TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Bret Victor: Seeing Spaces [video]

302 点作者 zindlerb将近 11 年前

20 条评论

jayvanguard将近 11 年前
Great presentation as usual. One fundamental tension I see in much of the work he does is between purpose-built and general-purpose tooling and environments.<p>The challenge in both the maker space as well as much of the visual learning and programming material he has done previously is that each of them is incredibly time consuming to adapt to each new different project. In the real world even similar tasks within projects in the same domain often have enough subtle differences that re-use is not possible or very costly.<p>That isn&#x27;t to say these are insurmountable but maybe much of the focus needs to be on meta-tooling that can accelerate the work of experts to build these purpose-built environments (as opposed to making generic tooling).<p>Inspiring stuff.
评论 #7881305 未加载
评论 #7880815 未加载
tsunamifury将近 11 年前
I think software engineer and the fundamentals of coding have always had a bias towards those who can conceptualize ideas in the abstract, then build with the assumptions that those concepts are happening regardless of their ability to see them.<p>This is fine, except that it limits those who need to tinker in order to find out how those concepts work. When the elements are visually recognizable and physically manipulatable, you can tinker without having to hold the entire chain on concepts in your mind. It reduces the load and increases the likelyhood of &#x27;playing around&#x27;.<p>I hope some day more of Victor&#x27;s ideas can be realized through the understanding that visualizing processes allows us to use more of our brain to design and develop or products -- not to mention stumble upon and explore unexpected outcomes.
评论 #7881668 未加载
评论 #7882926 未加载
评论 #7881387 未加载
primigenus将近 11 年前
It blows my mind that Bret keeps giving talks in public and sharing his ideas for free when pretty much each of them could have been used as a startup pitch in return for likely investment. But I guess he&#x27;s more interested in inspiring others than just committing to one idea for years. I&#x27;m glad we have him around.
评论 #7881062 未加载
评论 #7881086 未加载
评论 #7881061 未加载
greggman将近 11 年前
I love Bret Victor&#x27;s talks, blog posts, etc. They&#x27;re super inspiring.<p>2 things came to mind though.<p>1. It seems, possibly, the exact wrong time to make rooms with giant displays. With things like Google Glass and Oculus Rift as first gen (2nd?) VR&#x2F;AR you could project all of that info virtually and cheaply and be able to have all the visualization he describes wherever you are, not just at a makerspace that only a few people can use at a time.<p>2. I&#x27;m always super inspired by the Bret&#x27;s visualizations but when I actually try to figure out how they&#x27;d be implemented I&#x27;m clearly not smart enough to figure out how that would happen.<p>In this example in particular, he shows graph toward the end where the system tries every setting and graphs the results so it&#x27;s easy to pick out the best setting. How would that happen? How does the system know what &quot;good&quot; is? It seems to me it can&#x27;t know that. You&#x27;d have to program it which in itself would be pretty hard. Worse, most system have not just one adjustment but many. Just a few and there&#x27;d be tens of thousands of variations&#x2F;combinations to try to figure out &quot;best&quot;.<p>I&#x27;m not saying we can&#x27;t get there. Maybe the first step is building a framework that would make it easy to make systems like that with various kinds of visualizers, analysers, time-recording, searching features etc, and maybe somewhere along the way we&#x27;d figure out how to automate more of it.<p>I&#x27;d love to help work on such a system.
评论 #7882608 未加载
评论 #7882007 未加载
评论 #7882354 未加载
评论 #7884006 未加载
评论 #7884253 未加载
评论 #7883858 未加载
评论 #7882904 未加载
评论 #7893186 未加载
mrspeaker将近 11 年前
Every time I realize I&#x27;m guessing about (rather than directly seeing) the behavior of my code I think of Bret&#x27;s talks. I never actually improve my workflow, but at least now I&#x27;m angry about it!
评论 #7880832 未加载
cing将近 11 年前
In other words... a meatspace debugger? Cool idea, but I don&#x27;t quite buy the comparison to &quot;spaces laden with sensors and visualizations&quot; like the NASA control center, Large Hadron Collider, etc.. All of those spaces revolve around monitoring, not the design&#x2F;making process. In a similar vein, in my field of computational science, heaps of money has been invested in spaces for data exploration&#x2F;visualization [1], unfortunately, they are essentially useless for the scientific process.<p>[1]: <a href="http://en.wikipedia.org/wiki/Cave_automatic_virtual_environment" rel="nofollow">http:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Cave_automatic_virtual_environm...</a>
评论 #7883741 未加载
评论 #7887639 未加载
seanmcdirmid将近 11 年前
Some of this is extremely similar to Jun Kato&#x27;s research.<p><a href="http://junkato.jp" rel="nofollow">http:&#x2F;&#x2F;junkato.jp</a><p>More specifically see phybots:<p><a href="http://junkato.jp/phybots/" rel="nofollow">http:&#x2F;&#x2F;junkato.jp&#x2F;phybots&#x2F;</a><p>Kato leverages the overhead camera trick in this system, though in a bit different way. See &quot;A Toolkit for Easy Development of Mobile Robot Applications with Visual Markers and a Ceiling Camera:&quot;<p><a href="http://junkato.jp/publications/uist2009-sakamoto-andy.pdf" rel="nofollow">http:&#x2F;&#x2F;junkato.jp&#x2F;publications&#x2F;uist2009-sakamoto-andy.pdf</a>
评论 #7882850 未加载
pseud0r将近 11 年前
These kind of things would be really great for science labs also.<p>Let say you&#x27;re doing some medical research, growing some cell cultures and you add some compounds to the cell cultures to see what happens. Then something weird happened to some of the cell cultures, and you don&#x27;t know exactly what caused it. Perhaps that thing was really an important scientific discovery waiting to happen, but you missed it, because you didn&#x27;t have all the data.<p>The process is normally recorded with a lab diary, where you write down everything deemed important. The problem is, you&#x27;re not going to notice everything, and there is also a lot of things that you can&#x27;t see without more sensors that just your eyes.<p>The system Bret describes here is basically an automated lab diary. With enough sensors it could record much more data, much more accurately than a person, and it has a way to query the actual data rather than having to either manually browse through pages of text or searching through it with just a basic full-text search engine.<p>A problem with many scientific experiments is that you might a lot of measuring equipment and sensors for the thing you are doing an experiment on, but you don&#x27;t have the same thing for the experiment itself, to easily be able to debug the process and to see where something went right or wrong. Why was one lab able to reproduce an experiment, but another couldn&#x27;t? This kind of questions can be very difficult and time consuming to answer.
vanderZwan将近 11 年前
BTW, for good reading material on control rooms, look on Google Scholar for papers by Paul Heath and Christian Luff. They&#x27;re very thorough in their analysis of how people in control rooms communicate and &quot;spontaneously&quot; synchronise their actions.
cma将近 11 年前
I think spaces like this would do well to incorporate projected augmented reality ala CastAR: <a href="https://www.youtube.com/watch?v=GpmKq_qg3Tk" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=GpmKq_qg3Tk</a><p>You could collaborate, sharing the same view, or each individual could project different views, or mix and match.
Htsthbjig将近 11 年前
I agree with Bret Victor here. I already have something similar of what he is proposing. Not so great, but my prototype is real and works. You can make one of this using &quot;inexpensive&quot; TVs for most of the room. Cheap cameras with HDMI and framegrabbers, a PC with CUDA-OpenCL cards. Arduino sensors work anywhere with all OSes and super easy to use, albeit not very efficient.<p>My experience with years of embedded programming is that NO HUMAN BEING is made for working with the cold, brainless machine or metal if you don&#x27;t visualize your data.<p>Even the person who tells you she likes doing it, she can&#x27;t work on it for long periods of time without burning.<p>It is like climbing over 7.000meters of altitude. Humans could survive for some time with those conditions, but depleting internal resources fast.
hartror将近 11 年前
Some of the software shown in the first minute: <a href="http://vimeo.com/66085662" rel="nofollow">http:&#x2F;&#x2F;vimeo.com&#x2F;66085662</a><p>Some pretty tools in there.
vidar将近 11 年前
He just keeps knocking stuff out of the park.
评论 #7881151 未加载
Sarien将近 11 年前
I am totally in favor of good tools with good visual representations but those almost always have to be handcrafted for every specific problem. Which is probably why Bret has never delivered anything useful.<p>And if you&#x27;re going to talk about ideas and inspiration: Lighttable does nothing that emacs didn&#x27;t do 20 years ago except a little prettier.
kentpalmer将近 11 年前
A theory that I have been developing that might be a basis for understanding the possibility of Seeing Spaces is called Schemas Theory.<p>See <a href="http://SchemaTheory.net" rel="nofollow">http:&#x2F;&#x2F;SchemaTheory.net</a> for a draft presentation that is still in work. Audios are still in production for the tutorial.<p>Other papers on Schemas Theory are at <a href="https://independent.academia.edu/KentPalmer" rel="nofollow">https:&#x2F;&#x2F;independent.academia.edu&#x2F;KentPalmer</a> and <a href="http://emergentdesign.net" rel="nofollow">http:&#x2F;&#x2F;emergentdesign.net</a> and <a href="http://archonic.net" rel="nofollow">http:&#x2F;&#x2F;archonic.net</a><p>A good book on Schemas is Umberto Eco Kant and the Platypus.<p>Basically schemas theory tells us what it is possible to see and also give us the intelligible templates for our designs.<p><a href="http://kdp.me" rel="nofollow">http:&#x2F;&#x2F;kdp.me</a>
kgc将近 11 年前
I think Oculus technology would allow people to do all of this virtually with the physical portion being merely props. This would be a lot cheaper than doing everything for real.
ohwp将近 11 年前
I think this is how the NSA became the NSA as we know it today. When your task is to prevent terrorism you need to see. You need to see in time and detect patterns. So you need to store as much data as possible because.<p>So it&#x27;s good to stick to some boundaries. In the example of the robot: you could measure room temperature, because maybe the sensors are acting to it. Or you could measure the amount of people in the room because sensors could act to it. Heck, maybe the sensors are acting different to different people, so track there faces and store it. Well maybe sensors are sensitive to somebodies smell so track that too.<p>There are limits to what is useful to track.
nchlswu将近 11 年前
After a while of reading the replies, Iron Man came to mind.
nickbauman将近 11 年前
Bret Victor is the Leonardo DaVinci of the age. A curator, assembler and presenter of the great ideas of our time.
agumonkey将近 11 年前
Augmented Breality