API: Invented or discovered?

I found something that interested me in a Wikipedia article about the term API. The term, “application programming interface”, was first recorded in a paper called “Data structures and techniques for remote computer graphics” at a conference in 1968. However, Wikipedia then says the following.

“The idea of the API is much older than the term. British computer scientists Wilkes and Wheeler worked on modular software libraries in the 1940s for the EDSAC computer. Joshua Bloch claims that Wilkes and Wheeler ‘latently invented’ the API, because it is more of a concept that is discovered than invented.”

It is the last sentence that attracted my attention. There is a parallel question in Mathematics: Is Mathematics discovered or invented? Careful how you answer that one as there have been differing answers by both Mathematicians and Logicians. But, I have never seen it posed in Computer Science before.

I know next to nothing about APIs and so can’t give an opinion, but what do others think?

For the Wikipedia article and references, enter “Application programming interface” in Wikipedia search field.

Is Mathematics discovered or invented?

The mathematics of the universe IS, the terminology to describe it is invented.

Computer Science – in the devices we use this is simply patterns of 0’s and 1’s which could occur, and probably have, totally naturally.
Discovering one that is pleasing in some way is discovery and finding a means of reproducing it regularly is programming.

Is Mathematics discovered or invented?

This is my cue for recommending the book Mathematics: Form and Function by Saunders Mac Lane, ISBN-13: 978-1-4612-9340-8, published in paperback by Springer-Verlag in 1986. This book is intended to describe the practical and conceptual origins of Mathematics and the character of its development – not in historical terms, but in intrinsic terms. Steve squares the Platonic with the Logicist answers. Saunders wraps both in a Social answer.

Two things:

1. I wrote that while I was waking up¹
2. Unless someone drags some RO content in, this thread needs to be restarted in Aldershot

So, on the first point – clearly my subconscious² is “interesting”

¹ Sequence is get out of bed, go downstairs and feed cats and/or release them to the wilderness that is the garden, put kettle on, eat, twiddle with computer, make tea, twiddle with computer.

APIs are initially invented, but tend to evolve in ways never envisioned by their creators, and are discovered by programmers using them.

Does that cover all bases?

@ GavinWraith

This is my cue for recommending the book Mathematics: Form and Function by Saunders Mac Lane, ISBN-13: 978-1-4612-9340-8, published in paperback by Springer-Verlag in 1986.

You totally rock man, I wanted to answer this last night, but I couldn’t remember the book title with the details and then today boom you posted it, smashing :)

For Mathematics there are many different views as to the nature of its statements and how they are validated including Formalism, Logicism, Intuitionism, Structuralism and Platonism. Steve’s view

The mathematics of the universe IS, the terminology to describe it is invented.

is Platonism or close to it. The problem then is: Where are 100-dimensional or even infinite-dimensional, e.g. Hilbert spaces, to be discoverd; where do they exist? Let alone the infinite number of different size infinities that were ‘discovered’ by Georg Cantor. Plato’s answer, had he known about these topics, would be that they are ideal forms which we humans imperfectly understand. You might say a rather unworldly point of view, but one that seems to be true for many research Mathematicians.

Well according to various sources, in the English language:

• Discoveries apply to those that have been long existent, while inventions, to those that have never existed in the past.
• Discovery is made by providing observational evidence and attempts to develop an initial, rough understanding of some phenomenon
• To invent is to plan and produce something to meet specific purpose
• Invention does not necessarily involve exploration, whereas discovery necessarily involves exploration

So considering the above and the original definition of Application Programming Interface and restricting the judgement to solely Computer Science it’s clear that API are an invention. However, the nature of API especially when extending the analysis to mathematics and the intrinsic nature of a mathematical function (which undoubtedly recall the early forms of API) is in fact a discovery given that such properties have always been there…

Off to Aldershot