Comment by pdimitar
3 days ago
RE: make/CMake, consider yourself in a somewhat privileged bubble and IMO let's leave it at that. Every single time I had to deal with make or CMake, I regretted it: every time when I used them in anger and to more of their capacity, with barely any exceptions. The only safe usages of make were like 50-line files max. And even there I had to manually eyeball the .PHONY targets to make sure I know which sub-commands I can invoke that don't depend on file timestamps because of course somebody decided it's a good idea to liberally mix those with the others. Sigh. Don't ask.
I don't _hate_ them or anything. They are super solid tools and I have derived a lot of value out of them in a previous life. But they leave a lot of room for overly clever humans to abuse them and make life hard for the next guy. Which is exactly how I was exposed to them, dozens of times.
The downfall of all flexible tools or PLs is that people start using them in those obscure and terrible ways. Not the fault of the tools or the PLs.
If you can control 100% of the surface where you are exposed to make/CMake then that puts you in a fairly rare position and you are right to make full use of it! Go for it. Deep work and deep skills are sorely needed in our area. I am rooting for you. :)
> The issue is that I think it (Rust) is going to worm it way everywhere and I will be forced to deal with it.
Help me understand the actual technical criticism buried in the "worm its way everywhere", please. And in the "it's good for CV/Resume padding" statement as well.
It's a strange thing to say and it smells like a personal vendetta which is weirdly common on HN about Rust and to this day I have no idea why even though I have asked many people directly.
Rust has objective technical merits and many smart devs have documented those in their blogs -- journeys on rewrites or green-field projects, databases, network tools (like OP), and even others. Big companies do studies and prove less memory safety bugs over the course of months or years of tests. The Linux kernel devs (not unanimously) have agreed that Rust should no longer have experimental status there recently -- and people are starting to write Linux drivers in Rust and they work.
I am honestly not sure what would satisfy the people who seem to hate Rust so passionately. I guess it announcing full disband and a public apology that it ever existed? Yes this is a bit of a sarcastic question but really, I can't seem to find a place on the internet where people peacefully discuss this particular topic. (I have seen civil exchanges here on HN of course, and I love them. But most of the civil detractors ultimately simply admit they don't have a use for Rust. Again, that is very fair and valid but it is not an actual criticism towards any tech.)
> I don't _hate_ them or anything. They are super solid tools and I have derived a lot of value out of them in a previous life. But they leave a lot of room for overly clever humans to abuse them and make life hard for the next guy. Which is exactly how I was exposed to them, dozens of times.
I don't understand why many people on here say "humans", instead of people. It sounds like you are talking as if you are a grey alien on a space ship somewhere.
What you are complaining about is abuse of tools/language features. This can happen in any language and/or tools.
That doesn't mean the tool itself is bad.
> Help me understand the actual technical criticism buried in the "worm its way everywhere", please. And in the "it's good for CV/Resume padding" statement as well.
They are self explanatory. I don't like it when someone does this stupid game of not understanding common idioms.
> It's a strange thing to say and it smells like a personal vendetta which is weirdly common on HN about Rust and to this day I have no idea why even though I have asked many people directly.
It isn't. What typically happens is that a tool lets call it "Y" get used everywhere to the point where you cannot use "X" without "Y".
Ruby used to get used for CV padding. I used to work in a Windows/.NET shop and someone wrote a whole service using Ruby on Rails in a Suse Linux. That person left and got a job doing Rails shortly after.
> Rust has objective technical merits and many smart devs have documented those in their blogs -- journeys on rewrites or green-field projects, databases, network tools (like OP), and even others. Big companies do studies and prove less memory safety bugs over the course of months or years of tests.
Just because <large company> does something and says something is true doesn't mean it is or is suitable for everyone or it is actually true.
I have worked at many <large corps> as a contractor and found that the reality presented to the outside world is very different to what is actually happening in the building.
e.g.
I was working at a large org that rewrote significant portions of their code-base in new language instead of simply migrating their existing code-base to a new runtime.
I made plenty of money as a contractor, but it was a waste resources and the org lost money for 3 years as a result.
BTW they never fully transitioned over to the new code-base.
Company blogs and press releases will say it was a success. I know for a fact it wasn't.
> The Linux kernel devs (not unanimously) have agreed that Rust should no longer have experimental status there recently -- and people are starting to write Linux drivers in Rust and they work.
So I will need any additional toolchain to build Linux drivers. This is what is meant by "worming its way in". I have done a LFS build and it takes a long time to get everything built as it is.
> I am honestly not sure what would satisfy the people who seem to hate Rust so passionately. I guess it announcing full disband and a public apology that it ever existed? Yes this is a bit of a sarcastic question but really, I can't seem to find a place on the internet where people peacefully discuss this particular topic.
You are making assumptions that I hate Rust. I don't. I just don't care for it.
What I do hate is hype and this constant cycle of the IT industry deciding that everything has to be rewritten again in <new thing> because it is trendy. I have personally been through it many times now, both as an end user and as a developer making the transition to the <new thing>.
You are kind of ranting on general trends here, not to mention misrepresenting what I said which I cannot see as arguing in good faith (never said that big companies using something makes it good; I said that they did studies -- that's absolutely very much not the same as many other types of cargo culting that they do which I'll agree is never credible).
I am not pretending to not understanding anything as well by the way, I was trying to find objective technical disagreements and still can't find any in your reply. I am seeing a bit of curmudgeon-ing on several places though, so I am bowing out.
> You are kind of ranting on general trends here, not to mention misrepresenting what I said which I cannot see as arguing in good faith (never said that big companies using something makes it good; I said that they did studies -- that's absolutely very much not the same as many other types of cargo culting that they do which I'll agree is never credible).
I didn't misrepresent anything you said. You misunderstood what I said. I said companies will claim all sorts of things and the reality behind the scenes is very different. Having a study is one form of making claims. What works at one company may not work at another.
> I am not pretending to not understanding anything as well by the way,
Yes you were. What I said was plainly obvious.
CV driven development is a well known and understood phenomenon.
Technology stacks being subverted is a well known and understood phenomenon.
It is very annoying when people pretend not to understand basic idioms. It is a dishonest tactic employed by people online, I've been online now since the late 90s and have seen it done many times. You are not the first and won't be the last.
> I was trying to find objective technical disagreements and still can't find any in your reply.
I gave you them. Several in fact.
Part of engineering is understanding that resources aren't infinite and you have to make trade offs. So how resources (money, time, man power, compute) is used is technical. I make calls all the time on whether something is worth doing based on the amount of time I have.
This is often discussed on many blogs, podcast and books about software engineering.
What you want to do is narrow discussion down to the sort of discussion "well they found they found X more bugs using Y technique". Ignoring the fact that they may had to spend a huge amount of man power to rebuild everything and created many more bugs in the process.
> I am seeing a bit of curmudgeon-ing on several places though, so I am bowing out.
Yes I am disillusioned with the industry after working in it for 20 years. That doesn't invalidate what I say.