3 ms·
I'm not defending C. If anything, I'm attacking all of the so called "better" replacements. Specifically your call out that they should have existed by now.
by taeric 10y ago
I'm not defending C. If anything, I'm attacking all of the so called "better" replacements. Specifically your call out that they should have existed by now. I treat that argument as a fallacy nowdays. I'll call it the "wishful thinking fallacy."
I do this because if there is anything that calls into question the "science" part of computer science, it is the dogma of folks that refuse the empirical side of the field. It isn't like folks haven't tried to make better languages and toolchains throughout the years. You even acknowledge this. They just haven't delivered on their promises.
And, some of that is because you are ignoring all of the advances that have happened in C. If you are not using modern tools like valgrind/coverity/whatever, you are not comparing C fairly to contemporary languages.
- jerf 10y ago"If you are not using modern tools like valgrind/coverity/whatever, you are not comparing C fairly to contemporary languages." I personally consider C + strong static analysis tools to be an entirely different language. I fully acknowledge this is my personal opinion. Consequently, I consider it equivocation in this sort of context to be talking about C when convenient, then talk about C + strong static analysis tools when convenient. You're only programming in one or the other. My utter contempt is for C the language by itself. This is relevant because as near as I can tell, C has at least a 10:1 advantage in the field over C + strong analysis tools, and I'm probably off by at least a factor of magnitude. It is not realistic to pretend that C programmers are using these tools at scale. (Now, you can point at a ton of high profile C programs that do use these tools, because high profile C programs are exactly where they get used. But as near as I can tell they are firmly the exceptions, not the rule. And you can point at a ton of high-profile projects that don't use them, too.) If everybody started using all these tools, including the ones that you actually have to buy because there's no open source equivalent of, it is true that most of my issues with C would be resolved. It would also be true that programmers would be a lot less gung-ho about C, though, because as anyone who has deeply integrated these strong tools into their workflow can attest, C + strong analysis tools is a much more complicated language. Now, in my opinion, that complication is being revealed by the analysis tools, not created, and all C code is actually that complicated and the programmers are merely kept in the dark by C's extensive structural deficiencies that make it so it's only safe to write in with these rather sophisticated tools, but the end result is still you're working in a much more complicated, demanding language with a much more baroque compilation process. Claims about C's productivity, compiler speed, probably execution speed, ease of programming, probably quite a few other things, all go flying out the window if this is the solution, and C looks eminently more replaceable.
- taeric 10y agoWe have not escaped the "wishful thinking" yet. Do I want there to be a replacement? Yes. I do my coding in other languages. Have I grown weary of the attacks on C? Also yes. Are not enough people using advanced toolchains? Yeah. I agree with that. But, this is effectively the same complaint that not enough people are using higher level languages. With the exception that fitting some of the extra tools on there is easier than switching languages.
- pjmlp 10y agoLint was created for C in 1979. In Dennis own words: "Although the first edition of K&R described most of the rules that brought C's type structure to its present form, many programs written in the older, more relaxed style persisted, and so did compilers that tolerated it. To encourage people to pay more attention to the official language rules, to detect legal but suspicious constructions, and to help find interface mismatches undetectable with simple mechanisms for separate compilation, Steve Johnson adapted his pcc compiler to produce lint [Johnson 79b], which scanned a set of files and remarked on dubious constructions." Taken from https://www.bell-labs.com/usr/dmr/www/chist.html https://www.bell-labs.com/usr/dmr/www/chist.html We are in 2016, and static analysis is still ignored by the majority of C programmers that "just get work done".
- taeric 10y agoApologies if this is rude, but I'm not sure what this is getting at. I know that static analysis is old. I also know that I have been guilty of ignoring it sometimes. Especially as I was learning. I think it is still in the wishful thinking world to think "better tools" should have won by now.
- pjmlp 10y agoThe thing is that the C working programming class doesn't value quality and secure code, performance FTW everything else gets done later, if there is time that is. Even the language authors were honest and: 1 - created a tool in 1979 that most C developers ignore even in 2016, 37 years later; 2 - realized that C was due its date and took part in Limbo and Go design Yet lots of C coders continue to improve this NIST list, every single day: https://nvd.nist.gov/visualizations/cwe-over-time https://nvd.nist.gov/visualizations/cwe-over-time For me wishful thinking, is the mentality that there are C developers out there that manage to write memory corruption free without using such tools. Never met any since my first contact with C in 1992.
- kstrauser 10y agoFWIW, I worked at Coverity. I didn't write the analyzers, but I sat next to the team that did. It is an amazing piece of software and that team would regularly present new features to us that I would have sworn were provably impossible, like "you've just solved the halting problem" kind of things. I'm sure any of those engineers could glance at an IOCCC entry and say "oh, there's a missing semicolon in that part that looks like a cat's head". Seriously, it's black magic written by wizards and is the most impressive codebase I've ever seen. But. Holy crap, the idea that you have to use something like Coverity just to make C code not suck is a bitter pill. I mean, I'm glad it exists but it's terrifying that it has to and that it's still being developed. After decades of development, the Coverity team is still finding enough examples of new ways to write bad code that they can still improve it. I don't write C because I'm on a first-name basis with the only people on the planet I trust to get it right, and I am not among them.
- aseipp 10y ago> It isn't like folks haven't tried to make better languages and toolchains throughout the years. You even acknowledge this. They just haven't delivered on their promises. The question also has to consider, why did they fail? And what is the criteria for failure? That may be entirely the point: the languages may be better by some metrics, worse in others, but they cannot easily beat the inertia factor, as a result of being so entrenched for 4+ decades. Everyone is very aware of the inertia factor; it seems to permeate all levels of technical decisions in all kinds of fields, for good and bad reasons. If the failure criterion is "not used as widely as C", then, like, everything has failed, so it's a pretty bad criterion, I think. Ada has been around for a while now, and has seen a fair amount of high-assurance industrial use. It's really unclear if you can actually characterize Ada as a total failure for example, it lives on, although it's certainly less popular. But it was designed with the field in mind from day 1. I think a lot of this has to do with things like institutional knowledge and inertia, and vast amounts of time and money sunk into tooling for systems like C. That's not bad, the money and stuff obviously helped; but it makes the narrative that all these other things are strictly worse/never attempted a little bit muddier. Does the "empirical" lack of better replacements imply that better replacements cannot exist, or never existed? I think Ada alone specifically negates the core of that argument, they've certainly been tried and seen some level of success. Or perhaps we didn't actually try enough due to resources, maybe? Or maybe it suggests that other contributing factors, externalities, gave rise to such a "monopoly" of computing intellect? Also, it's easy to forget, but today, it seems general programming languages have enough demanding requirements in general (from things like expectations of package managers, to libraries, to editors), that actually supporting and promoting a general purpose language, to the point of wide usage, is ridiculously time and money consuming. Programming languages are a commodity, for the most part. Existing languages almost all either struggle to exist, have their primary developers financed by some corporations, or have existed for long enough to not simply drift into the void and become effectively immortal. Yet, there's almost no money in it as a field. It's insanely hard to make a business out of things like compiler tech or programming languages these days, outside of specialized fields and clients. Unless you have deep pockets and a willingness to throw money into the void for a while, bootstrapping a "competetive" programming language is going to be insanely hard, on any kind of reasonable timeframe. Rust is a good example: Mozilla financed a vast amount of the development for a long time, and sustained it even in the beginning stages when it was much riskier as an investment. That's likely one of the major reasons it's even remotely as popular as it is today, because they ate the cost, to catch up with modern expectations. All of these compounding factors are not going away. I agree there are few viable alternatives; but I don't agree it's so easy to explain it all away "empirically" with such a simple analysis as yours, that it was all either never tried or just "failed" (without defining what that even means). > And, some of that is because you are ignoring all of the advances that have happened in C. If you are not using modern tools like valgrind/coverity/whatever, you are not comparing C fairly to contemporary languages. The fact you suggest a 15+ year old, insanely complicated static analysis tool that's proprietary to ensure your code should basically be allowed to exist on the modern internet (because otherwise it will just be a nightmare hellhole security problem for someone down the road), yet continues to fail to find true high profile vulnerabilities that keep occurring, and still requires constant evolution -- all so that C can be "fairly" judged with modern languages... it's a bit alarming when you think about it. Well, I'm being slightly tongue-in-cheek. I mean, every time I write C, I heavily indulge in these tools as my insurance to ensure I'm doing things correctly. I even tend to use CompCert to find undefined behavior! Coverity is really good, no doubt. But when I do this, it definitely gives me pause at the investment I have to make. I don't sit around thinking "Yes, this is basically a great solution to the problem". It would almost be, like, I dunno... calling the story of the Titanic a success? Lifeboats hardly seem like the ultimate risk insurance when you're the one crashing into ice, especially when it seems we never have enough of them.