Bing, Google Squared, Wolfram Alpha, and Google Wave reviewed

Since these were all announced/beta-released/released recently, here’s a look at each and comparisons between them.

 

Bing

Microsoft says Bing is a “decision engine,” but really, it’s a search engine. One of their commercials starts with a woman saying “We really need to find a new place for breakfast.” Bing’s result for “Where should I eat breakfast?” won’t help you make that decision. First off, Bing doesn’t know where you are (for those who don’t know about geo-ip, click here to find out where you are). Bing also seems to confuse the word “where” with “why,” returning a list of results that describe why it’s important to eat breakfast. “Places to eat breakfast in (your zipcode),” does better depending on what zip code you enter, but in general returns search results, just like Google does.

That’s the problem with Bing: it’s too similar to Google, without doing anything that much better than Google.

I’ve already described some of the issues with their shopping experience. In short, I don’t think companies like PriceGrabber, Shopzilla and Shopping should pack their tents just yet.

Bing’s travel site seems to be getting some good and some mixed reviews. People like the interface, but perhaps that’s because Microsoft used a good template by copying Kayak.

Bing’s image search offers more related information, which is useful for specific things like celebrity results, where it offers biographies, quotes, news, etc., but less so for other image queries, where the options provided are pretty much the same as what Google provides.

Bing’s news layout seems dramatically less useful than Google’s.

Bing’s maps seem reasonable enough, but Google maps has a lot of user-added content and cool mashups, so there’s a large transition cost and no reason to switch.

In summary, Bing is nice enough, but it’s no revolution compared to Google, the way Google was compared to what came before it.

If you’re curious, the site Blind Search takes a query and shows the results for Bing, Google, and Yahoo in a generic format with no attribution and lets you vote for the one you like. This bypasses the many additional features of the sites — word definition, image search, maps, travel, calculator, etc. — but it does get to the main purpose of a search engine: search. When I looked the results were Google 45%, Bing 33%, Yahoo 21%, but apparently since then someone has been gaming the system and the owner took the results down. You can still compare for yourself, though.

 

Google Squared

Google Squared is an experiment in Google Labs that takes any query that can return a set of related items, and presents a table of information on those items. Ask it about arctic explorers and it will return Amundsen, Mawson, Scott, Peary, and others in the rows. The columns show the explorers’ names, a picture, descriptions, birth dates, dates of death, and places of birth. Each individual cell is sourced independently, so the grid may come from a few sites, or many. Information on the source and the alternatives are available by mousing over/clicking on a cell.

There are two things that make Squared interesting.

First, Squared works on a broad set of queries. Ask Squared about Fokker airplanes and it will tell you their manufacturer, the number built, and their length. “wwii battles?” Date, location, and result. Squared often doesn’t have all the information: it only has the outcome for three of the seven battles it lists. Still, it can be an interesting pastime to play with Squared to see what sets it can handle. “Stephen King novels” works fine; so does “Stephen King short stories.” A query I just tried expecting it to fail was “Vlad Taltos novels.” Vlad Taltos is the main character, not the author, but Squared nailed it. 

Google Squared often works on queries that wouldn’t seem to define a set at all. Richard Nixon returns a list of U.S. Presidents, interestingly none of whom are Nixon (although amusingly his picture displays for Andrew Johnson). Joseph Stalin returns a more eclectic list, including Hitler, Moscow, and Leninism.

The second interesting feature is that Squared can add columns. Squared’s results for astronauts shows a list including Alan Shepard, Neil Armstrong, and Buzz Aldrin — good results. It shows descriptions of them, their date and place of birth, and date of death (amusing since all but one have entries, but only Shepard has actually died). To the right of the columns displayed is the “Add” option. It’s a dropdown menu with options (hopefully) appropriate to the square displayed. For Fokker airplanes it’s Height, Crew, Wingspan, Maximum Speed, and Service Ceiling. For astronauts it’s Awards, Status, Education, and puzzlingly, Release Date and Author.

That Google’s algorithms are able to come up with (generally) reasonable suggestions for additional columns is interesting, but even better is that you can enter just about anything, and Squared will do its best to come up with that information for each of the entries. For example, entering “age” returns (74) for Shepard (his age at death, and in parentheses), and reasonable values (not necessarily current) for most of the rest. 

Squared is similar to the early days of Wikipedia: it shows a great deal of promise, with shaky results in many cases. Obviously the algorithms behind it are trying to solve a difficult problem, and there’s plenty of room for improvement. Squared tacitly acknowledges this by allowing you to remove any row you like from the results. For now the results are hit and miss, but a lot of fun to play with.

 

Wolfram Alpha

Microsoft says Bing isn’t a search engine, but it is; people call Wolfram Alpha a search engine, but it’s not — strange world, eh? Some people described Alpha as a Google-killer, but it’s not. This becomes clear when you search for something Alpha doesn’t understand, like Fokker airplanes, and Alpha responds that it doesn’t know what to do with your query. Some have come close to saying what Alpha is, but not quite. Wolfram Alpha is an online expert system. Taken on that basis, Alpha is a significant achievement. It’s a great tool, and fun to play with.

The obvious question is why its creators have avoided that moniker — perhaps to give the impression of coming up with something totally new.

One thing is clear, though: to use Alpha productively you need to read the documentation. Like any expert system, Alpha addresses specific problem domains; you need to know what they are to get productive results. You can get a general sense of the types of questions Alpha will answer by looking at the examples page. How to format the questions appropriately can be a mystery.

Alpha has excellent capabilities in math, physics, chemistry, space, health and medicine, history, and a variety of other subject areas.

As Alpha grows and adds new problem domains it will become more and more useful, but it will continue to be necessary to understand what it can and can’t do, and how to get it to divulge what it knows.

Several people have compared Google Squared and Wolfram Alpha, and that’s unfair to both. Alpha is a hand-curated expert system. It’s narrow, but presumably getting wider as it grows. Squared is an algorithm for extracting related information from the web as a whole. Alpha will often refuse to give an answer, but when it does, it is likely to be correct. Squared will more often give an answer, but there is a greater likelihood that the answer will be wrong, or partial. Alpha does an excellent job presenting any data it has using graphs, charts, and plots. Squared’s presentation format is limited to a basic table with whatever pictures and text it finds on the web. Compare their results for “tallest mountains.” Here’s Wolfram Alpha. Here’s Google Squared. Alpha’s results are more accurate: from the top down they’re the world’s tallest mountains. Squared’s results are all tall mountains, but not the tallest in order.

To get Alpha to display a table of information like Squared presents, it’s necessary to reformat the query to list the specific mountains. In addition, Alpha didn’t seem to like Kangchenjunga West, one of the mountains it had listed for the original query. As previously described, Alpha does a great job of presenting accurate information: the height, country, and year climbed for each of the mountains. It also displays a map locating the mountains, and gives the air pressure at the top of each.

Squared displays pictures and descriptions, along with the height, location, and the range in which each mountain lies. Other suggested columns include “First Ascent” and “Coordinates,” both of which seem to return good data. 

 

Google Wave

Google Wave is the toughest to review. It hasn’t been released yet (it’s supposed to be available before the end of the year), so this assessment is based on the demo video and documentation available. People have tried to define Wave in various different ways. Many seem to think of Wave simply as a new way to do email, which is fair since Google themselves have repeated said about the creation of Wave that they tried to think of what email would be like if it were invented today; but I think a more general definition is better, so here it is: Wave is a general communication method achieved by real-time collaborative wiki page editing with customizations to support various specific scenarios. Breaking that down:

Wave is a general method of communication. As such it can replace email, instant messaging, discussion boards, and perhaps even blogs, photo sharing sites, social networking sites, and even general web pages. A tall claim, and it remains to be seen what Google’s full intentions are. They seem to be focusing mainly on email and instant messaging, but the demo video and developer documentation imply much more.

Wave is real-time: when two people share a wave, each can see what the other types as they type it, character by character, within milliseconds. This is of course configurable. In the scenario where you’re composing an email (an email-like wave, actually) to a friend, you can set it to not appear to your friend until you’re done with it, so you can type out the various descriptions you have for that party last night before deciding whether it was outrageous or just crazy.

Wave is collaborative: when two or more people share a wave, each can edit the document. This can occur sequentially, in the same way that an email thread happens, or immediately, the way instant messaging does, or even in-place and simultaneously, like a few document sharing tools do, with multiple people’s edits appearing real-time at various places in the document.

Waves are fundamentally wiki pages, although the term “wiki” has certainly come a long way in the last decade. Originally a simple(r) mark up language designed to make it so anyone could understand how to edit a web page, and a server that allowed anyone to participate, wikis today often have WYSIWYG editors, and Wave seems to have a very good one. Wave also implements an extremely detailed permission system for viewing and editing, to replicate the functionality of email, instant messaging, and other forms of collaboration. Wikis generally include the entire edit history of the wiki page, and Wave raises the bar on this by having a history slider that moves back and forth through the versions as fast as you drag it.

Based on the demo video, Wave seems like a very credible alternative to email and instant messaging. Of the two, IM seems like the easier replacement. Many businesses use AOL Instant Messenger. I can see Wave replacing that by company decree about a week after Wave launches. Wave is better than existing instant messaging tools for several reasons:

• Wave is immediate. In regular instant messaging, you see a message that the other person is typing, but you don’t know whether they’re close to done or have gone off to get coffee. So you sit and wait.
• Wave supports easy formatting.
• Wave automatically archives your communications (many IM clients do this as well, obviously).
• Wave’s interface makes it easy to add several people to a conversation, and they can all see the history of the communication immediately.
• With Wave there is no distinction between real time (instant messaging style) communications and time shifted (email style) communications, so if someone has to leave to go to a meeting, they can just come back later to add their thoughts to the conversation.
• It should be easy to set up an in-house Wave server so your corporate email/IM-type information doesn’t leave the building.

None of these are knock-out blows to existing instant messaging, but taken together, they’re quite compelling. Replacing email is more complicated, but Wave offers advantages over email as well:

• Message threads are a pain, and no two clients handle them in exactly the same way. Once people start replying inline, all semblance of organization is out the window. With Wave a message thread is nothing more than a wiki page. If the conversation gets unwieldy, it can be edited to clean it up, with the entire conversation in the history if needed.
• Adding someone to the thread is easy, and they have access to the history as well.
• Email conversations sometimes tend toward real-time conversations. With Wave there is no distinction, so rapid-fire emails simply become an instant message conversation. There shouldn’t really be a distinction, and Wave doesn’t make one.
• Wave is another chance to get rid of spam. I don’t have a good answer for how to prevent it, but at a minimum Wave isn’t hampered by having to support legacy systems unconditionally. Wave will presumably have to interoperate with existing email somehow, but it doesn’t have to do so indiscriminately. 
• Wave supports rich formatting options, and since it works in (almost) any browser, it should support rich formatting uniformly, meaning that everyone can drop pictures into waves and know that their friends will be able to see them.

The question of interoperability deserves more discussion. To me, Wave will be a success if I can ditch email for it completely. That means that email from my parents on MSN needs to end up in my Wave inbox somehow — and hopefully that spammers can’t do the same thing. For that to happen I’d be willing to accept having a personal whitelist, and force anyone not on it to actually use Wave (with whatever new anti-spam capabilities it brings) to send me something.

Wave is (going to be) open source and highly extensible. The demo and screenshots show people taking polls, playing chess, posting photos to blogs, and accepting blog comments, all with waves. One other interesting aspect of Wave is that there is conceptually only one instance of any given wave. That means that if I turn a wave into a blog post, people can comment on it, and I will see those comments and be able to reply to them, directly through my standard wave inbox. I won’t need to go to my blog to respond.

In case it isn’t obvious, I think Google Wave is going to be revolutionary, and I can’t wait.

 

Conclusion

Bing: nice home page, but I’m not switching

Google Squared: easy, fun, may be useful, but double-check your answers

Wolfram Alpha: if you’re willing to learn what it’s good at, and how to use it, very useful within its area of expertise

Google Wave: could change the way we communicate, can’t wait to see if it does