The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Why so little innovation in programming languages, Hemisemipumo?
Poast new message in this thread
Date: June 14th, 2012 5:53 AM Author: Salmon adventurous cruise ship Aside from Objective C for iPhones, nothing. Same languages for windows 95 minesweeper design multithreaded massive backed infrastructures of google today.
http://www.tiobe.com/content/paperinfo/tpci/images/tpci_trends.png |
 |
Date: June 14th, 2012 8:36 AM Author: Grizzly Abode (retard) (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20885625) |
Date: June 15th, 2012 12:51 AM Author: Salmon adventurous cruise ship (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20890951) |
|
Date: June 15th, 2012 1:05 AM Author: Multi-colored state 1) what do you even mean by this? It's kind of a vague and nonsensical question
2) your premise isn't even true if I make some assumptions about what you're saying. There have been enormous advances in programming languages. Ever heard of Java and the JVM? That was a HUGE deal because it allowed developers to write non-platform specific code.
3) innovations don't come at a superficial "code" level, but rather in compilation or in VM efficencies if that is how the code is executed. Redesigning code syntax or using a new language because "it's new" would be like redesigning a hammer or something. C has been used for 30 years because it still works really well in a lot of situations. |
|
Date: June 15th, 2012 1:12 AM Author: Salmon adventurous cruise ship Java is bloated shit for 12 year olds, (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20890994) |
|
Date: June 15th, 2012 1:16 AM Author: Multi-colored state irrelevant and false. Nowadays java program execution speed is pretty similar to C due to compiler and JVM optimizations. Java is also still widely used. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891007) |
|
Date: June 15th, 2012 2:18 AM Author: dashing tank Innovation, I feel, is more often pulled than pushed through. So, I agree that languages or problem-solving methodologies arise in response to components of context and less so for the sake of being "innovative." I don't mean to come off as rude or arrogant but cannot figure out how else to say what I am only crudely able to sense or feel about this subject; and so, I am completely open to being corrected. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891168) |
|
Date: June 15th, 2012 1:12 AM Author: out-of-control erotic garrison travel guidebook techies can argue all they want, but computer architecture is the same as it has ever been. it's still based on serial processing of binary operations by transistors. we need to move beyond this model in order to truly revolutionize computing. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20890995) |
|
Date: June 15th, 2012 1:18 AM Author: Hairraiser location what would be revolutionary about it?
you're practically speaking more limited by human ability to create only a certain amount of code and brainpower at this point |
|
Date: June 15th, 2012 1:22 AM Author: out-of-control erotic garrison travel guidebook the setup of today's computer architecture requires coders to issue specific commands or build a specific framework for each task that their program carries out. this is insanely tedious.
it's possible to imagine a much more "genetic" architecture which would allow for "self-developing" code which can carry out tasks by itself.
for example, think about database programming. you have to define EVERY GODDAMN THING about the nature of the storage and retrieval of your data.
but imagine a system of computer coding and architecture which would allow you to basically say, "computer, i am going to hurl data at you, and it's up to you to sort and store it in the most efficient manner you can find. change things up if need be." |
|
Date: June 15th, 2012 1:27 AM Author: Multi-colored state this already exists to a large extent. That is the entire point of high level languages and APIs and libraries. Abstraction is an extremely important part of engineering at all levels, and I don't think software design tools are an exception.
I can see some situations in very limited contexts in which what you're saying is true. |
|
Date: June 15th, 2012 1:30 AM Author: out-of-control erotic garrison travel guidebook it goes well beyond that. i'm talking about something more like the computer being able to use and even define its own heuristics in order to achieve specific outcomes.
for example, think about malware. there is no way to give the computer a set of general guidelines regarding things that shouldn't be able to happen to it and would thus indicate a malware attack, which the computer would then automatically reject.
there's no way to code a computer so that - even though you never "mention" (via code) something like a buffer overflow attack - the machine would "recognize" at a general level that certain events were improper all on its own, and would not allow them to proceed. |
|
Date: June 15th, 2012 1:37 AM Author: Salmon adventurous cruise ship Skynet? (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891074) |
|
Date: June 15th, 2012 8:43 AM Author: Razzle bbw (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891569) |
|
Date: June 15th, 2012 1:22 AM Author: Multi-colored state I don't think most people in the know would disagree with that at all. Clearly the current paradigm has certain limitations. We've already reached the physical limit of being able to further shrink processors, which is why there's been a lot of movement toward parallelization in past decade- unfortunately, even that's a temporary fix because it requires that software be redesigned to utilize extreme multithreading, something that humans have a hard time with.
Whether the next computing medium will be bio goo or something else who knows. |
|
Date: June 15th, 2012 4:05 AM Author: Grizzly Abode I'm puzzled by your comment about shrinking. Ivy bridge is on the 22nm process node; sandy bridge was on that 32nm tip, etc. intel claims to have achieved a 15nm process. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891305) |
|
Date: June 15th, 2012 4:17 AM Author: Multi-colored state admittedly, I am not an expert on the hardware side of things so I'm not sure how intelligently I can talk about processor implementation.
However, my understanding is that intel has had problems with the physical implementation of chips- namely that certain components, if moved any closer on the die, would fuse together due to intense heat. Not saying that further size reductions are impossible, but most agree that this paradigm of improving processor clock speed in order to improve performance is not sustainable long term. |
|
Date: June 15th, 2012 4:25 AM Author: Grizzly Abode Im no great expert on hardware either brother. I will say goalposts have been shifting for as long as i can remember, and heat/electrons interacting/whatever have long been cited as reasons we'll never move past x, where x is a couple process nodes down from wherever we happen to be.. That said, we will hit a wall with silicon eventually and it cant be too far off now
BUT returning to the point you were responding to originally, isnt the issue less about silicon and transistors and more about something in the neighborhood of the EDIT halting problem? |
|
Date: June 15th, 2012 4:29 AM Author: Grizzly Abode Bumping this to highlight my edit (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891340) |
|
Date: June 15th, 2012 4:46 AM Author: Multi-colored state do you mean the halting problem?
That is certainly a major, major issue in computing which has a profound effect on the way software is designed. I can see how resolving that issue would allow for major optimizations in hardware. Unfortunately, it is proven to be undecidable so it is just a mathematical fact that any machine designer has to work with.
I think the real issue is balancing hardware implementation and ease of software implementation (this goes to my point about why multithreading isn't a sustainable solution). Great pains have been taken to abstract away the user's experience from lower level machine implementation via multiple layers.
When we have a multicore CPU, only so much optimization can occur through automated means (eg, the kernel can automatically schedule new threads). More importantly, these automated performance gains are logarithmic. The only real way to truly take advantage of 16 cores is through innovation on the programmer's part. Humans are just bad at that.
So you can see, a major obstacle is creating a paradigm for a computing device which allows for the limitations of designers. |
|
Date: June 15th, 2012 5:01 AM Author: Grizzly Abode Cr points.
Just realized I had in mind hemi's post in the next subthread up--I was posting in the wrong subthread. Recognizing buffer overflows before they happen etc--my impression is that the halting problem makes that impossible whatever the architecture unless (I guess) code is executed In a vm first or something of that nature |
|
Date: June 15th, 2012 1:32 AM Author: primrose impertinent indirect expression (non cs-degree holders pretending to be smart)
lol |
|
Date: June 15th, 2012 2:01 AM Author: Swollen apoplectic house reading party hemidemi = one of the most beloved and respected XO posters of all time.
you = greasy little chink bitter about columbia who tried to IP trap XO posters after they LOLed at your shitty competitor forum. |
|
Date: June 15th, 2012 2:13 AM Author: Hairraiser location (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20891152) |
|
Date: June 15th, 2012 2:17 PM Author: primrose impertinent indirect expression lmao!! (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893107) |
|
Date: June 15th, 2012 2:33 PM Author: bearded corner internal respiration let's not overstate things. while he's made some good points in this thread, he's easily prone to being out of his element
for example http://www.xoxohth.com/thread.php?thread_id=1963757&mc=60&forum_id=2#20828059 |
|
Date: June 15th, 2012 3:29 PM Author: Swollen apoplectic house reading party do you realize who he is? he has the largest collection of classic 180 threads of any current poster by a wide margin.
you also need to determine when he's just flaming the fuck out of a thread. |
|
Date: June 15th, 2012 3:31 PM Author: bearded corner internal respiration (pumo who is too easily impressed) (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893488) |
|
Date: June 15th, 2012 3:41 PM Author: Swollen apoplectic house reading party i am definitely on his dick here, but since he was universally recognized as a top 3 poster during XO's golden era and you seem to have no clue about it, i'll quite arguing with you. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893547) |
|
Date: June 15th, 2012 3:34 PM Author: Salmon adventurous cruise ship He's flaming cow goes moo. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893501) |
|
Date: June 15th, 2012 2:28 PM Author: 180 aquamarine forum mediation you realize the language is just a collection of build blocks right?
if a language is turing complete, you can create any sort of program with it that you can create with any other turing complete language. so they are equally "powerful" from a mathetical POV (in fact, languages in the same class can be considered "equal" straight up, because you can map one to the other).
"innovations" in language are usually just adding syntactic sugar, and not at all necessary. |
|
Date: June 15th, 2012 2:33 PM Author: Salmon adventurous cruise ship Concrete is much stronger than it was 30 years ago. Bridges are designed better. Why not the same with the building blocks of computer languages? (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893196) |
|
Date: June 15th, 2012 2:38 PM Author: 180 aquamarine forum mediation because the underlying mathematical principles don't change, as they are axiomatic. computer science is a mathematical science, not a physical science.
the very basic building blocks will always be the same, and that is machine code, or 0's and 1's. we have abstracted two layers above that now. the first layer is assembly, and the second layer being the current collection of "higher level programming languages".
it's possible that we will add more layers, see: functional programming as a more advanced paradigm of programming, but that's all we are doing. |
|
Date: June 15th, 2012 2:48 PM Author: Salmon adventurous cruise ship Your argument is garbage. We aren't using mulicore processors effectively at all. Programming language is not a solved problem. (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893261) |
|
Date: June 15th, 2012 3:10 PM Author: 180 aquamarine forum mediation no one is saying it's a solved problem.
and multithreading is an algorithmic problem, not a language problem. what you tell the processor to do is the same, there are just more processors in parallel and implications. |
|
Date: June 15th, 2012 3:18 PM Author: Grizzly Abode :D, in what sense is Java bloated? (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#20893418) |
|
Date: October 24th, 2017 5:33 AM Author: gold antidepressant drug whorehouse (http://www.autoadmit.com/thread.php?thread_id=1970439&forum_id=2#34515444) |