Wolfram Alpha won't ask and won't tell
Summary: It's the site's attitude toward Internet inputs, as opposed to Matt Asay's revelations on Internet outputs, which I am afraid marks Wolfram Alpha for failure.
In everything being written about Wolfram Alpha, here and elsewhere, an obvious point is being missed.
Its attitude toward the Internet. (Picture from the Richmond Library childrens' blog, Batavia, NY.)
Stephen Wolfram does not trust the Internet. He doesn't trust its inputs, he doesn't like its outputs. He doesn't like what it has done to "intellectual property." He does not like it, Sam I am.
Notice what I just did there? I alluded to a copyrighted work without attribution. Wolfram wouldn't like that.
Wolfram promises that all Wolfram Alpha results will be "vetted." The freewheeling attitudes of Google or even Wikipedia are forbidden.
It's the site's attitude toward Internet inputs, as opposed to Matt Asay's revelations on Internet outputs, which I am afraid marks Wolfram Alpha for failure.
The Internet, for all its messiness, for all its lies and lying liars, usually does come up with the right answer. Maybe it's not on the first reference, certainly not on every page, but if you page through a dozen good links on a subject, collected objectively, you're going to get around the answer to your question.
Wolfram Alpha takes a different, more proprietary approach. They control their inputs, and their results are essentially reports. Ask about Google vs. Microsoft and you get financial charts. Ask about me and you get nothing. Same with Matt. Same, for that matter, with Palamida.
What Wolfram Alpha wants to build is an authority, and my point is that on the Internet there is no unitary authority, no final answer. Questions on the Internet lead mainly to other questions, or to data, or to opinions, usually all three.
Personally I don't want a cul de sac. I prefer the superhighway. Wolfram Alpha won't ask the Internet for the data it needs and it won't tell you what it hasn't checked. That is not a strength.
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
wolfram alpha wants to be a brain
I see it more like an academic atempt to data mining rather than a search engine.
it is not a search engine
The fact is that Wolfram Alpha is not a search engine - they don't handle unstructured data and don't intend do as far as I can tell. It handles structured data such as stock quotes, weather data, mathematical algorithms etc. It uses these sources and combines the data together on-demand to give you the information you need.
About time for Wolfram Alpha
There is no authority on the Web?
What a crock.
There is right and wrong in a mathematical sense 1 + 1 = 2.
There is true/false in an emperical sense. My house is painted purple. The wavelength of purple is easily indentified and not even a color blind person can dispute the fact.
There is preponderance of evidence in favor of one thing or another.
Yes for some things maybe there isn't a right/wrong, but for the things that matter like science, research, fact-checking, Wolfram Alpha will be a cut above the normal nonsense that current search engines spew out.
RE: About time for Wolfram Alpha
i like your example of purple, because even something as seemingly objective and simple as the physical specification of a color is debated by scientists. furthermore, judgments and discrimination of colors vary across language groups. my point is that not only do scientists debate over things you may not realize are contentious (alpha could conceivably give you results summarizing a debate), but presupposing that an objective answer exists at all is also very dangerous and naive.
I liked your example
That wiki article, though informative and quite correct in the technical sense, isn't what most people want. Do most people want to read about the debates about definitions of purple, how it changed over time, and how it varies from violet? Maybe if I typed "purple in the year 1400" into google.
I bet more people want what Alpha gives you:
http://www88.wolframalpha.com/input/?i=purple
A pic giving a representation of what most people would think of as a "purple" in the possible purple range, and how to make an approximation in RGB because you are sitting at a computer and an RGB approximation makes sense.
What will make Alpha perfect, is if/when they include links at the bottom to the "related" information about the topic (history/alternative definitions/cultural variations/etc).
If I want, I can click on them. I don't necessarily need 20+ paragraphs of context to get the basic answer.
Colorful scientific calculator meets a 10 year old version of Encarta
Useless, elitist, and almost fascist "I hate society" and "I hate the internet" dogma.
Stephen Wolfram and the people behind Wolfram Alpha hate the idea that people on both the low end of the IQ spectrum and high end might actually talk with one another to figure things both useful and useless out. That is what information and knowledge is all about.
Stalin and Hitler would love Wolfram Alpha for all of the control and ambiguity it has over things. Texas Instruments should probably be suing them.
Again, colorful scientific calculator, and one with all the information plugged into the database creators intentions rather than a pure mathematical or scientific base.
re:colorful scientific calculator
RE: Wolfram Alpha won't ask and won't tell
It's not a search engine. It's an answer machine. Searching randomly on the Internet does often lead to other questions, but half of the time it's because you can't find or trust an answer to the one you originally asked.
I think the obvious point that you're missing is that there's more than one way to treat the Internet. Browsing and searching is one way to get information, and sure, you often learn things you hadn't anticipated. I'm all for the increase of knowledge. But sometimes you just want answers, and Alpha will supply those. I don't see why there's any sort of conflict between two very different and equally valid ways of gathering info, or why being sure the pre-screened answer you're getting is correct is something to be avoided. Maybe you're just afraid to trust Stephen Wolfram.
Re:Wolfram Alpha won't ask and won't tell
it is with the interface. It presents on-screen
like a semantic search engine, which it is not.
I believe a form-driven interface would enforce
the necessary categorization to get results.
90% of my attempts to search with Wolfram Alpha
resulted in "isn't sure what to do with your
input". Not even a "could you phrase that as a
question?" like the Firesign's cyber-President.
RE: Wolfram Alpha won't ask and won't tell
Here's a quote from Wolfram Research about their goals for WA:
"We aim to collect and curate all objective data; implement every known model, method, and algorithm; and make it possible to compute whatever can be computed about anything."
Can you say Hubris?
Wolfram Alpha = Statisticus Vomitus
Absolutely unreal
lying liars, usually does come up with the right answer.
Maybe it?s not on the first reference, certainly not on every
page, but if you page through a dozen good links on a
subject, collected objectively, you?re going to get around
the answer to your question.[/i]
Translation: The internet is unreliable unless vetted. So you
agree with Wolfram Alpha all the while loudly proclaiming
your disagreement. Do you even read what you write?
There is no contradiction in the post
Relying on a single source like Wolfram|Alpha as "definitive" information makes the Internet less rich. I believe that the same could be said of Google, which page ranking is beyond self-fulfilling bordering on self-congratulatory (the top-ranked search results are seen most, so therefore linked to most, so therefore returned most as top-ranked search results ...).
So, the argument of the article doesn't fall apart on grounds of self-contradiction.
However, I think the article falls apart because Dana seems to expect that Wolfram|Alpha is search engine. It's not. It's a very sophisticated scientific calculator. You punch in a calculable problem, and it spits out the reult.
Search on "Bill Gates", and it returns almost nothing. Search on "3x^2" and it tells you a whole bunch about that particular parabola.
To put it simply: Wolfram|Alpha is for solving problems, not for providing information.
Absolutely real
there, but there's also a lot of truth, and you
need both to reach an answer. You need all the
inputs to get a reasonable output.
RE: Wolfram Alpha won't ask and won't tell
1) For a lot of things, there isn't a large concensus. If you want to find out when the Battle of Antietam was fought, that has a definite answer. If you want to find out why the Civil War was fought, that doesn't have an easily agreed-upon answer. In fact, that's one of the central weaknesses of the Internet itself: Because the information isn't cut and dried, it's not that easy to find a <i>good, correct</i> answer to a question. Especially if it's not your area of expertise to being with.
2) It's not all that clear to me that Wolfram Alpha <i>does</i> have the ability to answer scientific queries, beyond a very limited scope. The site recommends, for example, that you look up AIDS. So I type in "AIDS morbidity", and I'm greeted with...nothing. Wolfram doesn't even understand the <i>question</i>. Same thing for "AIDS mortality". Type in "AIDS", and you see that one of the statistics listed is "death rate". Typing in "AIDS death rate" does, indeed, work, but that's not the way it would normally be phrased in scientific work. And words like "incidence" and "prevalence" are no help, either.
3) As the author of this article notes, you're limited to whatever Wolfram Alpha spits out at you, in terms of a report. There's almost no customization, rather than narrowing or broadening terms.
I suppose it needs time to develop. But with all the hype built up around it, I would have thought that there would be more spit and polish when they launched it.
Dana. So what? Google and Wikipedia haven't disappeared.
If I need concrete mathematical results to a high precision, or, a one-stop-shop for scientific data Wolfram Alpha delivers the goods. If, what I'm looking for is outside the scope of W|A, I'll google around for more info.
Conventional search engines can be big time wasters. When you have to wade through multiple pages, that are ranked according to popularity. Some obscure facts maybe so far down the rankings, that you quite often get distracted by other stuff which is thrown in your path, and forget what it was you were originally looking for.
An alternative. That's what W|A offers.
Here's an interview with Wolfram. (12mins)
http://www.channel4.com/news/articles/science_technology/wolfram+alpha+goes+live/3148457
It's a niche product
Internet behind it.
Yes.
Curated instant data/information, is preferable to stale pages with dead links. Increasingly common these days, so I'd rather steer clear of it.
Don't forget Wolfram Alpha is still at an embryonic stage. So, who knows what other services might spring up from the use of the W|A APIs, by 3rd party websites.
Interesting times. Even if it makes the likes of Google and Yahoo take a fresh look at what they *aren't* providing.
RE: Wolfram Alpha won't ask and won't tell
encyclopedias, if they don't know about Wikipedia.
The math tricks are nice but in many ways, it is just
plain "dumb as a post", as we used to say.
If you want a laugh, search it for "women over 30".
Believe it or not, the answer is ".03333 females". Is
someone "channeling" Douglas Adams?
RE: Wolfram Alpha won't ask and won't tell
Walpha, however expects that you will accept its authority on the subject. Its almost a primary teacher/ pupil relationship. I value my mind, I value my and others perceptions. For all that the web has issues, they are human issues. Long may humanity have them.