after a nuclear war, the remaining people will probably not be able to spin up a modern operating system on their improvised chips. How do you build a simple, reliable, legacy-free OS from scratch? What ideas 💡 and techniques should be passed down to those people?
If we think hard enough about this, I think we’ll agree that closed-source systems are basically designed to be almost impossible for people outside the sponsoring organization to reproduce (for an example, consider [ReactOS](https://reactos.org/), which launched as [a project to produce a system compatible with Windows 95](https://reactos.org/wiki/FreeWin95) and [then changed to focus on Windows NT](https://reactos.org/wiki/ReactOS/History), and after more than 25 years, is still not capable of being a daily use system.
But we may also determine that most open-source systems are likewise not designed in such a way that reconstruction is viable. The Linux kernel is *huge* these days.
Additionally, in my opinion, they’d probably want to use programming languages designed for readability, ease of learning, and error-reduction first (that is, more like #COBOL than #C, more like #Java than #CPlusPlus, more like !Smalltalk and #Lisp / #Scheme than #Perl / #Raku and #JavaScript) and then performance and low-level access.
I think it is a mistake to assume that one could start with a modern version of #gcc or #llvm or #msvc … because it is not a given that the software itself and someone who knew how to use it (and update, modify, and adapt it) would still exist.
A couple of things I've heard !Smalltalk people say killed their market: 1. Smalltalk's image-based system didn't mesh well with OS filesystems; 2. #Smalltalk vendor turmoil and their high pricing. (I wasn't even aware of the language's existence until the late #1990s. I took numerous programming courses over the years, but I never heard ST mentioned until another student mentioned it in a #Java course. ) CC: @clacke@libranet.de
Tried out subbing to the "BSD Now" podcast. Interesting, but the two episodes I've listened to are about four times too long and filled with babbling. Within 5-10 minutes, I've tuned out. #BSDNow
I think I'm going to unsubscribe, which will leave me with two: * Hacker Public Radio -- I've resubbed for the past couple of weeks and I have not burned out yet. Maybe I'm just a fan of @klaatu ;-) #HPR * Smalltalk Reflections -- This one is on another unannounced hiatus. I should probably take a look around and see whether there are any other !Smalltalk podcasts in currently active production.
I haven't really looked for them yet, but I'd be interested in trying out podcasts that focus mostly on !TclTk or !DLang.
> OMeta is a new object-oriented language for pattern matching. It is based on a variant of Parsing Expression Grammars (PEGs) which we have extended to handle arbitrary data types. OMeta's general-purpose pattern matching facilities provide a natural and convenient way for programmers to implement tokenizers, parsers, visitors, and tree transformers, all of which can be extended in interesting ways using familiar object-oriented mechanisms.
So it's even buzzword-compliantly up-to-date. The official implementation is in !smalltalk but there is one in !scheme too, and other languages.