4 ms·

He played this game with precompiled headers a decade ago and there's STILL barely any support for it in the GNU tools. When a debate around implementing preco

by tacos 12y ago

He played this game with precompiled headers a decade ago and there's STILL barely any support for it in the GNU tools. When a debate around implementing precompiled headers ten years after everyone else gets them requires lawyers to get involved, maybe freedom isn't ringing as clearly as we were promised. There's something fundamentally weird about the whole dev ecosystem wrt developer productivity. Next time you're sitting in front of an 8 CPU box with an SSD trying to solve a CPAN or library dependency, watching one C file compile once every three seconds, perhaps think about why things always wind up like this. I super respect Stallman's position but at some point you have to look around (and beyond!) and say "why are we working with crippled tools?" With few exceptions all the core projects are uncontributable at this point. And RMS isn't even the worst; he's just the most consistent and predictable.

ploxiln 12y ago
The tools really aren't that bad. There are many common open source c projects, most notably the linux kernel, which can easily use all your cores and compile tens of objects a second. Compiling crazy c++ codebases are painful for everyone, and in fact were a primary inspiration for the handful of people at Google who created Go, with a primary goal of fast compiles, at some reasonable expense to disk and run-time efficiency.
falcolas 12y ago
Because RMS' stance has nothing to do with your productivity, and has everything to do with giving you the right to look at every single detail of GCC and make it work better for yourself. If he did open it up for a proprietary component to make its way into GCC (while avoiding the licensing terms), and it becomes mainstream, you have lost that right to view and modify the software that runs on your box, and that is something he does not view as acceptable. Ultimately, it's his decision; it's based on his work, and the contributors agreed to his terms. That he makes it freely available with the condition that it remains freely available is a bonus, not a right.
- kps 12y ago
  > the right to look at every single detail of GCC and make it work better for yourself. And yet, the referenced discussion breaks down into RMS on one side and people who want to make gcc work better for themselves on the other.
  - falcolas 12y ago
    > And yet, the referenced discussion breaks down into RMS on one side and people who want to make gcc work better for themselves on the other. At the risk of someone building a closed backend off an artifact of the GCC frontend; at the risk of someone profiting off RMS' work without following his conditions for making that work free. It's important to remember that he isn't against Emacs getting the information it needs to do its work. In fact, this mirrors a common discussion in computer security, and system administration in general: you give applications and users only the permission they need, to protect against bad actors. If Emacs doesn't need the full AST, why would you give it to them?
    - tacos 12y ago
      You're certainly free to keep typing this stuff however the irony is that we all understand RMS clear, consistent position on the matter. The issue is that, by his own admission, he doesn't understand the other side.
      - falcolas 12y ago
        > we all understand RMS clear, consistent position on the matter. Then why are "we" all still using his software (which was given a license to match his philosophy) and complaining when he maintains his "clear, consistant position on the matter" and refuses to compromise his beliefs. To use a poor analogy, it's like going to church and complaining when they talk about God. There are lots of other options out there for editors, compilers, and OSes. If "we" really do understand his position, and disagree with it to the point that "we" want to make changes incompatible with his position, "our" time and energy would be better served working with an OS and software which better matches "our" ideology (like FreeBSD; awesome software under the BSD license). > The issue is that, by his own admission, he doesn't understand the other side. That seems like his problem, not ours. And he's not making it our problem; we're making it our problem by trying to use FSF software in a way other than what they intended. EDIT: Scare quotes around We and Our are used because I, personally, agree with RMS. I didn't used to, but over the years I have grown to appreciate the GPL, the FSF, and the openness these extreme ideologies have brought about for us. RMS popularity among newer programmers is always low (for the exact reasons outlined here, his software isn't "free" enough), and yet I find that after years of experience, we begin to appreciate that without his staunch support, we would be developing software in a very different world. A world which we have a taste of now, in the Apple, Microsoft, Google and Facebook walled gardens: I don't know how Swift works. I can't fix bugs in Swift's implementation, and I can't build upon Swift for the betterment of all, should I so desire.

by tacos 12y ago

ploxiln 12y ago

The tools really aren't that bad. There are many common open source c projects, most notably the linux kernel, which can easily use all your cores and compile tens of objects a second. Compiling crazy c++ codebases are painful for everyone, and in fact were a primary inspiration for the handful of people at Google who created Go, with a primary goal of fast compiles, at some reasonable expense to disk and run-time efficiency.

falcolas 12y ago

Because RMS' stance has nothing to do with your productivity, and has everything to do with giving you the right to look at every single detail of GCC and make it work better for yourself. If he did open it up for a proprietary component to make its way into GCC (while avoiding the licensing terms), and it becomes mainstream, you have lost that right to view and modify the software that runs on your box, and that is something he does not view as acceptable. Ultimately, it's his decision; it's based on his work, and the contributors agreed to his terms. That he makes it freely available with the condition that it remains freely available is a bonus, not a right.

kps 12y ago
> the right to look at every single detail of GCC and make it work better for yourself. And yet, the referenced discussion breaks down into RMS on one side and people who want to make gcc work better for themselves on the other.
- falcolas 12y ago
  > And yet, the referenced discussion breaks down into RMS on one side and people who want to make gcc work better for themselves on the other. At the risk of someone building a closed backend off an artifact of the GCC frontend; at the risk of someone profiting off RMS' work without following his conditions for making that work free. It's important to remember that he isn't against Emacs getting the information it needs to do its work. In fact, this mirrors a common discussion in computer security, and system administration in general: you give applications and users only the permission they need, to protect against bad actors. If Emacs doesn't need the full AST, why would you give it to them?
  - tacos 12y ago
    You're certainly free to keep typing this stuff however the irony is that we all understand RMS clear, consistent position on the matter. The issue is that, by his own admission, he doesn't understand the other side.
    - falcolas 12y ago
      > we all understand RMS clear, consistent position on the matter. Then why are "we" all still using his software (which was given a license to match his philosophy) and complaining when he maintains his "clear, consistant position on the matter" and refuses to compromise his beliefs. To use a poor analogy, it's like going to church and complaining when they talk about God. There are lots of other options out there for editors, compilers, and OSes. If "we" really do understand his position, and disagree with it to the point that "we" want to make changes incompatible with his position, "our" time and energy would be better served working with an OS and software which better matches "our" ideology (like FreeBSD; awesome software under the BSD license). > The issue is that, by his own admission, he doesn't understand the other side. That seems like his problem, not ours. And he's not making it our problem; we're making it our problem by trying to use FSF software in a way other than what they intended. EDIT: Scare quotes around We and Our are used because I, personally, agree with RMS. I didn't used to, but over the years I have grown to appreciate the GPL, the FSF, and the openness these extreme ideologies have brought about for us. RMS popularity among newer programmers is always low (for the exact reasons outlined here, his software isn't "free" enough), and yet I find that after years of experience, we begin to appreciate that without his staunch support, we would be developing software in a very different world. A world which we have a taste of now, in the Apple, Microsoft, Google and Facebook walled gardens: I don't know how Swift works. I can't fix bugs in Swift's implementation, and I can't build upon Swift for the betterment of all, should I so desire.