Hacker News Reader

Comment by gitemacs | Hacker News Reader

I disagree with Aaron. I would reserve the title of the greatest programmer in the world to Fabrice Bellard. He single-handedly wrote QEMU, FFMPEG, an LTE base station, a PC emulator in javascript, and countless other projects. Alone.

http://bellard.org/

View on HN

djb aims to write bug-free programs and actively works on improving the process to do so. While Bellard is a great programmer, I don't think Bellard aims to write bug-free programs.

See section 8 of http://cr.yp.to/cv/activities-20050107.pdf

There are more people who aims to write bug-free programs, but most of them seem to use formal verification. djb seems pretty unique in aiming bug-free programs in normal programming and getting close.

Re: Bellard software quality. Fuzzing found 1120 bugs in FFmpeg. While Bellard's code is only small part of entire FFmpeg, FFmpeg is very far from being bug-free.

http://googleonlinesecurity.blogspot.com/2014/01/ffmpeg-and-...

I was shocked (and impressed) when I looked at some of djb's code and saw that he put his "the return value of every syscall must be checked" view into practice so consistently that every single printf() (not only fprintf()!) was inside an if block that checked whether it succeeded in writing to stdout or not.

Of course, printf() does have a return value ("Upon successful return, these functions return the number of characters printed"), and it can fail... but it takes a considerable consistency of purpose to decide to deal with that possibility explicitly every single time.

out of curiosity, what do you (does he) do if printf fails?

From qmail/qmail-pw2u.c (linked somewhere above):

    void die_write()
    {
      substdio_putsflush(subfderr,"qmail-pw2u: fatal: unable to write output\n");
      _exit(111);
    }

    /* ... */
    if (substdio_puts(subfdout,uugh) == -1) die_write();
    if (substdio_puts(subfdout,dashcolon) == -1) die_write();
    if (substdio_put(subfdout,x,i) == -1) die_write();
    /* ... */

So what does he do when substudio_putsflush has an error?

Well, the process will still exit unconditionally immediately afterwards, attempting to signal an error via its return code. It's hard to do better than that in the context of a real operating system.

I think this says as much about djb's abilities as about C's limitations.

> See section 8 of http://cr.yp.to/cv/activities-20050107.pdf

Last sentence in the paper:

"I won't be satisfied until I've put the entire security industry out of work."

That's nothing short of amazing.

Another one from Section 8

> Bug-elimination research, like other user-interface research, is highly nonmathematical. The goal is to have users, in this case programmers, make as few mistakes as possible in achieving their desired effects. We don’t have any way to model this—to model human psychology—except by experiment. We can’t even recognize mistakes without a human’s help. (If you can write a program to recognize a class of mistakes, great—we’ll incorporate your program into the user interface, eliminating those mistakes—but we still won’t be able to recognize the remaining mistakes.) I’ve seen many mathematicians bothered by this lack of formalization; they ask nonsensical questions like “How can you prove that you don’t have any bugs?” So I sneak out of the department, take off my mathematician’s hat, and continue making progress towards the goal.

It's really quite different. Bellard is inspiration, he creates paths to new and interesting things. DJB's avowed goal is to put the security industry out of work by not having bugs through rigorous practices.

Bellard is what you use on your laptop to have fun, DJB is what you run in your bank to make the world go round.

While DJB may be more "bank" oriented, plenty of people have been making the world go round using Qemu and FFMPeg. I'd argue that Youtube (if I recall correctly it was and possibly is one of the major users of FFMPeg) has had just a big of an impact as online banking.

Well, FFMPeg makes the world go round, including for hackers: http://thenextweb.com/google/2014/01/10/google-says-helped-f...

I admire Fabrice Bellard. I find amazing the amount of software that he has produced. Some of his creations are remarkable being one-man efforts eg. a compiler, an emacs clone, an emulator, etc.

I spent quite some months learning about editors and one of my evening distractions was to write a new UI backend for qemacs. I found the design clever, quite flexible, etc, but in no-way easy to read or elegant. (To be fair, it did not help that he added a video player and a WYSIWYG html editor to it).

Later I came to a Vim clone called vis (https://github.com/martanne/vis) and I found the source code more close to those programs where you think "It can't be made simpler".

I wish future programmers read more code before writing just like literature writers do. It should be part of the education system and I regret not having read "classics" before starting writing.

I always had the feeling the djb's software is going to be interesting but never going to be widely adopted long term. I think this feeling turned out to be right, since I rarely come across environments running anything from djb, especially lately, and it's hard to spend a day without touching a few things involving code from Bellard.

I've spent quite a bit of time trying to do things with tcc, and - ugh, no.

tcc did start life as an Obfuscated C Contest entrant. It shows.

(It won, of course. http://bellard.org/otcc/)

Not single handed, look at how many contributors most of those projects have...

Fabrice Bellard is smart, though his code is hard to read imo.

Reply on news.ycombinator.com