural token values have a real, automatic check: a gate in the repository regenerates the derived token files whenever a source file changes, stages the result into the same change, and refuses the change outright if regeneration fails — so the exported file and its sources cannot silently drift apart. The qualitative rules sitting in the same token set have no equivalent. They are strings, not measurements. Nothing today reads a finished document and reports that its cross-references are inconsistently cased, or that a retired term appeared without its translation. That check is intended and it is not built, and until it is, conformance rests on a person reading carefully — which is the mechanism this paper opened by criticising.

We would rather state that plainly than let the preceding sections imply a completeness that does not exist. The structure is right, the structural half is enforced, and the qualitative half is currently a well-written rulebook in a filing cabinet that happens to be machine-readable.

7. What this changes for the reader

The first change is what "consistent" can mean. Today a firm claiming its documents follow a consistent house standard is asking to be taken at its word, because the standard exists as prose and the documents exist separately. Under this arrangement the standard exists as a file. Two documents of the same family either drew on the same entries or they did not, and that is a question with an answer rather than an impression.

The second change is for whoever examines the documents later. An auditor — an independent professional who examines a business's records and states whether they can be relied upon — or a buyer's counsel reading a stack of agreements in diligence, can be told not merely "these follow our house style" but "these were produced from this version of this rule set, and here is the rule set." The rules are versioned, so a document produced two years ago can be assessed against the rules as they stood two years ago rather than against today's. That is a materially different conversation from producing four documents and inviting someone to notice they match.

The third change is about correcting a mistake. When a drafting convention turns out to be wrong — as the number-display scoping was wrong — the fix is a change to one entry, with a written reason attached, and it applies to everything produced afterward. It is not a circulated note that some people read.

The trade-offs are real and belong in the same paragraph. Encoding a convention requires first deciding what it is, and legal drafting practice genuinely differs between jurisdictions and between good firms; the register handles this by marking what is settled and what is a house choice, but marking a disagreement is not resolving it. A rule held as data is enforced only where a machine actually checks it, and today half of this rule set has no such check. Placeholder values are placeholders. And there is a failure mode specific to this approach: a rule correct in one domain, applied by machinery to a document in another, does the wrong thing faster and more uniformly than a careless person would. We have seen that happen once, which is why the scope of a rule is now treated as part of the rule rather than as context a reader supplies.

8. An open invitation

This paper states a position we hold and a set of problems we cannot settle from inside one firm.

To researchers in legal informatics and computational contract drafting: the register described in Section 5 grades each convention by how firmly it binds — never varies, strong default, accepted alternative. We derived those gradings from the weight of filed precedent and from standard drafting references, which is a defensible method and not a rigorous one. Whether the gradings survive a proper corpus study of filed instruments, and whether the three-way grading is even the right shape for this kind of rule, is a question we would rather have answered than assumed.

To regulatory-technology researchers: the structured-filing requirements now in force for listed issuers in several jurisdictions govern the tagging of figures. The register in this paper governs the structure — headings, numbering, references. Whether these two should eventually be one mechanism, so that a document's structural grammar and its tagged content are governed together rather than by unrelated tools, is an open design question with real consequences for anyone who has to produce both.

To practitioners of comparative drafting practice: this register has been tested against two jurisdictions in depth and touches a third. Whether a shared base plus per-category differences holds up across jurisdictions whose formatting requirements diverge more sharply than these — or whether it quietly becomes a base that is shared by nobody and five variants that are really five separate systems — is the question we would most like a specialist to break.

And to anyone who has built a conformance check for qualitative drafting rules at scale: we would be glad to learn what actually works. We know how to check a margin. We do not yet have a reliable way to check, across a long document, that internal cross-references are consistently cased and that every retired term arrived with its translation, without producing so many false reports that people stop reading them.

9. Conclusion

The required shape of a regulated document should be governed, versioned data, decided once and reused — not re-decided document by document out of whichever previous document was nearest to hand. A prospectus, a statutory statement set, an agreement and a schedule each have a real form with real consequences, and that form is exactly the kind of thing a design system has spent a decade learning to hold as named entries in a file rather than as habits in a precedent bank.

The platform's print-and-document token set is built that way: a shared base of geometry, rule weights and type pairings; a set of real document families selecting from it; and one cross-cutting legal drafting register of a single shared base and five named variants, covering heading form, numbering, and cross-reference grammar. Most of that register's values are still placeholders, and the qualitative rules in it have no automatic check today, which is why a rule in it was able to be applied outside its scope and alter an operative figure before anyone noticed. We publish that rather than omit it, because a paper arguing that governed rules are better than remembered ones has no business implying its own are already finished.

References

Design Tokens Community Group. Design Tokens Format Module. World Wide Web Consortium Community Group. https://tr.designtokens.org/format/

British Columbia Securities Commission. National Instrument 51-102 — Continuous Disclosure Obligations. https://www.bcsc.bc.ca/securities-law/law-and-policy/instruments-and-policies/5-ongoing-requirements-for-issuers-insiders/current/51-102

British Columbia Securities Commission. Continuous Disclosure Obligations — overview. https://www.bcsc.bc.ca/industry/issuer-regulation/continuous-disclosure-obligations

Canadian Securities Administrators. National Policy 51-201 — Disclosure Standards for Forward-Looking Information.

European Securities and Markets Authority. European Single Electronic Format (ESEF) — inline XBRL requirements for annual financial reports. https://www.esma.europa.eu/policy-activities/corporate-disclosure/european-single-electronic-format

United States Securities and Exchange Commission. Rule 17a-4(f) — Records to be Preserved by Certain Exchange Members, Brokers and Dealers. https://www.ecfr.gov/current/title-17/chapter-II/part-240/section-240.17a-4

Contributors

Prepared by Woodfine Management Corp.

How this paper was produced

This paper is grounded in the operating and development work the preparing staff do themselves, and in what they have learned from the people they work with routinely: the graphic designers, web developers, and software developers and engineers who build the platform alongside them, and the architects and structural, building-services, and civil engineers they develop buildings with. No outside professional reviewed or approved this paper, and nothing in it is professional advice. It was drafted and edited with AI assistance under editorial direction, and the analysis and conclusions are Woodfine's own.

Disclosures

Woodfine Capital Projects Inc. ("Woodfine") is the author of record. PointSav Digital Systems, which builds the platform described, is currently a trade name of Woodfine, planned to become a wholly-owned Woodfine subsidiary upon incorporation; PointSav does not itself offer, sell, or solicit any security. The token set described in this paper is our own work, so this paper argues for an approach we have a commercial interest in. This work was funded internally; no external research funding was received. The descriptions of drafting conventions, filing requirements, and disclosure obligations are given in general terms to explain a design and are not legal advice on any party's obligations under any instrument or in any jurisdiction. This paper's content is provided for engineering, operational, and research purposes and does not constitute investment advice or a solicitation to invest in any Woodfine direct-hold solution. Some statements above describe planned or intended future work; in particular, the conformance check described in Section 6 as intended is not built, and most values in the drafting register described in Section 5 are provisional. Language such as "planned," "intended," "targeted," "may," and "expected" marks this forward-looking content, which is subject to change and does not constitute a commitment regarding future performance.

Data and reproducibility

Everything this paper reports about the token set was read directly from the published files in preparing it, rather than from a summary. The print-and-document token set is published as ordinary text files in a public repository under a permissive licence, in the standard interchange format described in Section 3, which means a reader can open them without any of our software. The named document families, the four-step rule weight ladder and its two extensions, the reading-face and form-fill-face pairing rule, the shared base and five variants of the legal drafting register, its per-category heading and cross-reference rules, the family-wide drafting rules, and the numeric-precision rule described in Section 5 are all present in those files with their own written descriptions attached. The correction described in Section 5 — the narrowing of a number-display rule so that it can no longer reach a signable instrument — is recorded in the file itself, including its reasoning; the document involved, and the figure it affected, are not described here and are not in the published file either. The automatic regeneration gate described in Section 6 is a short script in the same public repository. The document families' real production stylesheets are referenced from the token entries that describe them. No independent party has audited this token set, reviewed these drafting conventions, or verified these claims.

PointSav Digital Systems™ and Woodfine Capital Projects™ are trademarks of Woodfine Capital Projects Inc.