Rust "makes programming fun again".

People will adapt to suffering as long as it's not too bad. I.e. people adapted to asm (coming from punch cards), then to C then to whatever.
I have issues with the rust syntax but I'm guessing that I had to work exclusively with that for a bit I would start to think it was great :P
 
When Rust first came out, I bad mouthed it at first glance. Then I studied it a bit more and found good in it but never messed with it beyond that.
Turn the page to my recent looks and I'm struggling to understand how to use it. It's a bit foreign and complicated to me unlike when I was able to comprehend how to use it way back then.
I still see all this recent bandwagon stuff started up when the government suggested using it. Now everybody has to do it so you can be a cool kid (it seems).
 
I get the feeling Rust is "last years" now; been and gone... maybe zig is the new cool language, or nim; whatever, I'm still using C, but then, I'm not new and cool.
 
Yeah, I'm "industrial strength". I have two great virtues: I work reliably, and I'm stable, over time. Code you wrote in me 20 years ago, still compiles and runs today, with perhaps a few minor tweaks; and code you write in me today, will still compile and run 20 years from now. You could say I'm a "stable genius", to coin a phrase... 😁
 
When wasn't programming fun?
I think it's adding complexity and dependency to limit developer competition. "Memory safe development" --> "you can't access that because you naturally don't know what you're doing. Only approved professionals can go there."
 
The modern s/w dependency situation is somewhat analogous to global supply chains. You wouldn’t need them if you could find totally local sources for all the materials, components, machines as well as skills for whatever you produce. But global supply chains come with their own set of vulnerabilities, dependencies & risks. And for dealing with them you need new requirements, rules, tools, testing etc.
 
When Rust first came out, I bad mouthed it at first glance. Then I studied it a bit more and found good in it but never messed with it beyond that.
Turn the page to my recent looks and I'm struggling to understand how to use it. It's a bit foreign and complicated to me unlike when I was able to comprehend how to use it way back then.
I still see all this recent bandwagon stuff started up when the government suggested using it. Now everybody has to do it so you can be a cool kid (it seems).
My problem with Rust is it's obfuscated chain of custody i.e. backdoors in compiler risk.
I smell secret gov agencies rat.
Rust has longer trust chain. C has more compilers for cross/verification and longer track record.Also Rust has surface attack enhanced by the cargo ecosystem.Not an expert on compilers but seems legit concerns Rust hype has all the fingerprints of intelligence op.
 
So, a kind of honey trap, masquerading as a 'more secure' solution; and recommended/mandated by the goobermint and NSA. Nothing would surprise me.

Which reminds me of Ken Thompson's classic paper, on compiler security.

View: https://www.youtube.com/watch?v=1tMNssI56Hc
 
I think it's adding complexity and dependency to limit developer competition. "Memory safe development" --> "you can't access that because you naturally don't know what you're doing. Only approved professionals can go there."
You're not smart enough to figure it out yourself; now step away from the keyboard, citizen; we know what's best for you...

chained-kbd.jpg
 
I have seen a lot of writeups about Rust being unsuitable for quick prototyping because you need to get everything just right before anything works. Namely the borrow checker.

That sounds like the opposite end on the spectrum of languages you can specifically call "fun".
I have asked someone who has prototyped nontrivial operating systems in Rust....
So, a kind of honey trap, masquerading as a 'more secure' solution; and recommended/mandated by the goobermint and NSA. Nothing would surprise me.

Which reminds me of Ken Thompson's classic paper on compiler security.
While this is a possibility, I don't consider governments or their TLA agencies to be smart enough or pervasive enough to pull this off. They can't even do (what's supposed to be) their main damn job right!

The point of type & attribute systems is to be able to express constraints on values a variable can store or function can accept or produce. But the more you specify, the harder it gets to change things. There is a happy medium but what you need for the cottage industry of writing tools for yourself is likely to be very different from industrial strength mass produced software, many parts of which are from third parties. Just because the govt specifies some of this doesn't necessarily mean they have an ulterior 1984-ish motive! Almost always bureaucracies do things from a CYA point of view.
 
You're not smart enough to figure it out yourself, now step away from the keyboard, citizen...

View attachment 26784
We're good on the way. Keyboards are already sofrware-abstracted. You can't do a lot with a USB-keyboard and some wires. Still need PC-standard and a high-level USB driver to read or send something. Phones are ruled out completely. That's a touchscreen layer on 1 specified fake-open kernel. Everhting happens far from the metal.
IBM must do something. This wasn't the idea. 😆
 
Teams writing C and C++; even expert teams at Microsoft and Google with world-class security review processes; cannot eliminate memory safety bugs at scale. The bugs are too subtle, the language provides no systematic protection, and the attack surface is too large.
See, this is what gets me. Who wrote the compiler for Rust?
 
When wasn't programming fun?
When I walked into my first and only "agile shop". the model was generic coders taking whatever JIRA ticket the idiot lead decided to assign, expecting that every coder could do any ticket that came thru in the specified amount of time...and then being performance evaluated simply by ticket burndown rates...and lets not forget the "team standards" based code reviewed. based on 30 yo paradigms and which forbid use of any construct that the idiot lead and his golden haired boy didn't understand. The best way to describe that place was "untapped potential" they had good people, but misused them terribly, and those who didn't stroke the "idiot managers" egos left after a few months.
 
The point of type & attribute systems is to be able to express constraints on values a variable can store or function can accept or produce.
Sounds like <assert.h> to me. Yeah, I'm being facetious. Of course, Rust does have some technical merits, moving that type of constraint validation, from run-time to compile time.
 
When I walked into my first and only "agile shop". the model was generic coders taking whatever JIRA ticket the idiot lead decided to assign, expecting that every coder could do any ticket that came thru in the specified amount of time...and then being performance evaluated simply by ticket burndown rates...and lets not forget the "team standards" based code reviewed. based on 30 yo paradigms and which forbid use of any construct that the idiot lead and his golden haired boy didn't understand. The best way to describe that place was "untapped potential" they had good people, but misused them terribly, and those who didn't stroke the "idiot managers" egos left after a few months.
All they need is "warm bodies" who can do "baby steps". Sounds truly terrible. Don't think I could work in a shop like that. Not sure I'd last the week.
 
All they need is "warm bodies" who can do "baby steps". Sounds truly terrible. Don't think I could work in a shop like that. Not sure I'd last the week.
I lasted six months. They canned me because I refactored a 2000 line long struct copy module to be about 150 lines, using pointers instead of explicitly referencing every horrendously long nested struct assignment. Of course by this point I was outspoken about my displeasure with the stupidity of technical management.
 
I lasted six months. They canned me because I refactored a 2000 line long struct copy module to be about 150 lines, using pointers instead of explicitly referencing every horrendously long nested struct assignment. Of course by this point I was outspoken about my displeasure with the stupidity of technical management.
That deserves a :beer:for staying power, but you gotta do what you gotta do. Ah well, that kind of refactoring would reduce their KLOCs count, which probably means some bean counter thought that they had been less 'productive'... as we say over here, "never mind the quality, feel the width" 😉
 
Back
Top