Hacker News new | past | comments | ask | show | jobs | submit login

I'm wanting to learn APL (I'm leaning towards J or BQN) and Lisp as two languages/paradigms. APL for is expressivity and Lisp for its ML background.

More mainstream I'm wanting to learn TCL and lua.

I don't program professionally so don't need to learn anything to get a job.




I learned a little APL & J, and enjoyed doing anything mathy or with matrices. I didn't like them for generic office stuff.

I read 3-4 books on Lisp and can appreciate the power of it, but somehow I've never really enjoyed using it.

Lua was a bit too barebones for me.

Odly enough, Tcl is where I felt the most comfortable. It's partly because it's a command language, but I also found it full featured and lightweight. I didn't like that there is no official distribution though. There are commercial options and things on sourceforge and I'm not comfortable downloading from them. Anaconda Python has an option to include the Tcl interpreter, Tk GUI, and Wish as Python has a wrapper around Tk.


> I learned a little APL & J, and enjoyed doing anything mathy or with matrices. I didn't like them for generic office stuff.

Really? I found the boring stuff and things like web programming to be the most interesting in J.


About Tcl on Windows, personally i use the one that MSYS2 installs which i assume is the most "vanilla".


I read that Sourceforge changed ownership and the shady downloads are no longer a thing to worry about.


J or BQN are both fine choices for an array language. I prefer J's ascii-based syntax, but it has some rough edges (e.g. namespaces) that BQN seems to have solved more elegantly.


As a disclaimer, Lisp's ML background is different ML than what's mainstream today.


Symbolic reasoning AI is mainstream, it was just never related to ML. The debate was between symbolic reasoning and machine learning in achieving AI. Somehow AI and ML eventually got equated in the last decade (in that ML became the dominant AI approach).

Symbolic reasoning lives well enough in things like business rules, we just don’t associate it so much with AI anymore, nor is it done with LISP.


I thought that "Machine Learning" was also a topic for Symbolic AI.


They are generally posited as in opposition, though there is some research in trying to combine them.


Topics like "rule learning" existed already decades ago. Genetic Algorithms. Decision Tree learning. ...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: