[kick]

The above post didn't say unpopular, they said 'underpopular'. While that's a bit of a neologism, it just implies that it's less popular than it should be.

Also: "Lisp is the most fun you can have programming, yeah from the start," while subjective, is by most accounts, wrong. Especially when referring to Common Lisp; there could hardly be a more convoluted Lisp than Common Lisp.

[lispm]

I learned much of my initial Lisp programming decades ago with Macintosh Common Lisp. I found it to be incredible fun to work with a tight interactive environment on my Mac.

[mikelevins]

So did I--or, rather, I started with Coral Common Lisp, which is what it was called before Apple bought Coral Software.

I found it hugely fun--so much so that it took over my brain. I've been primarily a Lisp programmer ever since.

Coral Common Lisp offered an extremely easy way in to Lisp programming that was also tons of fun to work with, and an industrial-strength development environment for the full panoply of Mac applications. At AAAI in Detroit in 1989 I mentioned to someone that I was working in Coral Common Lisp and a passer by quipped, "Oh, the good Lisp."

It ran just fine on a 1 MB Mac Plus. You could start it up with a double-click. It launched with a Listener window and a blank Lisp source file (unless you launched it by double-clicking an existing Lisp file with some code in it.) You could save an image, copy the image file to a different machine, launch it, and you would see the same windows in the same state.

Creating a fully-functional Mac window looked like:

    (make-instance 'window)

You could populate the window with working widgets with similarly simple interactions. If you built something you wanted to keep using, you could turn it into a Mac application by doing

    (save-application "Foobar")

If you preferred constructing UI by dragging together widgets by direct manipulation, you could do that, too. Once your UI was assembled, you could tell the Lisp to save the Lisp source code needed to reconstruct it.

The built-in editor, Fred, was a species of Emacs, with the Lisp-oriented features that you would expect, but you'd never have to know that if you didn't want to, because it was also a fully-featured Tesler-style modeless editor at the same time.

The stepper, inspector, apropos, and other development niceties all had Mac graphic interfaces that made them easy to discover and tinker with. They were all written in Lisp, so you could use the Lisp reflection tools to crack them open and poke around inside to see how they worked.

Coral Common Lisp made it easy to get started using Lisp to build apps--about as easy at it can possibly be. At the same time, though, it imposed no arbitrary limits on what you could do with it. When I first met Dave Vronay he was working on the GUI for SK8, and gushed about how easy it was to write window-definition resources (WDEFs) in assembly language using CCL's LAP subsystem. He was an experienced arcade-game hacker, well versed in assembler, but Coral Common Lisp was his first fully-interactive assembler, able to define and integrate new assembly-language code while the program under development was running.

The CCL of today, Clozure Common Lisp, is still a great Lisp, but it isn't as easy or as inviting for newbies. It doesn't have the same graphical environment or the tight integration with the underlying system.

In part that's because when the Clozure Common Lisp project was created (under the name "OpenMCL"), its creators had the rights to the compiler and the Lisp runtime, but not to the Macintosh graphical environment. In part it's because modern macOS is quite a bit more complex than Macintosh System 9 and its predecessors, and development systems have a lot of additional hoops to jump through nowadays to do the kinds of things that MCL was able to do.

If a modern MCL existed, I expect a lot more people would find Lisp a lot easier and more fun to get started with. I think it's perfectly possible to create such a Lisp, but it would be a heck of a lot of work.

[lispm]

I used MacScheme earlier, which also was a huge amount of fun - even though I did not have the extension to access the Mac Toolbox (Toolsmith, IIRC).

MCL was a whole new world then. First with Object Lisp and later with CLOS. I used MCL before it was owned by Apple - at one time it was called MACL (from a market agreement with Franz, IIRC). But later LispWorks was another step up, because it was a big grown up full extended Common Lisp with everything from the commercial UNIX workstation Lisps (like Allegro CL, Lucid CL, LispWorks): it suddenly ran on small and simple to use Apple or Windows laptops, plus it had a Cocoa port.

But, as you describe, the simplicity and integration of MCL into the early MacOS was a lot of fun and there was a lot of tinkering by users. Some friends were still using MCL years after it was obsolete...

[mikelevins]

I started with the full version of MacScheme about the same time that I started with CCL. I loved them both, and at first I liked MacScheme better because Scheme was a smaller, simpler language, and easier to learn. CCL won me over in the end because it was just so much easier to get things done.

Later on, John Ulrich, who owned Lightship Software, the publisher of MacScheme, came to work at Apple and became a friend and colleague. I continued to use MacScheme off and on for years.

I’ve happily paid for Lispworks licenses many times now, but even Lispworks isn’t a good solution for everything I want to use Lisp for. A modern version of CCL’s graphical environment would still be a nicer place for newbies to start. You could even use Lispworks to build a product like that, if the license didn’t prevent it. (Lispworks doesn’t allow their product to be used to make Lisp development systems, which is understandable. They’re a small company who don’t want to put themselves out of business by enabling a customer to give their Crown Jewels away.)

[lispm]

yes, building a product to include development tools (compiler, ...) of LispWorks or Allegro CL would be expensive and not really attractive. I think Franz, Inc. has such a licensing agreement for Allegro CL, but their licensing model is different anyway.

In former times there was a market for add-one GUI or IDE tools. Lucid CL made money from licensing their Lisp to other companies, which then resold those with tools added on (like the SUN Common Lisp IDE). The Action! GUI designer was available for MCL. CLIM was an add-on product. The big Expert System development environments were available as add-on products to Lisp systems.

For a commercial customer something like LispWorks or Allegro CL might still be worth it - they are technically very stable platforms.

[mikelevins]

I tried to work out a way to reimplement my old programmable WYSIWIG word-processor, Alpaca, using Lispworks, but we've been unable to reach an agreement that works for me and for Lispworks. They don't want me to make something that someone could use to compete with their product, and I don't want to make a crippled version of Alpaca.

I wrote the original Alpaca in CCL, using its Cocoa interface. I used it to write a book, and a Japanese company approached me about licensing it for a product, but I fell ill before I could do anything more with it. By the time I recovered enough to look at it again, it had bitrotted. The macOS interfaces had changed out from under it.

I imagine I could fix it up and get it going again, but I'd really like to make a version of Alpaca that works on Linux and Windows as well as macOS. The obstacle is that I need a rich-text engine that can do a good job with page layout, and I don't know whether there's a suitable one, or which Lisp (that isn't Lispworks) would be best to use with it. Qt maybe? Perhaps with ECL? I guess I'd rather use CCL or SBCL. Anybody know of a text-and-page-layout engine or collection of libraries that will work well on macOS, Linux, and Windows with CCL or SBCL?

[mumblemumble]

Yes, thanks for putting it so concisely.

Personally, I do think Lisp is fun, and was having fun pretty much from the start. But, with how many people I've watched just bounce off of it, I simply can't bring myself to imagine that my experience is anything but unusual. Nor am I inclined to console myself with self-congratulatory stories about Blub.

[AnimalMuppet]

I suspect that certain styles fit the way certain people think, and don't fit the way that other people think. If Lisp fits the way you think, then learning it is a revelation - it's like being let out of prison. But if it doesn't fit, it's really hard. And those who find Lisp to be "the right way" (for them) expect that if everyone else really understood Lisp, they'd have the same magical experience. When everyone else does not have that experience, they assume that those people just haven't really understood Lisp yet.

[abrax3141]

I think that this is a function of how you were taught, not a personal preference. An excellent teacher can make pretty much anything interesting to pretty much anyone.

[abrax3141]

Ah. Okay. I misread underpopular as unpopular. My apologies.