One of the differentiating factors of Wolfram|Alpha: Gone.
Google's doesn't require a 250mb plugin to become interactive (just a webgl capable browser...). And it seems to be about 8 times faster than W|A...
W|A still does a lot of things other sites don't do, but it tends to be unusably slow, and inconsistent (running same query twice in a row will sometimes not give all of the same results).
I figure the end is probably closer than one might expect for W|A. They don't seem to be moving quick enough to stay infront of the curve forever. Also, the site is becoming littered with obtrusive ads, and that's a pretty negative experience. (I also hear Wolfram's company may be struggling.)
This should have been praise for Google's new tool, but it turned into a rant on Wolfram|Alpha... whoops.
Competition is good for both parties. Wolfram Alpha challenged Google to rethink what people search for to some degree, and Google's rigored performance should challenge Wolfram Alpha to become more performance aware and prevent them from taking too many features "premium".
And in the end, it is a win / win for users. I am always happy when the insanely smart people at Google have hard things to implement like this than how to integrate the posts of my friends on Google+ into my search results for a unicode table.
It's only a win-win if it doesn't do too much damage to wolfram alpha, although I guess if Google pushes too hard it will drive Wolfram to try to integrate more and more into Apple's ecosystem.
Wolfram Alpha also seems to bail out when a mathematical expression exceeds a certain complexity, which makes it fairly useless for the only thing I want to use it for - evaluating hairy integrals and other long expressions. Or worse, it just "interprets" it as a part of my original expression and calculates that, as if I just typed out the rest of it because I enjoy typing random symbols. It seems like such a waste to fire up Mathematica every time but I haven't found any comparable online tools.
I think W|A could have been a great tool for this kind of thing but instead it tried to cover every subject instead of doing one thing well.
The nail in the coffin for me on W|A was requiring a login in order to copy plain text results. (Even though the same data is copyable in the DOM, they actually go out of their way to obscure it!)
I don't think people understand how noxious this registration-obession is, especially when you need utilize it to access basic functionality.
I don't really go to W|A anymore unless there's something specifically math or statistics related that I'm looking for. That action killed any kind of casual use for me.
Of course, since this is a Google result, it can just fail and say "3D charts require a web browser and system that support WebGL." Which is super useful to me.
I can't tell if you're being sarcastic about it being super useful. If you are, it seems easier still to download Chrome to do a 3d render then it is to buy WA pro and download an apparently enormous plugin to do the same render
He might mean that this computer doesn't have a card/driver combo capable of running WebGL. This is actually fairly common from test Mozilla has carried out. Successfull WebGL context creation on Windows were bellow 70% (can't find the actual page with the real numbers right now).
But Chrome now supports software rendering through swiftshaders which should help those people.
I for one agree with you, even with the issue of webgl capable hardware is much better than having to deal with huge plugins. (I'm biased, I use WebGL for my current work).
This is actually fairly common from test Mozilla has carried out. Successfull WebGL context creation on Windows were bellow 70% (can't find the actual page with the real numbers right now).
[Disclaimer: I know absolutely nothing about the internals of WebGL, and I don't mean this to come off as a criticism.]
Does anyone know why it's so hard to get WebGL working on a lot of systems?
I'm just sort of surprised; I remember doing a lot of OpenGL projects many years ago (back when I still used Windows), and I never seemed to have any problems getting them running pretty much anywhere, regardless of graphics card or whatever. Even crappy systems with integrated graphics always seemed to do just fine.
Is there something about WebGL that makes it more difficult to support a lot of systems, or is it just that it's relatively new and nobody's gotten to adding support yet?
I simply didn't understand. His post doesn't say anything about WebGL not being supported by his GPU specifically, though obviously that's a problem. I guess I don't know what he wants? Yes, if your browser and system isn't supported... what do you want? Do we expect Google to have to support every user, for every nonessential, nonrevenue-producing feature?
Maybe I'm just missing something, but it seems I was punished for asking.
You basically equated the hassle of getting WebGL working if someone is on a browser without it to the hassle of getting Wolfram Alpha Pro.
What you are missing is that if you don't have Wolfram Alpha Pro, the regular plain old Wolfram Alpha still plots it for you. It gives you a 3D view and a contour map, and throws in a couple series expansion and the derivative. You don't get the fancy interactivity without Pro, but you get to see what the thing is.
At Google, on the other hand, you either get the full 3D interactive experience, or you get "3D charts require a web browser and system that support WebGL".
You were making unfounded assumptions (that enabling WebGL would be as trivial as installing Chrome) and at the same time ignored the fact that you get highly useful results from WA without any plugins or bleeding-edge tech. Having good fallbacks is important.
>You were making unfounded assumptions (that enabling WebGL would be as trivial as installing Chrome)
No, I made the point that installing Chrome is trivial and the message explicitly states that your system simply might not be capable. If that's the case, what is Google supposed to do? Just render an empty grid and make a "whah-wha" noise?
>at the same time ignored the fact that you get highly useful results from WA without any plugins or bleeding-edge tech
No I didn't. At all. I am not saying this is an "answer" or a competitor to WolframAlpha. This is a trivial feature that was thrown in. It's non-essential and I bet you'll never ever see Google pushing a browser plugin on any of their sites save for some existing exceptions (Talk). As you've noted, there are tons of places to go besides Google for a 3d graph.
>Having good fallbacks is important.
When it doesn't require confusing your user, or compromising standards to ask them to install a gargantuan plugin.
edit: If the whole point of this was merely to clean up the error message or simply suppress it on unsupported browsers, then sure, but like I said, I genuinely just didn't understand the sarcasm, my apologies.
>That's why you were corrected. Stop trying to defend this forever and move on.
Oh for god's sake, give it a rest. I thought (and have clarified this twice now) that it was simply a remark about (in)capable browsers. I clarified (also twice) that if it's about the error message on incapable computers, that it could be better written and I apologized for my confusion.
For many people getting that error, simply installing Chrome is a viable solution.
>If the whole point of this was merely to clean up the error message or simply suppress it on unsupported browsers, then sure, but like I said, I genuinely just didn't understand the sarcasm, my apologies.
That's why I clarified and apologized (hint, that's the post you just replied to). Stop trying to attack me forever (for what I even acknowledged at the time was probably a misunderstanding) and move on.
its because in chrome webgl is currently disabled by default. you can turn it on on about:flgs IIRC. or try any other release channel. It worked for me in canary.
I didn't see anyone mention this specific fear, but a lot of people talked about this in the context of competition with Wolfram Alpha. What I fear is that google (on purpose or not) will outshine Wolfram's offering, push W|A out of the game, and then in 3-6 years decide it is not a feature worth keeping up.
Just like code search, viable companies may fail due to Google toying with their entire industry, and then Google may just drop us back to the stone age in terms of that category of product. The fact that this is a fear at all points (in my mind) to some issues of monopoly.
Then again, maybe I'm just too cynical and worried.
I don't fear Google dropping us behind where we are, I fear another IE6.
IE6 back in the day was actually quite good. But because it was really good (among other other more shady things) everything else died off and it stagnated.
What I fear is google keeping us in a local maxima of search and having us stagnate there for a decade.
I'm guessing people took this as a sarcastic remark but I was genuinely thinking you were going to be able to tell me of some new social networks to try and point out their benefits; I try to keep on top of developments in this area. Oh well.
I'm not sure exactly what you mean, but most of the possible meanings I can think of aren't ones that would really stop us from having google be an (unfortunate) local maxima of search for too long.
I am thinking of laws that make it unlawful to publish certain links with an arbitrary property (containing or linking to copyrighted material) or laws that prohibits it to publish the result of statistical analysis (the whole auto-suggest debate). Note also that everytime that Google changes its ranking algorithm, some companies complain because, as they paid for SEO, they feel entitled to a high google rank.
I don't think W|A is going to disappear (I work on W|A). We do a lot more than plot functions of one and two variables -- after all, Apple chose us to power much of Siri's factual question answering.
We're pioneering automatic analysis of new kinds of input, from images to long-form text to raw data (what I work on). We can out-innovate Google, even if they do a really nice job of copying our old ideas.
Of course, sometimes, like with Google Squared, they don't do such a nice job.
Cool! Hey, a tip, and I think I mentioned this in the comment/feedback box on W|A as well:
For some reason, every time I need information that makes me think "hey this would be the typical thing I'd expect W|A to be able to answer, I've seen it answer a similar question before" it hardly ever works ...
And I think a lot of that is due to how most of the demos I've seen are somewhat USA-centered and W|A doesn't have the same data when I query it for a European (or other) country.
How does that work? Because when I use Google or DuckDuckGo and look for that nation's bureau of statistics, fact sheet, even Wikipedia, the numbers are right there. Why is that? I thought W|A is supposed to index the calculatable knowledge for me so I don't have to search for it but instead can just run calculations with it?
Or does W|A just import complete available datasets and does not scrape the web for such data (even when it's in tabular form on an official government or university site)?
In which case, please get some more of those from non-US sources. Loads of governments opened up their data sources:
http://ckan.org/ <-- Germany, Netherlands, UK, France, Canada or are you using this one already? It's CC-BY-3.0 so there's no reason not to.
What monopoly? Google doesnt remotely have a monopoly on scientific visualisation, and a possible monopoly on search doesnt really overlap with that.
The problem small companies have with Google is Google's _size_, so many of their side projects are bigger than somebody else's entire business. This absolutely doesnt have anything to do with their headstart in search. Whatever project somebody else successfully starts, Google can simply outspend them and reach the goal first, but this isnt specific to Google, you have this problem with any other big solvent company. WA would have the same problem if a producer of toilet tissue with more cash reserves than WA suddenly decided to compete with WA.
If you were to take a poll of earthlings asking them whether they've heard of Google and whether they've heard of Wolfram Alpha, Google would resoundingly win. People would also, by and large, be familiar with their search if anything.
If you then asked those people where they would search for an answer to a mathematical problem, or if they wanted to find the meaning of some math equation, it's not hard to believe that they'd google it. Even if they ended up on Wikipedia, they would probably use Google to get there. Before Google launched this feature, people would use the search to (hopefully) end up at Wolfram Alpha. Now, they are likelier (than before) to have their needs met by Google itself.
A monopoly on search is a monopoly on access to knowledge. If Google goes into competition with another provider of access to knowledge, they have a tangible head start. That's where I'm coming from -- I concur that search doesn't overlap with everything, but I think it overlaps at least a bit with this.
Well, this graph viewer is fun, but it's only a tiny sliver of what Wolfram|Alpha can do; if Google wants to "toy with" the industry it will require some more serious toying!
This looks... erm ... neat. A "sophomore OpenGL term project" kind of neat. Wolfram on the other hand does the graphs right [0], which hardly surprising given their 20 years of head start.
On a more general note, this reminds me of old Microsoft's tactics. Google should really stick to the search, but instead they throw together something that mimics competitor's feature. Something that looks more featureful and which is free, but upon closer inspection is effectively a half-ass effort, because it's an entirely different domain that's not their specialty.
My bet is that this is supported by Google as a "here's a neat hack" feature, not as a "this is a vital feature in our ongoing to-the-death competition with Wolfram Alpha".
i keep hearing this. what makes you think they aren't? Google has a huge staff and equally huge revenues, i'm sure that if they thought search quality could be improved by throwing more engineers at it, that's what they would do.
as long as they employ a lot of engineers, why not have some them working on cool things like this? i see this more as a demonstration of webgl and non-flash tech than an attack on W|A
Exactly, just because google is adding a bunch of cool free non obtrusive features doesn't mean that they are stifling innovation in their main field. Google is a large company with countless divisions, it isn't just one engineer who can only work on one project at a time or anything.
So if a user enters the following search query:
Sqrt(xx+yy)+3cos(sqrt(xx+y*y))+5 from -20 to 20
And gets a graph of the equation as the top result, are you claiming that this is a poor search experience? Are you claiming that this is a bad result?
If not, how is this feature not an improvement to search? How is this not "sticking to search"?
Note - I was not involved in any way with this feature, or with search quality in general
Who cares if it's trivial? It's functional, fast, convenient and doesn't require a bunch of cruft. Kinda damages your quasi-rant to find out that you simply... didn't compare the graphs over the same interval. I don't know why every small feature that Google adds is either apparently out of fear of someone who's hardly a competitor or is proof that Google is the next Microsoft (which is a useless statement anyway).
This all just reminds me how truly sad HN's markup is. I can't even escape the multiplication in the query for Google to get it to render properly (I replaced it with a Goo.gl link).
Word of caution: I left this open in a Safari tab without thinking about it, and soon found that my Mac's UI was hanging for ~20 seconds every minute or two. Took me half an hour to figure out what was going on.
Nope it's been on for a while but it will only work if you have good OpenGL drivers. (About ~50% of Linux users who try are able to view WebGL content from mozilla's numbers).
After fiddling with about:flags and using --ignore-gpu-blacklist on the command line, I got chrome://gpu to report that WebGL is fully enabled. It _still_ doesn't work, with this error reported in the console:
[75:75:1827873622829:ERROR:command_buffer_proxy.cc(110)] Could not send GpuCommandBufferMsg_Initialize.
I will say it's strange that Chrome thinks your computer is incapable but Firefox 11 is able to perform the task. It may be a bug with Chrome worth looking into.
Feels like a gimmick... Something MS used to do well - take the shiniest feature of a competitor & implement it. The graph plot is just the tip of the iceberg - W|A does much much more, and Google might drop this in a few months if they don't see traction.
I thought Page spoke to Jobs and got a lesson in 'focus' ?
If Google really comes up with a viable competitor to W|A, with all the datasets, math, etc., that's one thing. Just copying the shiniest toy feature is a little dull...
I spent a good minute reformatting the formula in my head, trying to figure out the significance of the equation, before I went, "oh, duh, it's the fact that it's rendering a formula in 3D within a Google result!"
Thank you, Google, for making me uninstall some nVidia packages from Ubuntu that were apparently stopping WebGL from working. The tool you built is also kinda nifty, and I might even use it when trying to help people online with mathematics.
Nonetheless, my day-to-day search traffic is still being offloaded to DuckDuckGo and my day-to-day image traffic is still going to Blekko. It's not you, it's me.
Blekko's good for images? I gave it a few tries a while back but I got distracted by the slash syntaxes and forgot what I was looking for :) I'll give it another try.
Today I mainly use Yahoo's image search for images.
Disabling JS on Google Search (with Opera's site-specific preference or NoScript) also does a lot of good to make this once-great search engine somewhat more usable (not for the results, of course, but at least it's not continuously stealing input focus and mucking up its own search results as you try to navigate with your browser's keyboard shortcuts).
As you can see, I call these "jumps" and I have a bunch of them. 'o drostie' takes me to http://www.drostie.org/ for example. 'wik Aharanov-Bohm Effect' takes me to the SSL-enabled Wikipedia pages (which used to be in a very ugly place, secure.wikimedia.org, but have now been moved to en.wikipedia.org like you'd expect). There's "def" to define a word, "ety" to look up its etymology, "ark" to look something up in the Wayback Machine, and of course, "s" to search. I like the idea of verbs in my URL bar.
This is impressive. Maybe the programmers who did this can speak to the financial charting side so their charts will plot a moving average using data that is actually off the left side of the chart, instead of omitting those points as if the data does not exist.
This is very nice; one addition that would make it even better is support for implicitly defined functions and relationships. If we could graph things like x^2+(y-x^(2/3))^2=1 directly, that'd be awesome.
Would be nice if it had a color bar (legend) to see which color is which value, though. The red/green/blue used below the graph for the axes ranges looks like one, and that is a bit confusing.
Google's doesn't require a 250mb plugin to become interactive (just a webgl capable browser...). And it seems to be about 8 times faster than W|A...
W|A still does a lot of things other sites don't do, but it tends to be unusably slow, and inconsistent (running same query twice in a row will sometimes not give all of the same results).
I figure the end is probably closer than one might expect for W|A. They don't seem to be moving quick enough to stay infront of the curve forever. Also, the site is becoming littered with obtrusive ads, and that's a pretty negative experience. (I also hear Wolfram's company may be struggling.)
This should have been praise for Google's new tool, but it turned into a rant on Wolfram|Alpha... whoops.