Authorship

"The author", states Bernard Cerquiglini, " is not a medieval concept"¹.

There are many participants in the production of a high-quality manuscript. Even if we ignore those who make the paper, the ink, and the binding materials, there are still a few people involved in putting material onto pages, namely scribes, illustrators, and various commentators and editors. All of these people can quite literally make their mark on the manuscript to significant effect. Artists could add images either adding to or distracting from the text, scribes and commentators could add clarifying remarks, editors could rewrite the texts to their liking. In one provocative instance of such an intervention, a scribe rewrites the ending of a poem:

So ends the knight with the lion:

Chrétien’s own romance, it ends here.

There is no more to say or hear,

Nor will more e’er be told anon

Unless some lies are added on.

becomes

The knight with the lion here is closed.

He who wrote it is called Guioz:

In front of Notre-Dame-du-Val

You’ll always find his market-stall²

This drastic change precisely illustrates the significance of these authorial insertions: works are always unfinished, "continuation was the first principle of this literature"³.

Even before the text reached the page, the question of authorship is still present. Many texts circulated as part of an oral tradition, and as such could change slightly between each retelling before finally being written down: oral transmission is necessarily fragmentary⁴.

If we consider a digital artifact such as the Wikipedia page for Poland⁵ (though any similarly storied Wikipedia page will do), determining its authorship is a nontrivial matter. As one scrolls down the page, one sees a large collection of data on Poland: its history, its culture, its economy, and more. If one clicks on the "View History" tab in the upper right corner, they will see a large list of changes attributed to 26 users at the time of this writing, and if one goes all the way back to the oldest revision⁶ one sees a sparse page written by the user WojPob which contains some basic information and the outline of topics that one might assume are expected to be added later. The present-day page is the result of over eleven thousand edits by over four thousand different users⁷.

Who, then, is the author of this page? Before the acts of WojPob, there was no content at all. Though WojPob started the page, each of the editors has contributed to the text in some way, even if they have only made "minor edit"s. So, considering the text of the Wikipedia article for Poland, it might be reasonable to conclude that all of the editors has a fair claim to authorship. However, as a complete digital artifact, the page has a less obvious author. As a Wikipedia page, there is considerable content in the margins and below the page, parts of which were created by some human designer, others are generated by a computer (such as the "Languages" section). Further, this page comes to a viewer from some server running the MediaWiki⁸ software, and is rendered by a web browser (in my case, Firefox ESR 52.3.0). The role of the coders of these pieces of software as authors is unclear. Further confusing the concept, it just so happens that as I write this, Wikipedia is currently attempting to raise funds. This means that on nearly every page, a box with a dark red border appears near the top, which contains a humble request from the founder of Wikipedia, Jimmy Wales, which inserts a new author into most every page. This box also appears when looking at a particular revision, so if one navigates to the 'original' version of the page penned solely WojPob, Jimmy Wales inserts himself as an author several times a year.

Considering the authorship of a significantly large piece of software highlights these struggles of authorship. curl, a "command line tool and library for transferring data with URLs"⁹, authored originally by Daniel Stenberg and additionally by others, is a fantastic example of the difficulty of understanding the authorship of software. On a computer I have access to, curl exists as a 180 kilobyte binary executable, and we can access the 28 megabytes of source code of curl ¹⁰, which "At least 1626 persons have provided code"¹¹ for. These 1626 are not the only entities involved in producing the binary I have. In order to transform the source code to the small binary, the code (of 1626 individuals) enters into a massive system of transformation and remediation. The act of compiling requires a mediator, whose decisions can have significant effects on the capabilities of the produced binary. If one follows a typical compilation¹², one has the option to determine how the binary is installed, whether particular features are to be included, and how to compile the software. Because the curl project does not provide binary downloads, this mediator is crucial to the existence of any curl binary. Compiling curl also requires several dependencies¹³, which are other software projects with their own similarly complicated sets of authors.

As a work, curl has a complicated authorship, but curl also serves to complicate the authorship of other works. curl is open source, and has been used in the products of over 250 companies¹⁴. The license for curl allows this, on the condition that curl's copyright notice be distributed with the product. It is unclear whether the authors of curl should also be considered authors of these derivative works, as most of these authors have zero association with the derivatives, and often are not aware of the use of curl in the derivatives. The confusion of authorship here has played out to very real effect, as curl's primary author Daniel Stenberg has received several emails from crowds as diverse as Instagram and Spotify users to owners of fairly recent cars who seek his help with various problems because they assume an authorial connection between Daniel and the products they own¹⁵.

The concept of authorship for software products, while more well-defined than that of manuscript authorship in some cases, can be equally unclear.