> What is $500 for a tool if it really did make you more productive?

What is $500 for a tool to you if you make lots of money? Sure, that is one question.

The more relevant question is: what is $500 for the tool, to the tool?

In 2020, what $500 for a tool is a kiss of death. Nobody learns it; no third-party ecosystem, no jobs, ...

OK, you can have some free-with-limitations student or personal editions or whatever. In the FOSS-dominated landscape, that's still a kiss of death. The entire quality and scope depends on the in-house development behind closed doors, which is funded by the $500 licenses.

I think, without the multiple efforts to have free Lisp over the past 3+ decades, Lisp would be history.

Microsoft Office, Microsoft Visual Studio and also Jetbrains wants a word with you.

Personally I honestly prefer

- LibreOffice (much better for getting styling consistent last I checked + not spreadsheet is not as obsessed with mangling everything into an American date

- VS Code (less overhead, more ergonomic and seriously: who develops things that needs old Visual Studio these days who hasn't done so for years?)

- NetBeans (yes, I use Jetbrains but only because I have to use Kotlin. NetBeans just works and the shortcuts makes sense to me.)

but that doesn't prevent the above mentioned products from earning lots of (hopefully) well deserved money.

> Microsoft Office, Microsoft Visual Studio and also Jetbrains wants a word with you.

These days I use LibreOffice on Windows (though very rarely). Microsoft Office isn't free and doesn't come preinstalled, so screw that.

Microsoft Studio has free versions. I have three versions of it on various hard drives; all were free, and with all I was able to produce a program and ship it royalty-free.

But to stay on point, you were arguing against a $500 tool because "nobody" would learn it and there would be "no third-party ecosystem" - that's emphatically not true for the tools above, regardless of your personal preferences.

None of these tools are programming languages or mainly programming languages, except Visual Studio, which you can easily use free.

The point isn't that nobody pays for any kind of software. I mean, we could whip out enterprise software as an example. "Hey, the company I work for pays 700K per year for SAP support; whatddya mean $500 will kill a programming language tool?"

Pycharm and VS both have free / hobby versions that is, arguably, of high quality.

It's in line with what you'd pay for Visual Studio. VS Pro licences are $45 a month and Enterprise licenses are $250/month. It's fair to ask $500 for hobby and $3000 for pro for a perpetual license for LispWorks

Turbo Pascal, Turbo C, Turbo Prolog, all things I got at an age when it was just a hobby in high school. In fact I can still look over at my bookshelf and see turbo prolog books.. some things you cant bring yourself to throw away!

     I use SBCL, just for hobby and it does everything I want but I agree that the price is not a lot for the tool if it offers something you want for that amount.

     We have got used to not paying for things and its not money we don't have, we have just spent it somewhere else because we have not had to spend it on tools but the price is not over the top - only if you have never had to pay for something.

Same here, despite piracy being quite common in 1990's Portugal, I bought student editions of Turbo Pascal and Turbo C++, as high school student.

Another point is that a lot of Common Lisp code seems to implicitly assume SBCL, so I don't know what, if anything, I can get through Quicklisp will run on Lispworks' little interpreter. That might not be a big problem if I were to ignore external libraries in favor of some preexisting codebase.

It would be interesting to find the places where Quicklisp systems depend on some quirk of SBCL that's not required by the standard.

I know of one case here: one depended on string constants having element type CHARACTER, when that is not required by the standard. It's ok for an implementation to cause a string like "abc" to have element type BASE-CHAR, for example.

Another example might be assuming that various standard functoins that return booleans return T for true. Almost all are not required to do that.

I used Lispworks alongside SBCL and other Common Lisps for about 16 years. Once in a while I had a problem with some quicklisp library in Lispworks, but it was infrequent.