TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Wolfram Language

134 pointsby oneiricabout 10 years ago

17 comments

one-more-minuteabout 10 years ago
tl;dr: The Wolfram Language is an <i>ok</i> 30-year-old language with a huge standard library. You can do really amazing things in a single line of code... as long as there&#x27;s a built-in function to do that thing. Hmm.<p>I actually really like Mathematica as a tool for symbolic&#x2F;mathematical stuff – they&#x27;re the best at what they do best. But all this Wolfram Language stuff is just way overblown, and realistically a proprietary standard library – however huge – just can&#x27;t compete with the open source ecosystems of Python and friends.
评论 #9555282 未加载
评论 #9553766 未加载
评论 #9553704 未加载
评论 #9555219 未加载
评论 #9554488 未加载
评论 #9555751 未加载
sukilotabout 10 years ago
Whenever Wolfram writes&#x2F;speaks, it&#x27;s really hard to wade through Wolfram&#x27;s overlong rambling about how awesome he is, to find the interesting content.
评论 #9554360 未加载
评论 #9553646 未加载
评论 #9554142 未加载
评论 #9554702 未加载
评论 #9554132 未加载
GuiAabout 10 years ago
The Wolfram Language is quite cool. Sure, it&#x27;s a shame it&#x27;s proprietary - but it has a lot to offer, a few neat ideas, and makes for great, visual demos such as the ones featured in that talk. The fact that you can get cool things running in a few lines of code, like the 3D stack of edges, is appealing - I don&#x27;t think there are many languages out there that allow you to do things like this in a few lines of code, and with such flexibility.<p>(it is quite a shame that Wolfram always has to assert how great they are, how revolutionary it is, how they invented everything, how unlike anything else ever done before their work is. It&#x27;s off putting. Let your work speak for itself, and leave the meta comments out. But I digress.)<p>I like the &quot;tweet a program&quot; concept. One issue with kids growing up mostly on mobile phones is that it makes them less likely to try to see what&#x27;s inside and program it themselves. One solution some have attempted is to have IDEs directly running on tablets&#x2F;phones, but those tend to be fairly clunky. Allowing people to experiment with computation in unconventional manners, e.g. over Twitter, strikes me as an interesting path. I had a similar project a couple years ago where you would tweet concise instructions to a bot that ran them into a virtual machine, for which the output was a 32x32 pixel image buffer and would get tweeted back to you. If you&#x27;ve played with forth salon or shadertoy, it&#x27;s the same idea but with a language optimized so as to get cool results in &lt;140 characters. Sadly, like many of my other bots, Twitter shut it down.<p>But yeah, given their aspirations for education etc., some open source components would be nice. It strikes me as a great exploratory&#x2F;prototyping language, although it&#x27;d probably be hard to maintain large codebases of it due to the fuzziness of the instructions and a certain opacity in the backend.<p>All that being said, the tail end of the talk (about immortality and the singularity and the box of trillion souls - basically, when the paragraphs don&#x27;t have code examples anymore) does fall into the standard techno-religious singularity crackpot speculation where things sound vaguely scientific - just enough to give them an air of credibility - but where the terms used are fuzzy and slippery enough to be manipulated into whatever direction the speaker wishes while maintaining an illusion of rigorous reasoning.
评论 #9553662 未加载
评论 #9553786 未加载
评论 #9554812 未加载
jamiiabout 10 years ago
Has anyone here used the Wolfram Language? How discoverable is it? How does debugging work? What about state management? All the demos are small, stateless calculations. What is it look like when you want to build long-lived, stateful systems - say a robot that builds an internal map as it moves.
评论 #9553886 未加载
Keyframeabout 10 years ago
From what was shown it is really cool. Not because of the language itself, but for integration with data (from W|A?). It&#x27;s like a CSI SQL.<p>Also, Wolfram is all nice for how he did this all by himself and how he let us use big tower of ideas and technology he built, also by himself, over the past 40 years.<p>edit:<p>on a related note, how does it compare with IPython&#x2F;Jupyter?
评论 #9553869 未加载
评论 #9554581 未加载
calhoun137about 10 years ago
Mathematica is very powerful, but it&#x27;s also very expensive and based on a proprietary code base. For me, this is unacceptable and the only solution in my judgement is to build our own.<p>I have been working really hard on exactly that, and in the past 3 weeks have made the most amazing progress. I don&#x27;t want to spill the beans just yet, but this is my second attempt at building a better mathematica, and this time around I am using python, the scipy stack, and sage.<p>I will be releasing the project quite soon, as free open source software, along with a website that is 100% free for everyone. I am really hopeful that I will be able to find like minded programmers who are excited about working on this project together.<p>If this sounds interesting to you, please reply in comments below, or else reach out to me @calhoun137 on twitter.
erispoeabout 10 years ago
What&#x27;s the business model of Wolfram? Do they make most of their reveue from Mathematica or do other things like Wolfram Alpha and Wolfram Language bring a significant chunk of it?
arturventuraabout 10 years ago
I&#x27;ve been trying to implement a open source version of the Wolfram language for FenixEdu but is extremely hard. For the most part is possible as long you don&#x27;t use stuff that depends of proprietary data sets. Also, after 30 years working on it, it is quite faster than anything else.<p>I truly would like to see the community come up with a open source version of it, but with such a huge size, the language would be near impossible for one single individual to implement a compatible version.
评论 #9553939 未加载
评论 #9554224 未加载
rambulatorabout 10 years ago
Mathematica is adequate for prototyping or detailed Science&#x2F;Engineering etc analysis provided datasets aren&#x27;t large and the codebase is small.<p>Mathematica is very fast for your initial development, but you run into massive bottlenecks when you jump from your small test cases to real world data. In other words you get something up and running quickly, then spend a massive amount of time manually (there is no proper debug environment) debugging edge cases and trying to optimize it for real world cases. The time savings aren&#x27;t really there for many projects, as you end up spending more time than if you had used something like C++.<p>The last version of Mathematica (v10) was released over a year ago without an update to the Eclipse plugin, making an already pretty archaic development experience impossible.<p>The front end debugger isn&#x27;t a &#x27;fair&#x27; debugger. When it works (which isn&#x27;t often) you are limiting to only running through small code snippets as its impossible to trace anything.<p>There is no profiler.<p>The number of builtin functions is huge but many are effectively deprecated in terms of their performance or existing bugs. Many of the functions like Map (&#x2F;@) aren&#x27;t optimized. The irony is that once you have optimized your code it ends up looking like that from any other language i.e. you have moved a lot of Wolfram&#x27;s high end functions.<p>I totally agree with others comments that code re-use is really difficult. Many of Wolfram&#x27;s comparisons compare code size with code complexity. Its not always to the case, but 20 lines of Wolfram language can often be very difficult to trace and understand.<p>Access to curated data is nice, but in many cases there is no information on the accuracy or tolerances for data and it has been truncated (rounded to the point of irrelevancy) or is missing elements. So like anything you still need to verify it or access it directly from other sources. Things like units cannot be used in functions&#x2F;loops as it slows runtime to a crawl.<p>But horses for courses, the language is exciting for small projects and maybe a jumping off point for people moving into full development of something.<p>One of the biggest issues facing Mathematica is simply that Wolfram Research&#x27;s focus is no longer Science and Engineering but rather the Internet of Things.
ENGNRabout 10 years ago
I&#x27;d like to use it commercially but it seems so expensive. $300 for only 3000 API calls.<p>I&#x27;d be happy to pay but it would need to be a marginal increment over the CPU + data costs. If that was the case they could get some serious volume and momentum.<p>With only 3000 API calls any cool little prototype you made would hit the limit within minutes of hitting the HN front page.
EGregabout 10 years ago
At first I balked at a computer language being so &quot;big&quot; instead of being, say, a functional language that can import yont of functions in a standard library.<p>But then I realized, this bridges the gap between computer languages and human languages, which we all use to communicate. Although &quot;keeping things organized&quot; in a regular computer language might be done more neatly in some &quot;computer oriented&quot; way, when it comes to actually USING IT to EXPRESS YOURSELF you&#x27;d need a professional to write he program for you.<p>Well, there still has to be clear, unambiguous documentation.
ameliusabout 10 years ago
I wonder if a purely functional language would have been more appropriate, given the fact that programs written in such languages are more amenable to automatic rewriting and reasoning.
bigdubsabout 10 years ago
The &quot;Language&quot; seems like a big kit of built in functions with dubious utility, with some awkward syntax thrown in to tie them all together.
评论 #9553645 未加载
Supersaiyan_IVabout 10 years ago
In case you don&#x27;t know, there&#x27;s &quot;Tweet-a-program&quot;:<p><a href="https:&#x2F;&#x2F;twitter.com&#x2F;wolframtap&#x2F;" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;wolframtap&#x2F;</a> some incredible Wolfram examples live here.
Thizabout 10 years ago
A cloud based open source organic language. That&#x27;s it.<p>That&#x27;s the greatest invention in the history of computing, and it is the final step in cracking AI&#x27;s nut.<p>Kudos Wolfram. You finally made it.
swehnerabout 10 years ago
I expected something new in the area of natural language processing.
yarrelabout 10 years ago
It&#x27;s proprietary.<p>Next.
评论 #9554074 未加载
评论 #9554437 未加载