Hacker News Reader

Comment by sifoobar | Hacker News Reader

Every coder should if you ask me; I consider C, Forth & Lisp to be the brightest shining stars in PL design. Each of them offers a unique approach cut down to it's core. If you prefer a more convenient and less dogmatic approach, I'm working on a hybrid scripting language [0] that might be of interest.

https://gitlab.com/sifoo/snigl

View on HN

Forth and Lisp are genius and real power tools for lone hackers. They are unsuitable for projects with many developers though. Forth in particular has driven the philosophy of not-invented-here roll-your-own to the extreme. That is completely fine for short code but "brightest shining stars in PL design" is too much.

> They are unsuitable for projects with many developers though.

Please go tell this to everyone on the GNU Emacs mailing list, I think I am not the only one who missed the memo.

I have to admit that I did not have emacs in mind. With 1.6M lines and 934 contributors we can consider it a sizeable project. https://www.openhub.net/p/emacs

Now, unsuitable does not mean impossible. It is just harder than with other languages. Which means you will hit a complexity ceiling earlier.

Also, Emacs is just a text editor and platform for anything. Java with Eclipse has eclipsed (hihi) it. Why hasn't the Emacs community made it into a real IDE with refactoring for example? They certainly have been challenged often enough. Why have multiple companies built new IDEs from scratch instead of building on Emacs?

Can you please stop trolling? You make assertions about GNU Emacs development while obviously having no clue about it. Your last question is nonsense (why don't those companies just not build on Eclipse instead of making new IDEs? Has it occurred to you that maybe it is because they want to make a new program?). Programs like Eclipse don't offer even a tiny fraction of the features that I use GNU Emacs for every day, and I only use a small fraction of the different GNU Emacs features and packages.

Have you ever tried working on an Eclipse plugin? You obviously have no experience writing extensions to GNU Emacs if you think that "It is just harder than with other languages." You are living in Java fantasy land.

There are literally dozens of Emacs packages for different programming languages offering refactoring support. The vast majority of Emacs development is done outside of the main repository. People decide to build on built-in GNU Emacs code, on other third-party packages, and sometimes third party packages get merged into the main repository. There is a huge group of different development communities around GNU Emacs that is unmatched by any IDE, and few other applications at all, right now. GNU Emacs is an overwhelmingly successful example of a project with a large number of developers, and among other things it shows that you do not need namespaces (modules) or OO for this.

Ever book a flight through Google? Common Lisp touches that.

So does COBOL.

Forth in particular has driven the philosophy of not-invented-here roll-your-own to the extreme.

To be fair to Forth and Forth programmers, Forth is very often used in environments where they are running on the iron and need to roll there own. There were environments where Forth used the local libraries like Forth on the early Macs[1]. You could access the Macintosh Toolbox just like the other languages.

They are unsuitable for projects with many developers though.

I don't know about that. I get the feeling that Forth would be fine but require a common, shared dictionary for the project which isn't far from a common class library used in other languages. I'm not sure it would require more interaction than something like Smalltalk which has been used in large projects.

1) Forth was an early Macintosh language because of how cramped 128Kbytes is.

Reply on news.ycombinator.com