What will programming look like in 2029?
4 comments
I suspect the languages and tools will look broadly similar. Not because they're currently any good, but because the next 8 years will be focused on different domains rather than tooling.
Writing contracts or transactions (monetary or other property) to a blockchain will become much more commonplace. Terms of service, especially around privacy and ownership of data, will be recorded in immutable blockchain records.
AR, and NLP will be "standard" in nearly all new apps. Apps will interact in more conversational ways using NLP and will trigger more real-world feedback via speakers, lights, video projection, etc. Boring stuff like movie theaters, coffee shops, malls, amusement parks, and some bureaucratic offices (e.g. DMVs) will have ways of personalizing your visit via the interaction between and app and the physical space. A lot of programming effort will go into deploying these systems. There will be at least one bad joke starting with "So a phone walks into a bar..."
Most programmers will be tasked with integrating all the old with this new stuff. There will be APIs, maybe some new serialization formats like JSON, and some new terminology to master, but most of the new will be in user experience.
Writing contracts or transactions (monetary or other property) to a blockchain will become much more commonplace. Terms of service, especially around privacy and ownership of data, will be recorded in immutable blockchain records.
AR, and NLP will be "standard" in nearly all new apps. Apps will interact in more conversational ways using NLP and will trigger more real-world feedback via speakers, lights, video projection, etc. Boring stuff like movie theaters, coffee shops, malls, amusement parks, and some bureaucratic offices (e.g. DMVs) will have ways of personalizing your visit via the interaction between and app and the physical space. A lot of programming effort will go into deploying these systems. There will be at least one bad joke starting with "So a phone walks into a bar..."
Most programmers will be tasked with integrating all the old with this new stuff. There will be APIs, maybe some new serialization formats like JSON, and some new terminology to master, but most of the new will be in user experience.
Yes I hope all the old C code will be replaced.
There might be some awesome C developers on the planet, no doubt. But most legacy, last millennium C code was written by ordinary guys, with no internet, stack overflow and code reviews. The last "experts" (only guys who understand it and can change things) who were writing it, will soon retire and then nobody can add major features anymore or it will take forever plus fluctuation suddenly raises (of course 0 unit testing was done, variable names at max 10 characters long, 6000 lines per file and sometimes function, at least 100 mutable global variables with undecipherable short names, raw pointer magic everywhere, and 3 times as many mutex locking as needed - just to be sure no data races happen, programs crashing on teardown, because of all the lifetime issues, etc.).
I hope more and more companies choose Rust for such ports in 2029. The ports to C++ from the last decade were mostly done by C++ devs with C practices mentioned above (sometimes the same guys while learning C with classes style), making it worse ("Singleton lets me program like C, pretending I'm doing OOP yay"). I've seen so much of this stuff that I lost a bit of my trust in technology, at least in certain areas like automotive and embedded devices with lots of generations. And I know there are a lot of awesome C and C++ programmers also. But sadly not in these generations of legacy code I had the honor to work with. And this stuff is waiting to be rewritten for sure. But please not in C, now that mankind has Rust (yes fanboy here, but you can rewrite it in Go too and I'm fine with it, heck use JavaScript... but please get rid of these legacy C codebases).
There might be some awesome C developers on the planet, no doubt. But most legacy, last millennium C code was written by ordinary guys, with no internet, stack overflow and code reviews. The last "experts" (only guys who understand it and can change things) who were writing it, will soon retire and then nobody can add major features anymore or it will take forever plus fluctuation suddenly raises (of course 0 unit testing was done, variable names at max 10 characters long, 6000 lines per file and sometimes function, at least 100 mutable global variables with undecipherable short names, raw pointer magic everywhere, and 3 times as many mutex locking as needed - just to be sure no data races happen, programs crashing on teardown, because of all the lifetime issues, etc.).
I hope more and more companies choose Rust for such ports in 2029. The ports to C++ from the last decade were mostly done by C++ devs with C practices mentioned above (sometimes the same guys while learning C with classes style), making it worse ("Singleton lets me program like C, pretending I'm doing OOP yay"). I've seen so much of this stuff that I lost a bit of my trust in technology, at least in certain areas like automotive and embedded devices with lots of generations. And I know there are a lot of awesome C and C++ programmers also. But sadly not in these generations of legacy code I had the honor to work with. And this stuff is waiting to be rewritten for sure. But please not in C, now that mankind has Rust (yes fanboy here, but you can rewrite it in Go too and I'm fine with it, heck use JavaScript... but please get rid of these legacy C codebases).
I'd say much the same as it is today. If I think how much it's changed in 10 years, it doesn't feel all that different. I'm still writing ruby most days. The focus will switch to even more web based. Likely more security though.
about the same as in 1980
For many still like vim I guess.
https://news.ycombinator.com/item?id=28656204