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

Alan Kay is one of my intellectual heroes. I've always been enamored by the Macintosh, which was influenced by the work at Xerox PARC. However, it wasn't until I started studying more about Xerox PARC within the past five years that I received a better understanding of the vision of Alan Kay's work. Over the past five years I've gone from a big Unix and C fan (who loved the NeXT heritage of macOS but who didn't program in Objective-C, basically treating macOS as a BSD that could run iPhoto) to someone who is in love with Smalltalk and Lisp. I just started reading "The Art of the Metaobject Protocol," a book that Alan Kay has spoken highly about.

One recent project from the now-shuttered Viewpoints Research Institute that I've taken particular interest in is STEPS, which is an attempt to develop a minimum-viable desktop operating system environment using a minimal amount of code. The STEPS researchers achieved this goal by defining domain-specific languages to describe and implement subsystems and applications. One of the things I'm curious about is how well could something like STEPS be implemented in a language like Common Lisp which provides macro support and also CLOS. I'm curious if the work done in STEPS could be applied to production desktop environments? The GNOME and KDE stacks rely on a lot of C and C++ code, respectively, and it would be interesting to see if the use of very high-level languages would result in a less complex stack. Granted, GNOME and KDE need to support complexities such as accessibility and internationalization, but I'm wondering if the lessons of STEPS can still be applied even to address these considerations.

I wonder what advice Alan Kay has for young researchers who want to do long-term, potentially revolutionary, open-ended research in an age that strongly favors short-term, directed research that tends to be incremental and iterative. Alan Kay came of age during a pre-Mansfield Amendment ARPA (where ARPA used to heavily fund basic research in computer science before the Mansfield Amendment of 1973 required ARPA to focus on funding research that was directly related to defense) and also during the golden age of industrial research labs, which ended in the 1990s. However, he has seen the landscape transform to the current situation.

Happy birthday, Alan Kay! He's inspired me to discover and encourage the beauty of computing, and his philosophy regarding research strongly resonates with me.




I share the same feelings.

UNIX seemed a better architecture than the archaic MS-DOS/Windows 3.1 (not so much when compared against Amiga).

However thanks to my university I was able to plunge into Xerox PARC universe and its influences on ETHZ, and the more I learned reading their papers and manuals, with playing with Smalltalk/V and Native Oberon, made me a very grumpy UNIX user and eventually back into Windows/macOS world a couple of years later.

Picking up on your examples, ironically with DBUS, gobject, KParts, gRPC one could replicate the experience of such workstations on GNU/Linux, even something similar to OLE (which goes back to Xerox PARC inline documents), or OS wide REPL instead of plain old UNIX shell, but given the fragmentation of Linux distributions it is in vain to try to offer such experience.




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

Search: