Click starts isbn-maschine
Click starts isbn-maschine

The isbn-maschine was launched for the first time on 08.03.2018.

On that day, the isbn-maschine displayed 359 books graphically. The isbn-maschine s constantly updated. The number of books was on

  • 01.01.2019: 384 books
  • 01.01.2020: 416 books
  • 01.01.2021: 443 books
  • 01.01.2022: 482 books
  • 01.01.2023: 519 books
  • 01.01.2024: 544 books

The work of art

» start isbn-maschine

Example pages:


last english pages added:

Anna Burns: Milkman (added May 20th, 2024)





Text Colour

The text of the page defines the text colour. For this, the text determines the value of the hexadecimal colour definition. The hexadecimal colour definition is based on the RGB colour space. The text determines the colour value for red, green and blue.

The first letter of the text determines the colour value for red and determines which letters are used to determine the colour values green and blue. The selection is made in German texts according to the order of the letters on a QWERTZ keyboard, in English texts of the QWERTY keyboard. The order starts with Q for the German keyboard and ends with ß. In English, it starts with Q and ends with M. The selection is made according to the writing direction of the Latin writing system.

The determination of the colour value depends on the number of occurrences of the selected letter in the evaluated sentence with the following formula in hexadecimal notation:

255 / [Number of occurrences Letter]

If the selected letter does not exist in the text, the value is "00". If the occurrence of all three selected letters is “2”, the formula is modified as follows:

(255 / 2) - 5

Background colour

The background colour is the complementary colour of the text colour, the individual colour values are determined using the following formula:

255 – (255 / [Number of occurrences Letter])

If the selected letter in the text occurs once, the value is "00". If it does not occur is the value is "FF"

Font size

The length of the sentence determines the font-size. Sentences with less than three words have a font size of 5em, sentences longer than that but shorter than 500 characters have a font size of 2.7em, longer texts have 1.5 em.

Display duration

The page is displayed one seventh of a second per word plus 6 seconds.


Below the sentence, the ISBN number of the book and the type of sentence are given. This links to the entry of the book in the online directory of the national library of the publishing country of the book.

Home page

The start page is determined by the day's date using the following logic. Each day of the month is assigned a letter in the order of the QWERTZ keyboard (1 = β, 2 = q, 3 = w, 4 = e, 5 = r, 6 = t, 7 = z, etc.). On each day of each month, a page is determined where this letter has the highest percentage of the characters of the sentence compared to all pages of the isbn-maschine. If several sentences have the same value, the page is selected by a random factor.

Following pages

The following page is automatically displayed according to the writing direction of the Latin writing system.


Directional arrows

Using the directional arrows, the user can influence the writing direction but not the display duration:

  • Arrow to the right: Writing direction according to Latin writing system first horizontally from left to right, then from top to bottom
  • Arrow to the left: the reverse of the writing direction according to Latin writing system, first horizontally from right to left, then from bottom to top
  • Arrow down: Writing direction according to Mongolian writing system, first vertically from top to bottom, then from left to right
  • Arrow up: Writing direction vertically from bottom to top, then from right to left

Pause / Start – button

With the pause button, the work of the machine can be stopped. The start button is used to restart the machine. It restarts the time evaluation of the page.

Page Name

The page name links to the entry of the book in the online directory of the national library of the publishing country of the book. The entry is output in a new browser window



The isbn-maschine includes the first, middle and last sentence of books the machinist has read (he still has some work to do).


A sentence in the sense of the isbn-machine is a word sequence with at least one word / character at the end of which stands a punctuation mark, on which the next sentence normally begins with capitalization or anything that appears to the machinist as closest to this idea.

First sentence

The first sentence of a book is defined as the first original sentence of a book. Dedications, headlines, prefixed quotes, in plays the list of participants are ignored by the machinist

Middle sentence

The machinist determines the middle sentence using the page numbers of the book. For this, he uses the following formula:  

([Page number last page] – [Page number first page] + 1) / 2 + [Page number first page]

If the result is an integer, the middle of the book is the space between the page determined and the previous page. If the previous page ends with a sentence’s end, this ending sentence is the middle sentence.

If the result is not an integer, then the middle of the book is the middle of the page of the integer before the decimal point. The middle of the page is determined by the machinist by counting the lines with text (dedications, headlines, prefixed quotes, in plays the actors' listings are ignored by the machinist) and divides them by two. If the result is an integer, the middle of the page is between the line number of the result and the next line, and the machinist chooses the sentence that covers the line break or that ends before the line break. If the result is not an integer, the machinist chooses the sentence that covers the middle of the line number of the result.

Last sentence

The last sentence is the last sentence of the text of a book. Afterwards, notes of the author to the text, attachments are ignored by the machinist. The last sentence is the last sentence of the story or in plays, the last sentence of what one would see in the performance.



The URL consists of the domain "isbn-maschine", "/", the ISBN number of the book and the name of the displayed sentence in the language of the book, (erster-satz, mittlerer-satz, letzter-satz, first-sentence, middle-sentence, last-sentence) separated by „-„


The title consists of the ISBN number of the book and the indication of the name of the sentence (first sentence, middle sentence, last sentence) separated by a space.


The keywords are the ISBN number, name of the page and the project name "isbn-maschine"


Each page contains the following content:

  • Link to home page
  • Sentence
  • ISBN number with page type
  • Four directional arrows
  • Pause/Start


By default, the isbn-maschine uses the ISBN number in ISBN-13 format. If there is no corresponding ISBN number for the book, the ISBN search engine uses the number format that leads to the entry of the book edition in the National Library of the issuing country.



The isbn-maschine is a machine for the presentation of phrases of literary works on web browsers.

The source for the phrases is the printed book.

Like every self-respecting machine, the isbn-maschine has a machinist. The machinist feeds the isbn-maschine with phrases out of books he has read. Doing so, he follows the rules laid down in the rulebook. This is where his control ends.

The presentation of the texts follows the rules. Neither machine, nor machinist or observer can manipulate the presentation.

The books have complete control.

(The machinist is indeed aware of the fact that the isbn-maschine is actually an automation).


At the beginning, I collected about 50 books and applied different calculation models to the first and last sentences to find out which distribution logic produces the most different colour images.

It turned out that the best way to define the divisor of the three numerical values of the RGB colour space was to examine the occurrence of individual letters.

As a result, the selection of these letters had to be rule-based. For this, I chose the typewriter layout of the respective language of the text.

I limited the texts used to books that I had read and still own or at least have access to the edition I have read. Sometimes I buy again now.


In 2015, I was looking for a way to give texts an autonomous pictorial competence by applying rules that arise from the text. I was looking for a set of rules that could automatically create pictorial works from texts.

The search is still ongoing, but over time, I experimented with translating numerical occurrences of text features and translating them into colour values.

For the translation, the RGB colour space "p { color: rgb(255, 0, 0); } /* integer 0 - 255 */" seemed targetable.

It was now necessary to find out which system could be used to get texts to generate meaningful and varied colours in the RGB colour space. Dividing the maximum value of 255 by a number to be defined was the most sensible model.

I soon realised that this was only possible with a limitation of values.

This led to the idea of using sentences for this model.

This led to the idea of using the own library for this task and presenting this on the internet.

» diese Seite auf Deutsch


other pages: