3 ms·
Instead of replying in a manner I haven't heard since middle school, you could be writing a much-delayed explanation of your claim that a stack-frames-and-addre
by mannykannot 12y ago
Instead of replying in a manner I haven't heard since middle school, you could be writing a much-delayed explanation of your claim that a stack-frames-and-addresses explanation of pass-by-value would necessarily be "exponentially more complicated" than your C++-based approach.
- jjnoakes 12y agoI guess I am the middle schooler here because I'm going to point out that "you started it". I won't even dwell on the hypocrisy of you trying to call me out for my rudeness right before you post this middle school nonsense. If you don't think stack frames and addresses are far more complex a learning tool for this than references and pointers, that's your opinion. I disagree, and I think many would, but there's no factual claim here for you to refute, no matter how much you are grasping for one in this thread. Go ahead and explain it your way to folks. See if it finally sinks in for those who didn't get it from the JLS or from the article. I'd be happy to watch the attempt. I don't think it'll work. Prove me wrong.
- mannykannot 12y agoIf it were just a matter of opinion, you could have left it like that a long time ago, but instead you have presented a long series of factual claims intended to show that yours is the only reasonable opinion, and I am merely responding to them. For example, your claim of exponential complexity was not just a matter of hyperbole, you backed it up with an explanation: "...because you need to either stick to one specific implementation of one specific language runtime, or you have to deal with explaining them all." If that is no longer your position, say so, and we can move on. Another recurring factual claim is that there is a set of people that will understand this issue only if it is presented in a very specific way. Putting aside, for the moment, the question of how well that worked out in practice, let's look at what characteristics you have ascribed to your target: 1) They didn't get it as presented in the original article; 2) They have read what the JLS says about it, but did not understand it there; 3) They could not be led, through a suitable explanation, to follow what the JLS says; 4) Any explanation based on computing basics would be beyond their grasp; 5) Despite all this, they understand (and not just think they understand) the more complex issue of argument-passing in C++. It seems unlikely there is anyone in this set. In particular, while there seem to be people in group 3, I doubt any of them have made it into group 5. As for an alternative explanation, we have, for example, eslaught's post: https://news.ycombinator.com/item?id=9249285 https://news.ycombinator.com/item?id=9249285 The first paragraph gets to the essential point with admirable directness, without using any comparison to a language that has a particularly complex array of argument-passing cases.
- jjnoakes 12y agoIt's always been a matter of opinion. Most people don't take conversational prose literally and argue against it as if it was a fact that needed refuting. For example, it is your opinion that C++ argument-passing is more complex; it is mine that it is more explicit and easier to reason about. Notice how I don't take your claim and call it a fact and write pages about how you must back up your claim with some more evidence before I'll buy it.
- mannykannot 12y agoI think many people would agree that, since you challenged me to present an example, and I did, then it is quite reasonable for me to ask if it has changed your opinion that any alternative to your approach would be infeasibly more complicated, and if not, why not.
- jjnoakes 12y agoI'm sorry, where did I say that "any alternative to my approach would be infeasibly more complicated"?
- mannykannot 12y agoIt is true that while you have tried repeatedly to imply there is no viable alternative, you have not explicitly ruled out some so-far unmentioned alternative. Nevertheless, you were very explicit when you made this argument against one class of alternatives: "Sure you could do it that way. But it becomes exponentially more complicated because you need to either stick to one specific implementation of one specific language runtime, or you have to deal with explaining them all. Neither is ideal." So it is reasonable for me to ask if the example you asked for changes your mind on this issue.
- jjnoakes 12y agoI asked you to write up an example using stack frames and addresses independent of any language. You have yet to do it. The comment you linked to is much more reasonable than that and only introduces a little more complexity. But it is still more complex in my opinion because it uses both stacks/locals and heaps/objects as well as referencing C++, and I don't think someone who confuses Java for a pass-by-reference language will understand that any better. If your goal in all of this is to try to change my opinion, you will need to keep trying. Why you would spend all this time on it is beyond me, but since I enjoy correcting your fallacies and you seem to enjoy grasping for ways to appear like you've accomplished something with all of your replies, we could be doing this for quite a while. Keep it coming.