I don't like the comparison either. Because the problem with conventional economics and digital goods isn't the "public good" thing, it's the marginal cost of production.

It costs nothing to produce one more unit of a digital good (or add one more customer to a digital service). Prices in economics are affected by supply, and supply is limited by the marginal cost of production. If there's no marginal cost, then there's no limits on supply, and price tends to zero.

Startup economics agrees with this. If the cost of acquiring a customer via paid advertising is $9.99 and the lifetime value of a customer is $10, then the smart thing to do is to spend every single dollar you can raise on paid advertising (I exaggerate slightly). The value we place on the actual production of that service for each new customer is zero.

This completely messes up traditional economic models.

> It costs nothing to produce one more unit of a digital good (or add one more customer to a digital service).

Having sold digital goods for years, I can attest that this is not true. For example, customers have questions, which consumes one's time. They do returns. Fraud costs. Accounting costs. Taxes on transactions. Bandwidth costs. Hosting costs. Transaction fees from your merchant service account. Etc.

I rolled all of that up into the cost of customer acquisition. The operating costs (bandwidth, hosting, customer service, fraud insurance) are tiny when spread across the entire customer base.

I guess it would be more accurate to refer to the cost of producing one more month of service rather than one more customer. Even then there are costs, but they're tiny compared to the costs of providing a physical service.

Contrast the costs of running Netflix now vs Netflix when it was distributing DVD's in the post (or the costs that Blockbuster had). Adding one more customer to Netflix costs so little now that it's effectively zero, compared to the direct costs of sending out DVD's in the old days.

Just running someone's credit card costs you 3%. This is why gas stations offer discounts for cash.

Ever notice that online companies hide from their customers? As in, no phone number, no email address, no way to contact support, they just direct you to a FAQ. This is because customer service is very expensive to provide.

> No where in here is the web service a public good.

It could be, it just depends on who owns the web service. Could be a corporation, could be a social-minded B corp, or could be a non-profit.

https://letsencrypt.org/ is run by a non-profit, and would be considered a public service. Consider how much "profit" they're removing from the economy by offering the service for free, and yet one could argue the value of every site able to obtain an SSL cert to receive free TLS support is greater than that profit.

The point the author was making was that computer files are fundamentally different than physical goods. But software engineers don't produce "computer files" for sale. They make products and services. Whether it's a file or not is meaningless, and the abstract idea of a "computer file" being a public good is inaccurate.

Well, it does get a bit murky. A book is a physical good. An ebook is not. But if I print an ebook out, it is now a physical good.

TL;DR Things are complicated. We keep trying to find physical world analogies for the digital space, and that just isn't going to work well.

Books aren't really physical goods, economically for all intents and purposes they are very similar to digital goods as the cost of reproduction is near zero which is why books like digital goods fall under copyright law.

And what happens when physical goods become similar to digital goods and can be replicated at zero marginal cost?

It won't happen tomorrow, or a year from now, but it'll happen in our lifetime.

I'm betting on UBI / Star Trek, others see violent revolution / Fallout 4.

Essentially, humanity is freed from 'work'

An ebook is still a physical good in a sense. The data (or some portion of it) rests on a medium in your possesion (even if only temporarily). As we all know, there are significant costs involved in creating, advertising, distributing, maintaining said digital good. The costs and subsidies (e.g. open source software, government research, corsortiums) are just distributed in a vastly different way, so much so that to some it may seem to be a free lunch (despite being far from it).

The cost to create and maintain a public good are not relevantly to its "publicness". Only the cost to access it is.

Clean air and national securty also cost money to maintain, but have no user fees.

Digital objects that are torrentable are public goods. This includes media files and local machine software but not web services.

You could torrent a web service if the app was dockerized, the docker containers were part of the torrent file, and the underlying datastore (files or a DB) were also included.

being a public good isn't a binary state, it's a continuum. I think there are internet services that approach the theoretical definition of a public good (non-rivalrous and non-excludable) close enough to consider them actual public goods.

Edit: I'll give you am example that might fit the definition: DNS

DNS could easily become excludable, if it limited service by IP/MAC address. Most people get DNS from their ISP, for example.

Yeah to be honest, I considered that myself as well, if we take that approach then 100% of internet services do not qualify as public goods (and never will), since it's always possible to limit the service in such a manner.