"It's called Dylan because it's so cool that nobody understands it." -- a manager in the Newton Group
[Kind of tounge-in-cheek; I didn't see that it was particularly difficult to grok]
The Newton was going to be programmed in Dylan from the metal on up. The week that it was decided to write it all in C++, there were a lot of glum ex-Dylan programmers wandering the halls, clutching copies of the C++ ARM.
One of them (from the Bauhaus) group is working with us now on some changes to look at adding alternate syntax support to bring back s-expression syntax. :)
Although he didn't move over to C++ Newton, he went on to SK8 and other things...
Maybe I had a pretty weird upbringing, but the design of Dylan is one of a rather small set of things that gives me a particular kind of wholesome, full-body happiness, and a feeling of real warmth, and like somebody out there truly understands me.
I think it's something like what they call "love." But no -- it's like my mind is getting hard.
Functions as first-class values, lexically-scoped closures, map, curry...objects, methods, classes & class inheritance...immutable objects that can have their pointers optimized-away by the compiler.
This sounds ahead of its day for a language intended for industry.
Pretty sure if they were to add a whitespace sensitive syntax ala python/coffeescript and get it compiling to js they'd be riding to victory. From what I can tell there is actually a project https://github.com/turbolent/ralph to bring dylan to the web.
Later version of Dylan are a much better language, Algol-syntax not withstanding. Better still, is GOO: a very elegant little language that would make for a perfect Dylan or Common Lisp core.
Now that we have someone looking at adding an s-expression syntax as an alternative to OpenDylan, it has been interesting reading through my copy of the original Apple Dylan book from 1992 (originally owned by P Tucker Withington given the signature inside the cover). A lot did change for the better between 1992 and 1996. I think sealing is a big example of that. I think the 'block' construct is a bit nicer than the prefix-Dylan equivalent as well.
It is also interesting to see how some elements of the current Dylan syntax make more sense when you see the roots of it all in the prefix-Dylan syntax.
Recognizing that you're probably just shilling for your own product ...
We use http://sphinx-doc.org/ and author in ReStructuredText with some extensions. From that, we output HTML, PDF and ePub.
There are some issues with the ePub (they display fine, but don't validate), so it looks like sometime soon, I'll be hacking on Sphinx to fix some of the bugs in the epub generation.
[Kind of tounge-in-cheek; I didn't see that it was particularly difficult to grok]
The Newton was going to be programmed in Dylan from the metal on up. The week that it was decided to write it all in C++, there were a lot of glum ex-Dylan programmers wandering the halls, clutching copies of the C++ ARM.