I'm just a designer/developer so I don't have deep knowledge of most of the tech world and I'm sure there are some crazy stories out there I am unaware of. But as I said, I'd be more than happy to hear them.

There’s some excellent voip and sip solutions. Asterisk has been packaged into many a reliable flavour.

The other end of the spectrum is Microsoft that says "Here's what you get in an E3 license, take it or leave it, because trying to put together an enterprise license that only has what you want to pay for is 1) gonna be more expensive in the end, and 2) we make it as hard as possible to do that.".

Some of this, too, is due to Cisco’s longevity in the market, the sheer number of product lines they’ve gone through, and the number of companies they’ve absorbed.

You buy a company, you can’t just cancel all the existing licenses for those products, so you have to keep those licensing systems and pricing packages around for those customers that keep renewing year after year. Only after a while, it becomes more trouble than it’s worth to integrate those into your existing licensing packages (especially if you’re figuring you’ll get around to EoLing that stuff when no one’s using it anymore), so they just stick around forever.

You kill off a product line and replace it with something, existing customers say, “Great, I’ll buy the new thing with the options I had on the old one”. Except those don’t map cleanly, so now you’re developing odd licensing packages for people who want the Super-Duper Routing 9000 but only with the same leather trim and manual shifter they had on the Routing 8000. If the customer is large enough you’ll bend for them, especially the big national ones with clauses in their contracts that require the forfeiture of dragon-blanching amounts of gold if you ever can’t sell them exactly what they bought last year.

If the customer is large enough you’ll bend for them

If the customer is large enough Cisco will produce a custom IOS build to your spec. Made some consulting gigs really interesting.

Most foundries won't do that for fonts (at least for web).

Now, completely without benchmarks, I can tell you that my workplace has migrated from Oracle to Postgresql and performance is essentially the same: a few things were a little faster on Oracle, a few a little faster on PG. All revenue is directly or indirectly dependent on the database systems, which have been continuous for 20+ years: this is not a side-project.

PG administration is much, much faster. Starting an Oracle database always involves a wait similar to booting an elderly operating system. Starting a PG database is fast. Things that Oracle will want to up-charge you to an Enterprise Edition license for, PG has for free. Want to spin up a new machine? No need to negotiate a new license with PG. The cost of the Oracle license would routinely be so large that you need to buy awesome hardware to justify it. You can install PG on anything, and justify it.

I ask because NVMe has neutralized a ton of performance differences across data related products.

I'm not in that kind of world so I'm not familiar with Oracle business practices. And I agree, there's usually no middle ground and it's both annoying and bizarre.

This is a general problem in many businesses. Back in the early days of blogging, Joel Spolsky wrote an entire article about how there was basically no software that cost more than $1,000 and less than $75,000. This was more than a decade ago, the numbers today would be different because we care about LTV not sticker price, but the gap would still exist.[1]

Of course, heuristics like this are not strictly true, but the general shape of his argument is still correct: There are three kinds of monetization models for software: Free, cheap, and dear:

1. Free as in beer. Open source, etc.

2. Cheap. $10 – $1,000, sold to a very large number of people at a low price without a salesforce.

3. Dear. LTV of $75,000 – $1,000,000, sold to a handful of rich big companies using a team of slick salespeople that do six months of intense PowerPoint just to get one goddamn sale. The Oracle model.

Net result is that some products are sold to people to scratch their own itch, and some products are sold to businesses that can justify the ROI of spending a large amount of money, and between the two it is hard to find reasonable options.

We see this in all the products that start at a low price and solve a simple problem, like DropBox and 1Password. Then they pivot to "The Enterprise" and you are now talking subscriptions, seat licenses, and all sorts of fancy enterprise features to justify the price.

I used to see this in hotels when I travelled on business. When my clients were paying, they put me up in fairly luxurious hotels that charged me $3 for a room service coffee. When I was paying, I stayed in Motel6-type places that had free coffee in the room or at the end of the hall. I think everyone eventually offered free coffee in the room, but the last time i traveled there was a similar dynamic with WiFi: The naive would expect the cheapest accommodation to try to rent you internet access, but it was the other way around. "You have an expense account? Ha! Pay for WiFi!"

Same dynamic: If you are selling to business, you have to have the sales force, customer success organization, and features business want to buy (like back-end integration with business expense and planning systems). That all costs you money, which you recoup by charging an arm and a leg.

TL;DR: The moment you start selling to business, you are on a path to charging huge money for something that when sold to individuals, ought to be affordable.

[1]: https://www.joelonsoftware.com/2004/12/15/camels-and-rubber-... "Camels and Rubber Duckies"

One model you'll see though in this market segment and VARs and other channel providers. They'll have set of enterprise vendors they work with, but take on the sales and service aspect themselves. So while it might not make sense to have dedicated sales and customer success for a $5k piece of software - if you are effectively bundling 10 of these from different vendors + profserv, a $100k/deal starts making sense of the VARs, and offloads the costs from the vendors who are then just giving bulk pricing to the channel.

My very first personal ISV "hustle" was a classified ad product for desktop publishing. I priced it at $750, and discovered that my cost of sales and support made this untenable. Now I know why small desktop publishers were doing classified ads by hand, and the really big companies had mainframe- and minicomputer-based systems costing five or six figures in the 80s and 90s.

Same dynamic, and all it took was my car being repossessed to make me realize there was something wrong with my business model.

Years ago, Philip Greenspun wrote a post noting that there was no real middle ground between a magazine article and a book. (This was in the context of the web maybe addressing this--though it mostly hasn't, longreads and things like that aside).

Something either fits into low single-digit thousands words length or it needs to be a 250+ page book. In some ways, shorted attention spans make this sort of thing even worse. We still have books but, in my day job, we've mostly found that the 3,000 word white papers we used to crank out have largely been replaced by blogs and blog series.

Unless you have something to sell, and presto, so many essays and blog posts are really advertisements for the author or their employer.

ETA: And now I see it's in the context of non-fiction, ignore this. But I remember seeing novella-sized non-fiction books around about a decade ago, they were just not that common.

I can't get past thinking "why would you do that", especially for something you're having to pay for to use, and could break your web site if something goes wrong. That seems like the poster child for "hey, find an alternative, any alternative, that doesn't rely on a paid third party for something as trivial as how a header looks".

* any Serif font (Times, etc, they all look the same except for 1-pixel differences)

* any Sans serif font (Arial, Helvetica, it's all the same thing)

* Serif monospaced (Courier)

* Sans serif monospaced (Consolas etc)

* Silly unreadable fonts like Papyrus. Sure, technically, there are lots of these, and they do look different, but there's no reason to use one of these, unless you just want to make your site difficult to read.