DEV Community: ARLing

Free Peppol invoice validators compared: what each one checks

ARLing — Sat, 10 Oct 2026 08:30:00 +0000

Originally published at https://arling.sk/notes/free-peppol-invoice-validators-compared/

Free Peppol invoice validators compared: what each one checks

There is no single best free Peppol validator, but there is a clear rule: for the official verdict you need the current Peppol BIS Billing 3.0 schematron plus the UBL 2.1 XSD schema; the KoSIT Validator has a Peppol configuration for release 3.0.21 and ecosio's free validator checks schema and schematron, although we could not verify which release it runs live. The European Commission's free validator checks EN 16931 only, not the Peppol or XRechnung rules. Our own check at arling.sk is free, runs in your browser and explains every finding in a sentence, but it does not check the XSD schema or the UBL-CR rules, and on 25 September 2026 its Peppol rules were written from release 3.0.20, not the mandatory 3.0.21. Below is a side by side table of seven free tools, what "valid" means under Peppol, and our own test in which three broken invoices passed our checker without a single finding. We also say where each tool sends your file, because some keep it on a server for a while and some never see it. We read every page on 25 or 26 September 2026 and did not test the output of the other tools.

Seven free validators side by side

Validator	Where your file goes	XSD schema	Rules it runs
European Commission eInvoice Validator	uploaded to the Commission service, also via REST and SOAP API	not verified	EN 16931 release 1.3.16 for UBL invoice, CII invoice and UBL credit note; no Peppol, no XRechnung
ecosio, Peppol and XML validator	uploaded, processed non-persistently and deleted after the result	yes	Peppol BIS 3.0, Peppol PINT, UBL, CII, XRechnung; live rule release not verified by us
KoSIT Validator	stays on your computer, Java 11 or later, Apache 2.0	yes	XRechnung configuration, and a separate Peppol BIS configuration for release 3.0.21
ValidateFin UBL validator	stays in your browser	yes	EN 16931 and Peppol BIS 3.0, around 200 rules by its own count
Peppol Validator, peppolvalidator.com	uploaded, files and results kept for a limited time for shareable links	not verified	311+ rules by its own count, UBL and CII
efaktura.tools	stays in your browser by its own description	not verified	UBL 2.1, EN16931, Peppol BIS 3 by its own description; we could not see details without a browser
ARLing e-invoice check	stays in your browser	no	332 rule codes from EN 16931, Peppol BIS Billing 3.0 (written from release 3.0.20) and XRechnung 3.0, plus own checks such as IBAN

If the invoice data is confidential, the column that matters most is the second one: tools that run in the browser or on your own computer never receive the file.

What "valid" means for a Peppol invoice

Peppol's own rule is short: "A message is considered compliant to this BIS if no fatal errors are detected when validating with the current validation artefacts.". Warnings do not make an invoice non-compliant, a receiver must not reject a compliant invoice and may reject a non-compliant one. Validation has two layers: the XML schema checks which elements occur where, how often and in which order, and the schematron checks meaning, conditional requirements and calculations.

The word "current" is where free tools differ. Release 3.0.21 was published on 20 May 2026 and is mandatory from 17 August 2026; among other changes it removed rule BR-CO-25 and approved the profile urn:peppol:bis:billing_with_response. In release 3.0.21 the EN 16931 schematron for UBL has 979 checks, 281 of them fatal, and 676 of its warnings are UBL-CR rules about elements the core invoice does not use.

What we measured ourselves (25 September 2026)

We changed one thing at a time in the English sample invoice of our own checker and ran eleven variants. Eight were caught, three passed without a finding, and the release gap showed up as well.

Variant of the sample invoice	Our result	What the official rules say
Total with VAT one cent too high	BR-CO-15 and BR-CO-16	both fatal
Currency code EURO instead of EUR	BR-CL-04, BR-CO-15, eleven times PEPPOL-EN16931-R051	among others the fatal PEPPOL-EN16931-R051
Typo in the IBAN	own check ARL-IBAN	not checked by the Peppol schematrons for a Slovak invoice
IssueDate and DueDate in the wrong order	no finding	XML schema error
Extra ProfileExecutionID element	no finding	warning UBL-CR-003
CustomizationID shortened to urn:cen.eu:en16931:2017	no finding, Peppol rules skipped	fatal PEPPOL-EN16931-R004
Profile urn:peppol:bis:billing_with_response	error PEPPOL-EN16931-R007	approved in release 3.0.21

The last four rows are our weaknesses, and we would rather you read them here than find them out at your buyer's access point. The IBAN row is the other side: the official Peppol schematrons check an IBAN checksum only in two German rules, and they do not check that a Slovak 0245 identifier is followed by a ten digit tax number, while our own checks do.

We also ran ten XRechnung files through the KoSIT Validator 1.6.3 and through our checker. KoSIT reported an XML schema error for an invoice without an invoice number, where we reported rule BR-02, and a BR-DE-27 warning for a missing phone number that we did not report; we reported a typo in the IBAN as an error and a Leitweg-ID without check digits as a warning, which KoSIT accepted.

When you do not need us

If you only need the final yes or no before sending, use a validator that runs the current release with the XSD schema, for example the KoSIT Validator with the Peppol BIS configuration for release 3.0.21 on your own computer, or ecosio in the browser after checking which release it states. If you need the European standard alone, for example for a CII invoice, the Commission's validator is the reference, but it does not run Peppol or XRechnung rules. Your access point provider delivers the invoice into Peppol; we do not deliver into the Peppol network at all.

Our checker is useful when you want a plain sentence and the path to the element for every finding, IBAN and Slovak tax number checks, and a preview of the invoice as a readable document, all free and without your file leaving the browser. The only paid part is downloading an invoice you build in our form: 2.90 € excluding VAT for one invoice or 9.90 € excluding VAT for 30 days.

FAQ

Is an invoice with warnings but no fatal errors compliant?

Yes. Peppol BIS Billing 3.0 treats a message as compliant when no fatal errors are found with the current validation artefacts, even if warnings are reported, and a receiver must not reject a compliant message.

Does the European Commission validator check Peppol rules?

No. On 25 and 26 September 2026 it offered only UBL Invoice XML, CII Invoice XML and UBL Credit Note XML against EN 16931 release 1.3.16; there was no validation type for Peppol BIS Billing 3.0 or XRechnung.

Does ARLing send my invoice into the Peppol network?

No. The XML is created and checked in your browser and is never sent to our server; delivery into the Peppol network is done by your access point provider and is not part of our offer.

Conclusion

Check your UBL invoice for free, read every finding in plain language and build the XML if you need to, in our e-invoice checker and generator; for the final verdict before sending, run the file through a validator with the current Peppol release and the XSD schema as well.

ARLing also has a free e-invoicing tool.

Puzzles for reMarkable and Kindle Scribe: what to check

ARLing — Fri, 09 Oct 2026 08:30:01 +0000

Originally published at https://arling.sk/notes/puzzles-for-remarkable-and-kindle-scribe/

Puzzles for reMarkable and Kindle Scribe: what to check

Puzzles for a reMarkable or a Kindle Scribe are PDF files. You copy one to the tablet and write on the page with the pen. Before you pay, check three things. The page should be 3:4. The links should answer your finger, including the ones near the top edge. The seller should say how each solution was checked. A free sample on your own tablet tests the first two.

On a Kindle Scribe, send the PDF with conversion off, because conversion removes the links.

This note covers where these puzzles come from, which kinds suit a pen, six checks and the way onto each device. It ends with what we measured in our own ten e-ink books on 25 September 2026. One result counts against them. Their Solution and Contents links sit 8.2 mm below the top edge and are only 3.7 mm tall, which is small for a finger. That is also the edge where a planner maker collected reports of links that stopped working on the Scribe.

Where puzzles for these tablets come from

Write-on books in the Kindle store. These are for the Scribe only. After an update in May 2023, Amazon began selling books you write on directly, most of them crosswords and sudoku. The eBook Reader's roundup links to the collection.
Free sudoku to print. Krazydad has thousands of free sudoku in PDF booklets of eight puzzles, with hints and answers. Pick the layout with one puzzle to a page.
Free crosswords. The Guardian has thousands of free crosswords, both quick and cryptic.
The catch with paper files. Pages cut for A4 or US Letter do not have the 3:4 shape of the screen. Shown whole on a reMarkable 2, an A4 page is drawn at 71 percent of its size, with a white band down each side. Why an A4 PDF does not fill an e-ink screen has the arithmetic.
Free packs made for the screen. Pixel Wizard Press gives away 13 packs of 10 linked puzzles for the reMarkable. Each pack has a second PDF for the 9:16 Paper Pro Move.
Paid e-ink books. Most are digital downloads on Etsy. Pixel Wizard Press sells editions of 100 puzzles there at 5.99 USD each. Ours are listed with prices at the end of this note.
Daily puzzles in a browser. Our daily logic games are free and need no account. They cover the same ten kinds as our e-ink books, so you can find the kind you like before you buy anything.

Which kinds suit a pen

The write-on books in the Kindle store are mostly crosswords and sudoku. Our e-ink books hold grid logic puzzles. In nine of their ten kinds, the pen does one of three jobs:

A digit or a mark in a cell. Killer Sudoku, Anti-knight Sudoku, Kakuro, Nonograms, Star Battle, Nurikabe and Numberlink work this way.
A loop through the cell centres. That is Masyu.
Lines between dots. That is Slitherlink, where the grid of cells is not even drawn.

Cells you write in with a pen need room. Our e-ink pages allow cells up to 30 mm, against 22 mm in print, and keep one puzzle to a page.

Six things to check before you pay

A free sample on your own tablet answers the first three at once. Try its links on a computer first, then with a finger on the tablet. Our note on testing PDF links on your e-ink tablet goes through it step by step.

The page is 3:4. The reMarkable 2, the Paper Pro and the Kindle Scribe all have 3:4 screens, so a 157 x 210 mm page fills each of them. The Paper Pro Move is 9:16 and needs a file of its own.
The links near the top edge work on your tablet. Planned Pages, a planner maker, collected reports from Kindle Scribe owners. After a firmware update, links near the top and right edges stopped working. A margin of about an inch helped. The reMarkable keeps roughly the outer 13.5 mm for its tool bar and the bottom edge for quick browsing. Boox keeps the corners for page turns.
The links are big enough for a finger. On the Scribe most of the screen turns the page. A near miss on a small link turns the page instead of jumping. For scale, the buttons in our Puzzle Post are 9.0 mm tall. The top links in our e-ink books are 3.7 mm.
The seller says how each solution was checked. Looking at a grid will not tell you whether it has exactly one solution. What you can check is whether the seller says how they made sure. Our books keep a puzzle only when our solver finds exactly one solution, and that solution is printed at the back.
There is a free sample. Pixel Wizard Press gives whole packs away. Our e-ink book sample is ten pages with 14 links. The Puzzle Post sample is 28 pages with 91 links, Solution links included.
The file suits your transfer route. The reMarkable takes files up to 100 MB over USB and through its web app. None of its apps lists ZIP, so unzip a download before you import it.

Getting the PDF onto the tablet

reMarkable. Click Import in the desktop app, or drag the file into its window. In the mobile app, tap Import. In the web app at app.remarkable.com, drag the file into the list. Over a USB cable, drag it into the web interface.
Kindle Scribe. Use Send to Kindle with conversion off. Converting turns the PDF into Kindle's reflowable format and destroys the fixed layout and the internal links.
Boox and Supernote. On a Boox, use a USB cable and the device folder. On a Supernote, use its upload page.
Tapping. Links answer a finger, not the pen. The reMarkable needs a finger tap, the Scribe a finger tap or the pointer tool.
Writing. You write over the page with the pen, so a puzzle PDF needs no form fields. Our e-ink books have none.

If a link does nothing, open the same file on a computer. If it works there and only links near the edges fail on the tablet, suspect the gesture zones of the device, not the file.

What we measured in our own e-ink books

On 25 September 2026 we opened the ten PDFs of our e-ink bundle with PyMuPDF and counted what is in them. What the tappable contents page does for a reader is in our note on hyperlinked puzzle PDFs.

Size and content. Each book is 379 pages of 157 x 209.9 mm. It holds 300 puzzles of one kind: 100 easy, 100 medium and 100 hard.
Links. Each book has 1200 internal links, and every one points to a page in the same file. There are 0 external links and 0 form fields.
On each screen. The page shows at 100 percent on the reMarkable 2 and the Kindle Scribe, and at 114 percent on the Paper Pro. On the Paper Pro Move it shrinks to 58 percent, with 40.5 mm of the screen height left empty. We have no Move layout.
Where the links sit. Solution and Contents start 8.2 mm below the top edge and are 3.7 mm tall. Contents ends 8.0 mm from the right edge. In nine books, 600 of the 1200 links start within 25.4 mm of the top edge. In the Hashi book it is 652.
Our Puzzle Planner. It has the same pattern. Its Day, Week, Month, Puzzles and Solution links sit 7.9 mm below the top edge and are 3.7 mm tall. In all, 2343 of its 4117 links start within 25.4 mm of the top edge.

By our own checklist, the books and the planner fall short on points 2 and 3. A 3.7 mm link is small for a finger, and these sit inside the inch that the Planned Pages reports point to.

Puzzle Post was built later, with a bottom bar. Its buttons are 9.0 mm tall and 15.9 mm above the bottom edge. None of its 598 links starts within 25.4 mm of the top or right edge. The books and the planner have not been rebuilt that way yet.

Nobody at ARLing has opened these files on a reMarkable, a Kindle Scribe, a Boox or a Supernote, because none of us owns one. The counts describe the files. Only your tablet can show how it treats a tap near the top edge.

FAQ

Can I write on a puzzle PDF on the Kindle Scribe?

Yes. A PDF sent with Send to Kindle takes the pen. After update 5.19.2, The eBook Reader and Good e-Reader reported in March 2026 that PDFs copied over USB take it too. Amazon's documentation still said otherwise at the time. What you write stays on the Scribe and does not show in the file on a computer.

Why does nothing happen when I tap a link on my reMarkable?

The reMarkable follows links on a finger tap, not on the pen. If a finger tap still fails, try the same link on a computer. If it works there and only links near the edges fail on the tablet, the edge belongs to the tool bar or to quick browsing.

Will a puzzle book made for the reMarkable 2 work on the Paper Pro Move?

It opens, but the 9:16 Move draws a 3:4 page at 58 percent of its size. That leaves 40.5 mm of the screen height empty, so grids and links get small. Look for a separate Move file. Our e-ink books do not have one.

What to buy, and how to try it first

Of our own files, Puzzle Post is the one laid out for checks 2 and 3. Its 9.0 mm buttons sit in a bottom bar, away from the top edge. Issue 1 has 100 puzzles in ten kinds. A subscription is 3.90 € a month or 29 € a year, and Issue 1 alone is 4.90 €.

Our ten e-ink books hold 300 puzzles each, one kind per book. The kinds are Killer Sudoku, Hashi, Anti-knight Sudoku, Star Battle, Numberlink, Nonograms, Slitherlink, Kakuro, Masyu and Nurikabe. They cost 4.90 € each in our Etsy shop, or 12.90 € for all ten on the e-ink bundle page. The Puzzle Planner puts one of 372 puzzles on each day page of an undated year, for 9.90 € on Etsy. Both keep their links at the top, so tap those first.

Try before you pay. The free e-ink book sample is the first ten pages of the Star Battle book, with no account and no e-mail. Put it on your tablet and tap Contents at the top of a puzzle page with a finger. The Puzzle Post sample has 10 puzzles and the bottom bar, Solution links included. Prefer paper? The printable books for A4 and US Letter are on our puzzle books page, at 4.90 € each or 19.90 € for all ten.

ARLing also publishes printable puzzle books.

Peppol Invoice Validator: What It Checks and What It Misses

ARLing — Thu, 08 Oct 2026 08:30:01 +0000

Originally published at https://arling.sk/notes/peppol-invoice-validator-what-it-checks/

Peppol Invoice Validator: What It Checks and What It Misses

A Peppol invoice validator should run three layers: the UBL 2.1 XML Schema, the EN 16931 rules and the Peppol BIS Billing 3.0 rules. An invoice is compliant when no fatal error comes back, and warnings alone do not change that. Before you trust a clean result, check which profile the validator detected and which release of the rules it runs. We broke one sample invoice eleven ways: our own browser check flagged eight, one of them only with a warning where Peppol has a fatal error, and it missed a swapped element order, an extra ProfileExecutionID and a shortened CustomizationID.

This matters to anyone who sends invoices over Peppol, and in Slovakia it becomes routine: e-invoices between domestic VAT payers are mandatory from 1 January 2027 and travel over the Peppol network through certified providers. Slovakia uses Peppol BIS Billing 3.0 directly, with no CIUS of its own. Below you will find what each layer checks, how to read fatal and warning, every result of our test with the free ARLing Peppol validator, and which free validators to use, including where ours falls short.

Three layers, three different questions

The first layer is the UBL 2.1 XML Schema (XSD), and it checks structure only: which elements occur where, how often and in which order. The free validator from ecosio lists it as the second of its three stages: well-formed XML, the XML Schema, then the Schematron rules. Our browser check skips this layer, which is why a swapped pair of dates went through it unnoticed in the test below.

The second layer is EN 16931, the European core invoice model, with its BR rules and the BR-CO rules for totals and consistency. The third is Peppol BIS Billing 3.0, which sits on top of EN 16931 in the UBL 2.1 syntax and adds its own PEPPOL-EN16931-R and PEPPOL-COMMON-R rules. Both rule sets are published as Schematron files on the Peppol BIS Billing 3.0 page.

Some countries keep their rules inside the Peppol file itself: of the 166 checks in release 3.0.21, 102 are national rules for Germany, Denmark, Greece, Iceland, Italy, the Netherlands, Norway and Sweden. Germany also has XRechnung, its national CIUS with its own BR-DE rules, for example BR-DE-15 for the buyer reference. There is no Slovak rule in the file. The Slovak national requirements (PASR) mainly set the network identifier, schemeID 0245 followed by the ten digit tax number DIČ, and nothing in the official files checks that the DIČ is really there.

Fatal or warning: which findings stop an invoice

A sender shall not send a message that is not compliant, a receiver may reject one, and a compliant message shall not be rejected.

In the current EN 16931 Schematron for UBL, version 1.3.16, we counted 979 checks: 281 fatal and 698 warning. Of those warnings, 676 are UBL-CR rules, 21 are UBL-DT rules and only one is a BR rule, BR-51. The Peppol file of release 3.0.21 is the other way round: 166 checks, 139 fatal and 27 warning.

A UBL-CR warning usually means the file carries an element the core invoice does not use, for example "[UBL-CR-003]-A UBL invoice should not include the ProfileExecutionID".

We broke one sample invoice eleven ways

The test file was the made-up English sample from our e-invoice check: invoice 2026-0142, Peppol BIS Billing 3.0, reverse charge, 1500.00 without VAT. Unchanged, it passes with no errors and no warnings. On 25 September 2026 we changed one thing at a time and ran our check in Node over every variant.

Only the total with VAT raised by one cent, to 1500.01: BR-CO-15 for the total itself and BR-CO-16, because the amount due no longer equals the new total.
Currency code EURO instead of EUR: 13 errors, BR-CL-04, BR-CO-15 and eleven times PEPPOL-EN16931-R051, one for every amount that still says EUR.
Empty invoice number: BR-02 and PEPPOL-EN16931-R008, "Document MUST not contain empty elements."
A typo in the IBAN: ARL-IBAN, our own mod 97 check. The official Peppol rules check the IBAN only when supplier and customer are both in Germany, and then only as a warning, so a validator with those rules lets this through.
Issue date written as 11.09.2026: ARL-DATUM and PEPPOL-EN16931-F001, "A date MUST be formatted YYYY-MM-DD."
An element left unclosed: XML-01 and nothing else, because no rule can run on broken XML.
A Slovak endpoint with schemeID 0245 holding SK2020000000, a VAT number instead of the DIČ: ARL-SK-DIC-TVAR, again our own check. No official rule checks that a ten digit DIČ follows 0245, so a validator with the official rules lets this through.
An unknown CustomizationID: only one warning in our check, ARL-PROFIL, and only the EN 16931 core rules ran. A validator that runs the Peppol rules stops this file with the fatal rule PEPPOL-EN16931-R004.

Two things to take from this. A single mistake can fill a report: the EURO file produced 13 lines from one wrong code, so fix the first error of each kind and run the check again. And one cent is enough, because BR-CO-15 compares the totals with no tolerance, and that test is the same in version 1.3.16.

Three mistakes that passed without a single finding

Swapping the order of cbc:IssueDate and cbc:DueDate gave zero findings in our check. Element order is the job of the XML Schema, and our browser check does not run it.

Adding a cbc:ProfileExecutionID element also passed with zero findings, because we do not run the UBL-CR group. A validator with the official EN 16931 files would list it as the warning UBL-CR-003.

The third one is the dangerous kind. We shortened CustomizationID to urn:cen.eu:en16931:2017, and our check did exactly what it is built to do: it read the field, detected plain EN 16931 and evaluated 230 rules instead of 289. There was no error because the Peppol rules never ran, while the official fatal rule PEPPOL-EN16931-R004 requires the Peppol value in that field. So read the profile line and the rule count before you believe a report with zero errors.

Which free validator to use, and which release it runs

Free validators you can use today, each with a different scope:

The ecosio validator checks well-formed XML, the XML Schema and the Schematron rules, lists Peppol BIS 3.0, UBL, CII and XRechnung among its formats, and says it deletes a file once the result is displayed. It uses the validation engine of Philip Helger, and the upload page of his Peppol Practical site now points there. We could not confirm which release of the rules its form offers, so look for 3.0.21 before you rely on the result.
The eInvoice Validator of the European Commission checks EN 16931 in UBL and CII with release 1.3.16, but offered no Peppol BIS Billing 3.0 or XRechnung type when we looked, so it stops at the core rules.
The KoSIT Validator, with the KoSIT configuration for XRechnung, is a program for Java 11 or later that you start from the command line; it runs the XML Schema and the Schematron rules and computes an acceptance status.

Whichever you pick, check the release. Peppol publishes its rules in numbered releases: release 3.0.21 came out on 20 May 2026 and has been mandatory since 17 August 2026, together with version 1.3.16 of the EN 16931 artefacts. Compliance is measured against the current artefacts, so a validator on an older release answers a slightly different question.

The changes look small but they bite: BR-CO-25 was removed, R007 now accepts only a list of approved business process identifiers that includes a new profile 02 and two French profiles, and PEPPOL-COMMON-R052 and R053 changed from warnings to errors. One trap caught us too: on 25 September 2026 the docs.peppol.eu files page served release 3.0.21, while the master branch of the GitHub repository still held 3.0.20.

Our own check was built from those GitHub files on 10 September 2026, so it runs release 3.0.20. In practice it still reports BR-CO-25 for an invoice with a positive amount due and neither a due date nor payment terms, and it reports R007 for the new profile 02 identifier and for a French profile, although release 3.0.21 allows them. It also lets urn:fdc:peppol.eu:2017:poacc:billing:02:1.0 through without a finding, a value release 3.0.21 does not approve.

FAQ

Is an EN 16931 validator enough for a Peppol invoice?

No. Peppol BIS Billing 3.0 adds its own PEPPOL-EN16931-R and PEPPOL-COMMON-R rules on top of EN 16931, and the European Commission validator we tried offered only EN 16931 types. Use a validator that also runs the Peppol Schematron.

Can I send an invoice that still shows warnings?

Yes, as long as none of the findings is fatal: a receiver shall not reject a compliant message. Most warnings in the EN 16931 file are UBL-CR rules about elements the core invoice does not use.

Why does my validator still report BR-CO-25?

Most likely because it runs rules older than release 3.0.21. The release notes of 3.0.21, mandatory since 17 August 2026, list BR-CO-25 as removed, and the rule is not in the current file. Our own check reports it for the same reason, as described above.

Which free Peppol validator also checks the XML Schema?

The ecosio validator describes a three stage check: well-formed XML, the XML Schema, then the Schematron rules. The KoSIT Validator runs the XML Schema too, but from the command line. Our browser check does not run it.

Does any official rule check the Slovak 0245 identifier?

Only that 0245 is an allowed scheme. In the Peppol file of release 3.0.21 the code appears only in the list of allowed electronic address schemes, rule PEPPOL-EN16931-CL008, and no check verifies that the ten digit DIČ follows it. Our check reports a VAT number in that place as ARL-SK-DIC-TVAR.

Does a clean result mean the buyer will accept my invoice?

Only for what was actually checked. A clean run with the current artefacts means the file is compliant with the BIS. A partial check like ours, which skips the XML Schema and the UBL-CR, UBL-SR and UBL-DT rules, does not guarantee acceptance by the buyer or the network.

Check your own file

Load your XML into the free ARLing Peppol validator: it reads the file in your browser, shows the detected profile and the number of rules it evaluated, and lists every finding with the rule code, the value from the file and the original wording of the rule. It still runs release 3.0.20, so if it reports BR-CO-25, or R007 for profile 02 or a French profile, the mistake is in our rules, not in your file. It also runs two checks the official files do not make on a Slovak invoice, the IBAN checksum and the DIČ after 0245. Before a large batch goes out, also run the file through a full validator such as the free one from ecosio, which checks the XML Schema and the Schematron rules. The same e-invoice tools also draw the XML as a readable document, so you see the invoice the way a person reads it.

ARLing also has a free e-invoicing tool.

Why an A4 PDF does not fill an e-ink screen, with the arithmetic

ARLing — Wed, 07 Oct 2026 08:30:18 +0000

Originally published at https://arling.sk/notes/why-an-a4-pdf-does-not-fill-an-eink-screen/

Why an A4 PDF does not fill an e-ink screen, with the arithmetic

Open a printable puzzle book or a printable planner on a reMarkable or a Kindle Scribe and it looks wrong in a way that is hard to name: there is a white band down each side, the type is smaller than it should be, and the page feels further away than a page of a real book. Nothing is broken. The file was set for a sheet of paper, the screen is not that shape, and the reader is doing the only sensible thing with a page that does not fit, which is to shrink it until it does. This note does the arithmetic, because once you have seen the numbers you can tell from a product listing whether a file was made for your device or for a printer. It also says what else changes in a page built for glass, and when the paper file is still the one you want.

What the glass actually measures

A reMarkable 2 and a reMarkable Paper Pro both have 1404 by 1872 dots, and at 226 dots per inch that is 157.8 by 210.4 millimetres. A Kindle Scribe has 1860 by 2480 dots, and at 300 dots per inch that is 157.5 by 210.0 millimetres. Both come out at a ratio of 1.333, which is 3:4.

That is the number to hold on to. Our e-ink page is 157 by 210 millimetres, a ratio of 1.338, so it fills the display instead of floating in the middle of it. Boox readers of that size share the same 3:4 shape, which is why one page size serves several devices.

What happens to an A4 page on that screen

A4 is 210 by 297 millimetres, a ratio of 1.414, and US Letter is 215.9 by 279.4 millimetres, a ratio of 1.294. Neither is 3:4, so neither fits.

Take the common case. A reader showing a whole A4 page on a screen of 1404 by 1872 dots fits it by height and scales it to 0.708 of its size. The drawing ends up 148.8 millimetres wide, 9.0 millimetres of width stay empty, that is 4.5 millimetres down each side, and everything on the page is drawn at 71 percent of its size, so text set at 13 points arrives as 9.2 points. The band at the side is the visible half of the problem, and the type size is the half you feel after a while.

What else changes when a page is set for glass

Page size is the start, not the end. An e-ink page has no binding and nothing is trimmed, so the margin is 8 millimetres on all four sides and does not alternate; there is one puzzle to a page even on a small grid; the cell ceiling goes up from 22 to 30 millimetres because you write into it with a stylus; there are four solutions to a page instead of six; there are no page numbers, because the reader does its own paging; and the contrast is stronger, because the display has sixteen shades of grey.

Numbers can be measured after the fact rather than promised. In our planner the smallest type is 8 points, the smallest link is 2.9 by 3.4 millimetres, a day cell on a month page is 18 by 11 millimetres, and no page overflows. In the novels the text is 13 points on an 18 point line, 22 lines to a page, with nothing below 10 points anywhere in the file.

When the paper file is the one you want

There is no argument against A4. There is only an argument against using an A4 file on a screen that is not A4 shaped. So the same content ships as both: A4 is 210 by 297 millimetres, US Letter 215.9 by 279.4 millimetres, which makes US Letter 5.9 millimetres wider and 17.6 millimetres shorter, and both sheets carry the same block, so nobody has to change the scale in a print dialogue.

In the reading books that block is 139 by 194 millimetres, centred on the sheet with a page number under it, so the pages, their order and the links are identical in all three files. Print at 100 percent rather than scaled to fit, and what comes out of the printer is what was measured.

FAQ

Is 157 by 210 millimetres a standard paper size?

No, it is a screen size. It is the 3:4 shape of a 10.3 inch e-ink display, which is why a page of that size fills the display without bars at the sides. For paper we use A4 and US Letter instead.

Will an e-ink file print properly?

It will print, but it is the wrong file for the job: it has an 8 millimetre margin on every side with nothing for a binding, and no page numbers. That is why the print formats exist as separate files with their own margins.

Does a bigger screen change the answer?

The ratio is what matters, not the diagonal. The reMarkable panels, the Kindle Scribe panel and Boox readers of that size all come out at 3:4, so one page shape covers them. A device with a different ratio needs a different layout, and we have not built one.

Where to look next

If you want to see a page built this way, the ten printable puzzle books come with an A4 and a US Letter PDF of the same puzzles, at 4.90 € a book and 19.90 € for all ten. The e-ink files, set on the 157 by 210 millimetre page instead, are the ones described above, and a monthly issue in that format sits at https://arling.sk/puzzle-post/.

ARLing also publishes printable puzzle books.

What the Douay-Rheims Bible is, and why we set that text

ARLing — Tue, 06 Oct 2026 08:29:50 +0000

Originally published at https://arling.sk/notes/what-the-douay-rheims-bible-is-and-why-we-set-it/

What the Douay-Rheims Bible is, and why we set that text

If you have looked for a Catholic Bible text you can print, put in a journal or set on a tablet without asking anyone's permission, you have met the Douay-Rheims, usually without an explanation of what it is. Here is the short version. It is an English Catholic translation made from the Latin Vulgate rather than from Greek and Hebrew, and the version almost everyone means by the name is the Challoner revision of 1749 to 1752, in its American edition of 1899. That edition is old enough to be in the public domain, which is the practical reason it turns up everywhere, and it numbers the Psalms the way the Vulgate does, which is the reason a Catholic reader finds it familiar. This note sets out what it is, what its licence rests on, and what we did and did not do to the text we set.

What it is

A translation from the Latin Vulgate into English, Catholic, and revised by Bishop Richard Challoner between 1749 and 1752; the text we use is the American edition of 1899. Two details tell you what kind of translation you are holding: it renders the divine name as "the Lord", never "Yahweh" or "Jehovah", and the title of a Psalm is counted as its first verse, which is why Psalm verse numbers in it run one ahead of what a modern Bible shows.

The numbering follows the Vulgate as well: Vulgate 9 is Hebrew 9 and 10, most of the psalter differs by one, and the two systems meet again at Psalm 148. A book can print both numbers and remove the confusion, which is what ours does.

What the public domain status rests on

The record of the translation at eBible.org says verbatim "The Holy Bible in English, Douay-Rheims American Edition of 1899, translated from the Latin Vulgate", gives its status as "public domain" and marks it as redistributable. That is one source, and one source is not enough for a claim about rights.

Two independent ones agree. Project Gutenberg carries the same work as ebook 1581, described as "Bishop Richard Challoner A.D. 1749-1752", and the CrossWire project distributes the same revision as its DRC module under a Public Domain licence. Richard Challoner himself died in 1781.

What we did to the text, and what we did not

The book follows the digital edition of eBible.org, identifier engDRA, with the source files dated 3 November 2022; only the markup is removed, and the annotations of the printed edition are not included. Removing markup is not a change to the text.

There is one exception, and it is worth describing exactly, because it is the kind of thing a publisher is usually silent about. Digital transcriptions contain typing errors. The first book corrects twelve obvious ones, for example Manesses for Manasses in Matthew 1:10, and only where two other transcriptions of the same printed edition, one on drbo.org and Project Gutenberg ebook 1581, agree on a different reading; the list of corrections is printed on the last page of the book. Comparing all 828 verses of the first book with the Gutenberg transcription after those corrections gives 652 verses identical, 99 differing only in punctuation and 76 differing in wording with no obvious error, and those 76 are left exactly as they are.

Why we chose it, and what we do not claim

Three reasons, in order of weight. The text is in the public domain, so it can be sold as part of a book. It is Catholic and numbers the Psalms after the Vulgate, so it matches the tradition the book is written for. And it exists in a machine readable edition, so a program can set it without a human retyping anything.

Now the limits. No ecclesiastical approval is claimed for our book. The approbation of the 1899 edition is mentioned only in the description of the CrossWire module, eBible.org does not mention it, and it is a historical detail about the source rather than an approval of anything we made. The book itself is a private aid to prayer, not a liturgical book, and it does not replace the Mass, the Liturgy of the Hours or spiritual direction.

FAQ

Is the Douay-Rheims the same as the King James Version?

No. The Douay-Rheims is a Catholic translation made from the Latin Vulgate, in the Challoner revision of 1749 to 1752 and its American edition of 1899, and it numbers the Psalms after the Vulgate.

Can I print or republish the text myself?

The record at eBible.org gives the status "public domain" and marks it redistributable, and both Project Gutenberg and the CrossWire DRC module carry the same work as public domain. What you do with that is your decision to check, not ours to promise.

Did you modernise the spelling?

No. Only the markup is removed and the annotations of the printed edition are left out. Twelve obvious transcription errors are corrected, with the list printed in the book, and 76 further differences in wording were left untouched.

Where to look next

If you want to read a few pages of the text as we set it, days 1 to 7 are a free sample on the Morning Quiet page at https://arling.sk/morning-quiet/, with no account and no e-mail address. The first book, 31 days of Psalms and Matthew with the text printed on the page, is a digital download of three PDF files at 4.90 €. It has not been opened on a physical reMarkable, Kindle Scribe, Boox or Supernote, so the sample is the place to start, and our other titles, including the ten printable puzzle books, are set on the same page.

ARLing also publishes printable puzzle books.

Reading a novel of 953 pages on an e-ink tablet: what actually matters

ARLing — Mon, 05 Oct 2026 08:29:51 +0000

Originally published at https://arling.sk/notes/reading-a-953-page-novel-on-an-eink-tablet/

Reading a novel of 953 pages on an e-ink tablet: what actually matters

A long novel is where the small decisions in a file stop being cosmetic. At 953 pages you will open the book a few hundred times, and anything that costs you two seconds each time costs you an hour of the reading. The things that matter are not the ones listings talk about. They are the length of a line, the size of the type, whether the contents can be reached without paging back to the front, and whether the page is the shape of your screen at all. Cheap PDFs of public domain novels usually get the last one wrong, which is why the type arrives small and shy of the edges. This note goes through each of those in turn using a book we set, says what we measured, and ends with the thing most sellers of public domain texts leave out.

The line, the type and the page

Our classics are set on a 157 by 210 millimetre page, the 3:4 shape of a 10.3 inch e-ink display, so the page fills the screen rather than sitting in the middle of it with bands at the sides.

On that page the text is ARLing Serif at 13 points on an 18 point line, 22 lines of text to a page, with no hyphenation, which works out at an average of 64 characters a line. Line length is the quiet one of those numbers. A line of about that width is what lets the eye find the start of the next line without hunting, and a file that reaches it by shrinking an A4 layout onto your screen will not.

Nothing in the file is below 10 points, the fonts are embedded as complete outlines rather than as bitmap substitutes, and a diagram is always kept whole on one page.

Getting back to the contents

A book you cannot navigate is a scroll. Every page of ours carries a Contents link in an 8 millimetre footer beside the page number, with a row 7.5 millimetres tall so a finger can hit it, and the contents itself links to every chapter and part, each entry 8.8 millimetres tall with long titles wrapped onto two lines.

Measured in the finished file, Ben-Hur has 1043 internal links and 91 link targets, with 0 dead links and 0 pages overflowing, across 8 books and 81 chapters. The other title is longer and is issued as five volumes following the five parts of the source: 479, 523, 504, 414 and 367 pages, 2287 in all, 117 chapters, 2415 links and 133 targets, 0 dead.

The band for the pen

A tablet with a pen invites writing in a book, and a novel with no room for it wastes the pen. At the foot of every page there is a 33 millimetre band for notes with a dotted rule, under the text and above the footer.

Since the band is only useful if you can also type into it on a computer, there is one form field over it on every text page: 935 in Ben-Hur and 2257 across the five volumes, one per text page.

What we could not test

No file of ours has been opened on a physical reMarkable, Kindle Scribe, Boox or Supernote, so everything above is measured in the PDF and not on glass. Greek words in one of the titles are set in a fallback system font, which is a limitation of the typesetting rather than of the text.

The samples exist for exactly this reason: the first three chapters of Ben-Hur are 26 pages with 31 links and 21 fields, and the first three chapters of the other title are 52 pages with 57 links and 47 fields.

The part sellers leave out

The same texts are offered free of charge as EPUB by Project Gutenberg and Standard Ebooks. If you read on a phone or a reader that is happy with EPUB, that is very likely the right answer, and we would rather say so here than have you discover it afterwards.

What is paid for in a file like ours is the typesetting for an e-ink page, the band for notes, the page numbers and the linked contents in a PDF, and our own listing says so in its own words. The text itself is public domain and unchanged.

FAQ

Why is the novel split into five volumes?

Because the source itself is in five parts, and the volumes follow them: 479, 523, 504, 414 and 367 pages, 2287 in all. Each volume has its own title page, edition page and contents.

Is the text abridged or modernised?

No. It is a public domain text from a plain text transcription, set without changes; the page sizes and the links are ours, the words are not.

Will it print?

Each title also comes as A4 and US Letter files that carry the same 139 by 194 millimetre block centred on the sheet, so the pages, numbers and links are identical in all three files.

Where to look next

The two titles are The Count of Monte Cristo and Ben-Hur: A Tale of the Christ, both public domain, both set on the same page, at 4.90 € a title in our Etsy shop, https://www.etsy.com/shop/ARLingPuzzles. Free samples of the first three chapters are part of each listing. If you would rather have something to solve than to read, the ten printable puzzle books hold 100 puzzles each with solutions at 4.90 € a book.

ARLing also publishes printable puzzle books.

Putting a printable puzzle in a parish bulletin or a school newsletter

ARLing — Sun, 04 Oct 2026 08:29:52 +0000

Originally published at https://arling.sk/notes/putting-a-printable-puzzle-in-a-parish-bulletin-or-school-newsletter/

Putting a printable puzzle in a parish bulletin or a school newsletter

Filling the back page is a monthly problem with a small set of honest solutions, and a puzzle is one of them: it is read by people who ignore everything else on the sheet, it takes no editing, and it gives you a reason to print next month's issue. The practical difficulties are dull and specific. A picture pulled off the internet lands in your column at the wrong width and goes soft when it is scaled. Grey fills turn into mud on a parish photocopier. And the question nobody wants to ask out loud is whether you are allowed to print it at all, which for most puzzles found by searching is either no or unknown. This note works through the width, the printing and the licence, in that order.

The width decides everything

Newsletter layouts are built on columns, so the first question is not what the puzzle looks like but how wide it is. A puzzle block sized for one column drops in; a square image sized for a screen has to be scaled, and scaling a grid of thin lines is what makes it go soft.

Our weekly sheet is built around a block that measures 80 millimetres, which fits a single newsletter column, and it is delivered both as a page and as the block on its own, as an image 945 dots wide at 300 dots per inch. The image is the part that matters for a layout program: you place it at its measured width and never touch the scale.

Line work, not grey

A puzzle for print is line work in black and white. The instruction printed on our own sheet says it in five words: black and white, no grey. Grey fills are the first thing a photocopier ruins, and a grid whose thin rules have been rasterised at screen resolution is the second.

That is what 300 dots per inch is for. At that resolution the lines of the grid survive the printer instead of turning into a smudge, which is why the block is delivered at that resolution and at a stated width rather than as a screenshot.

Print at 100 percent

The most common way to ruin a good file is the print dialogue. Our sheets say the rule on the page: print at 100 percent, not scaled to fit.

Scale to fit is the setting that silently shrinks a page by a few percent to clear a printer's unprintable margin, and it takes the measured widths with it. The way around it is to ship the right paper instead of asking anyone to rescale, so each week comes as an A4 and a US Letter file, since A4 is 210 by 297 millimetres and US Letter is 215.9 by 279.4, which makes US Letter 5.9 millimetres wider and 17.6 shorter. The solution goes on the second page, so you can print it in the next issue or keep it.

What a licence has to say

Four things, and if any of them is missing, the answer is no. Whether you may print it in your publication. How many copies. Whether you may put it on your website. And whether you may resell it or claim it as your own.

Our weekly sheet answers them in one line printed on the sheet itself: print this puzzle in one publication, in as many copies as you print of it, not for resale as a puzzle product, and web use is part of the Pro plan. The one off pack for publishers answers the same questions in a written licence: a non-exclusive commercial licence that permits your own books, magazines, newspapers, newsletters and educational materials with no limit on printed copies, and prohibits redistributing or reselling the source files or the pack, uploading it to a marketplace and claiming authorship of the puzzles.

Two limits you should know before you rely on it. The same non-exclusive pack can be bought by other publishers, so it is not an exclusive right. And the duplicate check behind it covers the archive and books we compared, not worldwide novelty. For a single puzzle in a club newsletter there is also the free route: personal and non-commercial use is free with the credit line, and commercial use is what the subscription covers.

What we have not done

No sheet of ours has been printed on a real printer or set in anyone else's newsletter, nobody holds a subscription to the weekly sheet, and revenue is zero. So the widths above are measured in the files and in the build, not in a finished parish bulletin.

What does exist is the material: 52 weeks are built, from Monday 21 September 2026 to Monday 13 September 2027, 156 puzzles in 624 files. Every one of them carries the line that the puzzle has exactly one solution, found by the program that made it.

FAQ

Can I print the puzzle in as many copies as I print of my newsletter?

Yes, that is exactly what the licence line on the sheet says: print it in one publication, in as many copies as you print of it, with no resale as a puzzle product. The publisher pack says the same for your own books, magazines and newsletters, with no limit on printed copies.

What about putting it on the parish website?

Web use is not part of the basic plan; it is part of the Pro plan. For non-commercial personal use the free route needs the credit line on the picture.

What size image do I need for a single column?

A block 80 millimetres wide, supplied as an image 945 dots wide at 300 dots per inch, placed at that width without scaling.

Where to look next

If you want files to inspect before anything else, three evaluation samples are free, and the full pack of 25 puzzles on 7 by 7 grids comes with matching SVG, PDF, PNG and JSON files and the written commercial licence at 19 € including VAT. If what you want is one puzzle a week rather than a pack, the weekly sheet is at https://arling.sk/puzzle-post/bulletin/, 19 € a month for one puzzle or 39 € for three levels.

You can check your own SEPA pain.001 file for free with the SEPA pain.001 Doctor.

Public domain in practice: what it means for Ben-Hur and Monte Cristo

ARLing — Sat, 03 Oct 2026 08:29:54 +0000

Originally published at https://arling.sk/notes/public-domain-in-practice-ben-hur-and-monte-cristo/

Public domain in practice: what it means for Ben-Hur and Monte Cristo

Public domain is one phrase covering two different questions, and mixing them is how people get themselves into trouble. The first question is whether the work itself is still under copyright. That one has an answer you can reach with a date of death and a calendar. The second is whether the particular copy you are holding may be used, which has nothing to do with the age of the novel and everything to do with who typed it out, how, and under what terms. A transcription of an old book can carry conditions even when the book carries none. This note works both questions through for two specific titles we set for e-ink tablets, with the dates, the rule that applies in the European Union, and the licence clause that decides the second question.

Question one: is the work free

Ben-Hur: A Tale of the Christ is by Lew Wallace, who lived from 1827 to 1905, and the original English text was published by Harper and Brothers in New York in 1880. The Count of Monte Cristo is by Alexandre Dumas, who lived from 1802 to 1870, with Auguste Maquet, 1813 to 1888; the anonymous English translation was first published by Chapman and Hall in London in 1846.

In the European Union and in Slovakia, economic rights last 70 years after the author's death, and for an anonymous work 70 years from publication, under Act No. 185/2015 Coll., section 32. Dumas died in 1870, Maquet in 1888, Wallace in 1905, so every one of those terms ran out long ago. Project Gutenberg states the American position on both works in the same words: public domain in the USA.

Question two: is this copy free

This is the part that is easy to miss. The text you download is not the 1880 book. It is a transcription somebody made, and Project Gutenberg attaches terms to its own transcriptions.

Read them and they are clear enough. Under section 1.C of its licence, anyone may copy, distribute and build on a work that is unrestricted by copyright in the United States, as long as all references to Project Gutenberg are removed. Under section 1.E.8, a work that keeps those references may be sold only with a royalty of 20 percent of gross profits. In other words: strip the branding and the underlying public domain text is yours to use, or keep the branding and pay for the privilege.

What that means for a file you sell or make

We took the first route, and the consequence is visible in the files. The PDFs on sale contain no mention of Project Gutenberg at all: the edition page calls the source a public domain digital transcription, and the check that runs over the finished PDF verifies that the forbidden words are not in it.

That is not the same as hiding where the text came from. The source, with the ebook number, the address and the checksum of the downloaded file, is recorded in our own documentation, in the file manifest and in the listing itself, outside the work. The two records, ebook 2145 for Ben-Hur, released in 2000 and last updated in 2023, and ebook 1184 for The Count of Monte Cristo, released in 1998 and last updated in 2026, are both marked public domain in the USA.

What may be changed, and what we did not change

Public domain means you may change a text. It does not mean you should, and the honest thing is to say exactly what was touched. In ours the sequence of words in the PDF equals the sequence of words in the body of the source file, and that is checked twice, once at download and once in the finished PDF.

What is left out is the bookkeeping around the text: the transcription's own header and footer, the opening block with the title, author and contents list, and the lines carrying illustration file names. Line breaks are reflowed, words marked with underscores in the transcription are set in italics, and common indentation of multi-line blocks is removed. The spelling and punctuation of the transcription stay as they are, and nothing is added, cut or modernised.

The part that is easy to leave out

The same texts are offered free of charge as EPUB by Project Gutenberg and Standard Ebooks. If that suits your reader, take it.

What is paid for in a file like ours is the typesetting for an e-ink page, the band for notes with a pen, the page numbers and the linked contents in a PDF, and our listing says so in its own block. Also on the honest side of the ledger: neither file has been opened on a physical reMarkable, Kindle Scribe, Boox or Supernote.

FAQ

Can I sell a public domain novel myself?

The work being out of copyright is only the first question; the second is the terms on the copy you started from. Project Gutenberg's licence allows use of the underlying public domain work if all references to Project Gutenberg are removed, and otherwise asks a royalty of 20 percent of gross profits.

Why does your PDF not credit Project Gutenberg?

Because that is the condition of using the text without a royalty, and a check over the finished file verifies that the words are absent. The source is recorded in the manifest and in the listing instead, outside the work.

Is the text the original or a modern edition?

It is a transcription of a nineteenth century edition, unchanged in wording, with the spelling and punctuation of that transcription kept.

Where to look next

Both titles are set on a 157 by 210 millimetre page with a band for notes and a linked contents, at 4.90 € a title, and each listing carries a free sample of the first three chapters. If you want a puzzle rather than a novel, the ten printable puzzle books hold 100 puzzles each with their solutions at 4.90 € a book.

ARLing also publishes printable puzzle books.

Psalm numbers, and how a 31 day reading plan decides where to stop

ARLing — Fri, 02 Oct 2026 08:29:53 +0000

Originally published at https://arling.sk/notes/psalm-numbers-and-how-a-31-day-reading-plan-is-cut/

Psalm numbers, and how a 31 day reading plan decides where to stop

Two things quietly defeat a home made Bible reading plan. The first is that Psalms have two numbers. A Catholic book will print Psalm 22 where a Protestant book prints Psalm 23, both are correct, and a plan that mixes the two systems will send you to the wrong page about a hundred times in a year. The second is cutting: a plan has to end each day somewhere, and if it ends in the middle of a sentence or splits a Psalm in a place the Psalm does not want to be split, the reading feels like homework. This note explains the numbering in full, because it is a table and not an opinion, then shows the two rules a plan needs to cut well, with the counts they produce over a whole psalter and four Gospels.

Why Psalm 23 is also Psalm 22

The Douay-Rheims and the Vulgate number the Psalms differently from the Hebrew. Psalms 1 to 8 and 148 to 150 are the same in both. Vulgate 9 is Hebrew 9 and 10. Vulgate 10 to 112 is Hebrew 11 to 113. Vulgate 113 is Hebrew 114 and 115. Vulgate 114 and 115 together are Hebrew 116. Vulgate 116 to 145 is Hebrew 117 to 146. Vulgate 146 and 147 together are Hebrew 147.

So for most of the psalter the two systems differ by one, and they differ because two Psalms are joined in one system and split in the other, twice each way. A book can solve this by printing both numbers, which is what ours does: the Hebrew number first and the Vulgate number in brackets, for example Psalm 23 (22), Psalm 9-10 (9) where they merge and Psalm 116 (114) where they split. Verse numbers always follow the Douay-Rheims, and in that edition Psalm 118, the Hebrew 119, has 176 verses in 22 stanzas.

Where a day should end

A plan needs two rules, one for the Psalms and one for the Gospel, and they are not the same rule because the two kinds of text do not break the same way.

For the Psalms: a Psalm of up to 40 verses is one section, and a longer one is split into the smallest number of sections of up to 40 verses, with the cuts on stanza boundaries where there are any and on verse boundaries otherwise. Over the whole psalter that comes to 160 sections.

For the Gospels: a chapter is split into sections of 8 to 20 verses aiming at about 13, chosen so that as many cuts as possible land on a paragraph boundary. That gives 297 sections, of which 285 end on a paragraph or a chapter boundary. The twelve that do not are the honest cost of never letting a day run past twenty verses.

What 31 days actually come to

A day holds one Psalm section and one Gospel section. In the first book that works out at an average of 15.8 Psalm verses and 12.7 Gospel verses a day, and one day takes 5 to 10 pages once the text, the four steps and the evening page are laid out.

In content, the first book covers Psalms 1 to 31 in the Hebrew numbering and Matthew 1 to 12, with Psalm 18 spread over two days and Psalms 9 and 10 read on one day because the Vulgate counts them as a single Psalm. The text is printed on the page rather than referenced, from the Douay-Rheims Bible, Challoner revision, American edition of 1899, which is in the public domain.

Dates, weekdays, and starting again

A plan tied to a calendar year is a plan you can fall behind. Ours is not tied to one: the days are numbered 1 to 31, with no year and no weekday, and each morning page has a line for your own date.

Days 7, 14, 21 and 28 are written for Sunday, the Lord's Day, each with a page for the week, and a fifth week page closes days 29 to 31. Start on a Monday and the Sunday pages fall on real Sundays; start on any other day and nothing breaks, and you can start again when you reach the end.

FAQ

Which number should I use when someone gives me a reference?

Ask which Bible they are reading from. For most of the psalter the Vulgate number is one lower than the Hebrew number, and a book that prints both, such as Psalm 23 (22), removes the question entirely.

Why is Psalm 9 longer in a Catholic Bible?

Because Vulgate 9 is Hebrew 9 and 10 joined, which is also why our first book reads it over a single day.

How long is a day's reading?

In the first book, an average of 15.8 verses of a Psalm and 12.7 verses of a Gospel, laid out over 5 to 10 pages. No day runs past twenty verses of Gospel, because that is the upper limit the cutting rule allows.

Where to look next

If you want to see a plan cut this way, days 1 to 7 are a free sample on the Morning Quiet page at https://arling.sk/morning-quiet/, with no account and no e-mail address. The first book is a digital download of three PDF files at 4.90 €, and it is a private aid to prayer rather than a liturgical book. Our other titles set on the same page size, including the ten printable puzzle books, sit next to it.

ARLing also publishes printable puzzle books.

Testing PDF links on your e-ink tablet

ARLing — Thu, 01 Oct 2026 08:29:54 +0000

Originally published at https://arling.sk/notes/pdf-links-that-still-work-on-kindle-scribe-remarkable-boox/

Testing PDF links on your e-ink tablet

A hyperlinked PDF is useful when you can open a day or a puzzle from its contents and find your way back. There are two parts to checking that journey. One is inside the file: does the link lead to an existing page? The other is in the reader: how do you activate it, and is that comfortable on the device you own? We can report measurements from the file, but we have no recorded physical test of these books on a reMarkable, Kindle Scribe, Boox or Supernote. A useful first step is therefore a small sample. Try its contents, destination and return link on a computer, then repeat the same route after transferring it to your tablet. This note explains what that comparison can tell you, what we have measured in our own files, and what remains yours to try.

How the file gets to the reader

Download a sample and use the normal PDF transfer method documented for your device. Then open the transferred file and repeat the navigation you tried on the computer. This is a proposed comparison of your own workflow, not a claim that every transfer service preserves every PDF feature.

Start with the original PDF: the links we measure are PDF annotations inside that file. If your transfer route changes the file or your reader handles it differently, test the result rather than assuming that a successful click on the computer settles the question. Likewise, a page size alone does not establish how comfortably navigation works on your display.

Why a link inside a PDF is a real thing

Our books are written as ordinary web pages with ordinary links, and Chrome keeps them when it prints to PDF, as link annotations with a destination inside the same file. An annotation is a rectangle on a page with a destination attached, so it travels with the file to any reader that acts on it.

A book therefore has a shape you can navigate. In our puzzle books the contents page opens a puzzle, puzzle pages link to their solutions and the contents, and solutions link back to their puzzles. In the e-ink edition of Morning Quiet, direct inspection counted 592 internal links. Its shorter sample contains 139.

When a link does not respond

A failed tap is a reason to investigate, not a diagnosis on its own. Check how your reader activates PDF links and which mode you are using. Try the same route in the original sample on a computer, and note whether the problem is opening the contents entry, reaching the solution or returning to the task. We do not have measurements that establish one common cause across these devices.

Our reading book puts navigation in a centred bar near the bottom. Its measured clearance is at least 29 millimetres from the left and right page edges and 15 from the bottom. Those measurements describe the PDF page, not the edge of your device's glass. They do not establish a universal safe gesture area, or whether your particular reader expects a finger, a pen or another action.

What the numbers actually prove

Ask which file was checked and what the check covered. Our direct inspection of the Morning Quiet e-ink book found 214 pages and 592 internal links; the sample has 52 pages and 139 links. Neither file contained an external link or an internal link whose destination page fell outside the document. That checks the existence of target pages, not whether every destination is the one a reader expects.

The e-ink edition of Puzzle Post Issue 1 has 145 pages and 598 internal links, again with no destination page outside the document. When we add text fields to the reading PDFs, we also check that each page retains its link count and renders identically at 72 dots per inch. These are useful checks on the file and the processing step. Neither substitutes for trying the controls on a tablet.

What we have not tested

We have no recorded physical test of these files on a reMarkable, Kindle Scribe, Boox or Supernote. The page measurements and link counts above are not measurements of those devices. We also have no recorded physical-device test of the Puzzle Post files.

The free Morning Quiet sample is a useful place to start: days 1 to 7, 52 pages and 139 internal links. Transfer it first and try the route you intend to use. If you also want to write, test that separately by saving an answer and reopening your copy. A navigation test does not establish whether form fields or annotations will save correctly.

FAQ

Does converting a PDF for my reader keep the links?

We have not verified every conversion route. Our measurements apply to the original PDF link annotations. Start with that file, then test the result of your own transfer or conversion before relying on it.

Why are the buttons at the bottom and not in the corner?

The reading book uses a centred bottom bar with measured clearance from the PDF page edges. That is a layout decision. It does not prove that this position avoids every device's gesture controls; the sample lets you try the arrangement yourself.

How do I know a file's links were checked at all?

Look for the file and the scope of the check. For example, direct inspection found 592 internal links in the Morning Quiet book and 139 in its sample, with no target page outside either file. Then use the sample to check the actual contents, destination and return route in your reader.

Where to look next

Try the free Morning Quiet sample on your own device, with no account and no email address required. For the puzzle side, look at Puzzle Post. If you prefer paper, the ten printable puzzle books hold 100 puzzles each with solutions, at 4.90 € a book.

ARLing also publishes printable puzzle books.

PDF form fields: typing, pen notes and what to test

ARLing — Wed, 30 Sep 2026 08:29:55 +0000

Originally published at https://arling.sk/notes/pdf-form-fields-in-a-journal-or-puzzle-pdf/

PDF form fields: typing, pen notes and what to test

A journal PDF and a puzzle PDF can pose the same quiet question: where do you put your answer? A ruled line is part of the page drawing. A form field is a separate box that a supporting reader lets you fill with a keyboard. Pen notes and text annotations are another way to add an answer, using the reader's own tools. They do not require the author to have put a form field under that particular line. A click that does nothing therefore tells you less than it first appears: the file, the reader and the selected tool all matter. This note explains how our fields are added to a finished file without changing its empty appearance, which puzzle types receive cell fields, and how to try typing, saving and reopening an answer before you commit to a book.

What a form field actually is

A form field is a box you can type into with a keyboard in a supporting reader. Our reading PDFs get theirs after the PDF is finished: a script walks the printed file and adds a one line text field over suitable grey writing lines. An empty field with no border and no fill draws nothing, so our comparison can check that the file still looks as it did before.

Finding the lines is the fussy part. With the settings used for Morning Quiet, a candidate is horizontal or a thin filled rectangle, its grey is near the value 0.541 in each channel, its stroke is at most 0.8 of a point thick and it is at least 20 millimetres long. A field must also avoid overlap with a link annotation on the same page. These are choices in our script, including a configurable minimum length, not rules that every PDF must follow. Avoiding links keeps a field from landing on top of a navigation row in the contents.

Why typing and pen notes are different

The distinction is between filling a prepared field and adding an annotation. Firefox documents both interactive fields and separate text and drawing tools; Adobe Reader describes Fill & Sign for non-interactive forms too. It would be wrong to say that a computer cannot write into a PDF without fields. Fields provide prepared places for an answer, while annotation tools offer another method whose behaviour you need to check in your reader.

The e-ink edition of Morning Quiet has 1038 fields on 118 of its 214 pages. Its free sample has 229 fields on 26 of its 52 pages. The number of pages containing fields is not the length of the book. Those counts come from inspecting the finished files; they do not establish how any particular tablet handles the fields.

Which grids can use fields in their cells

In a puzzle book, a text field would not suit every kind of answer, so not every type gets cell fields. We use them for Killer Sudoku, Anti-knight Sudoku, Kakuro, Nonograms, Star Battle, Nurikabe and Numberlink, where a digit, a mark for a filled cell, a star or a path number can record a cell's state.

Masyu gets no cell fields, because its answer is a loop between cell centres, and Slitherlink gets none either, because its answer sits on the edges between the dots. Hashi also needs drawn connections rather than text boxes. These types have note fields, while solution pages get no fields at all. In the e-ink edition of Puzzle Post Issue 1 there are 4457 fields: 4157 in grid cells and 300 over the ruled notes lines, with 598 links. These are counts for that edition; A4 has more note fields.

How to test an answer in your own reader

Open a known sample in a reader that supports forms and try a place intended for an answer. Type a short entry, save your own copy, close it and open it again. Check whether your entry survived. If clicking does nothing, check the active tool and the reader's support; that click alone does not prove that the file has no fields. Test annotation tools separately if those are how you intend to write.

There is a second question worth asking a seller: whether adding the fields changed the file. Our script compares the number of link annotations on every page and the pixels of every page rendered at 72 dots per inch before and after adding fields. A difference fails this check. In Morning Quiet the report records 592 links before and after, with identical pixels. This checks the processing step, not the behaviour of every reading application.

What we have not tested

Fields are measured in the file, not on a device. We have no recorded physical test of these files on reMarkable, Kindle Scribe, Boox or Supernote. Reader support and the visibility of the grey writing lines on an actual display remain things to try on your own device.

If you plan to type, try your own reader first with a free sample. The Morning Quiet sample contains 229 fields for that purpose. A successful test describes your file and reader, not a promise about every other combination.

FAQ

Do the fields change how the page looks?

Our empty fields have no border and no fill. The processing report confirms identical rendered pages before and after adding them. That is a measured result for the file, not a statement about how every reader highlights an active field.

Can I write with a pen and type on the same page?

Fields and annotations are separate features. Whether your reader lets you use both, and preserves both when saving, needs a test in that reader. Try one short entry of each kind in the sample and reopen your saved copy.

Why does my Masyu page have no cell boxes?

Because its answer is a loop along the edges between cell centres, not a mark inside a cell, so a box in the cell would be misleading; Slitherlink is left out for the same reason.

Where to look next

Try the free Morning Quiet sample first: days 1 to 7, with 229 fields and no account or email required. The full book has 1038 fields. The monthly issue with fields in the grids is Puzzle Post, whose e-ink edition has 145 pages and 598 internal links. The ten printable puzzle books are the print side of the same work, 100 puzzles each with solutions at 4.90 € a book.

ARLing also publishes printable puzzle books.

Lectio divina: the four steps, and what a journal is actually for

ARLing — Tue, 29 Sep 2026 08:29:56 +0000

Originally published at https://arling.sk/notes/lectio-divina-four-steps-and-what-a-journal-is-for/

Lectio divina: the four steps, and what a journal is actually for

Lectio divina is old, simple and easy to describe badly. It is a way of reading a short passage of Scripture slowly, four times over, with a different question each time, and it has an order: lectio, meditatio, oratio, contemplatio. People who sell journals for it tend to skip straight to the ruled lines, which is the least interesting part, and the lines then invite you to treat a prayer as a worksheet. This note goes the other way. It says what each of the four steps asks, where the sequence comes from, and then why a page built for it gives two steps room to write and deliberately leaves the other two blank. At the end it says plainly what such a book is not, because that matters more than any feature: it is a private aid to prayer and nothing more.

The four steps, in order

The order is lectio, meditatio, oratio, contemplatio. Read, think, pray, rest. Lectio is the reading itself, slowly enough that the words stop sliding past. Meditatio is turning over the line that caught you. Oratio is saying something back. Contemplatio is the part with nothing to do in it.

That last step is the one a journal can ruin. In our reading book the pages are laid out to respect the difference: Lectio has no lines, Meditatio has three, Oratio fills the rest of the page with lines 8 millimetres apart, and Contemplatio has no lines at all. The blank is not an oversight, it is the step.

Where the sequence comes from

The four step progression was formalised by Guigo II, a Carthusian monk and prior of Grande Chartreuse who died late in the twelfth century, in his book The Ladder of Monks, which is considered the first description of methodical prayer in the western mystical tradition. The practice is older than the description of it. The Rule of Saint Benedict, in chapter forty-eight, sets times and a manner for lectio divina, and that is where the daily rhythm comes from rather than from any book of pages.

What a page can honestly do

A page can do three small things. It can put the text in front of you so that the reading is not preceded by ten minutes of looking for the passage: each day in our book carries the text of one Psalm and one Gospel passage printed on the page, not a reference to look up. It can hold the four steps in order so that you do not have to remember them. And it can leave you alone in the last step.

It can also stay out of the calendar. The days are numbered 1 to 31 with no year and no weekday, there is a line for your own date on each morning page, and days 7, 14, 21 and 28 are written for Sunday, so a start on a Monday puts the Sunday pages on real Sundays. Begin on any day, stop, begin again: the book has no opinion about that.

What such a book is not

It is a private aid to prayer, not a liturgical book. It does not replace the Mass, the Liturgy of the Hours or spiritual direction, and no ecclesiastical approval is claimed for it. Any journal that suggests otherwise is selling you something other than a journal.

Two more pieces of honesty about ours. The headings, instructions and questions in it are our own editorial text prepared with the help of a language model and approved by the publisher, while the Scripture is a public domain translation and the typesetting is done by our own program. And the book has never been opened on a physical reMarkable, Kindle Scribe, Boox or Supernote, so if you plan to use it on a tablet, start with the free sample.

FAQ

Do I have to do all four steps every day?

The method has four steps in that order, and that is what the page holds. What you do with it on a given morning is between you and God, and a book cannot referee that; it is a private aid to prayer, not a rule.

Why does the page give Oratio so much more room than Meditatio?

Because they ask for different amounts of writing: Meditatio has three lines, Oratio fills the rest of the page with lines 8 millimetres apart, and Lectio and Contemplatio have none. The layout follows the steps rather than filling the sheet evenly.

Which passages does a day use?

In our book each day has one Psalm and one Gospel passage printed in full; the first book covers Psalms 1 to 31 in the Hebrew numbering and Matthew 1 to 12. Psalms carry both numbers, for example Psalm 23 (22), because the Douay-Rheims and the Vulgate count them differently.

Where to look next

If you want to see a page laid out this way before spending anything, days 1 to 7 are a free sample on the Morning Quiet page, with no account and no e-mail address. The book itself is a digital download of three PDF files at 4.90 €. Our other titles set on the same page size, including the ten printable puzzle books, sit alongside it.

ARLing also publishes printable puzzle books.