Friday, December 21, 2007

Remember the Googley Milk


There's so much Google in Remember The Milk, a popular task management service, that you may wonder why Google didn't acquire it. Remember The Milk has a simple interface, can show your tasks inside Google Calendar, Gmail, iGoogle, and makes your tasks available offline thanks to Google Gears.

The integration with Gmail is available only in Firefox (through an extension) and only in the new version of Gmail. "Remember The Milk for Gmail is a Firefox extension that allows you to manage your tasks in Gmail (complete, postpone, and edit tasks), add new tasks (and connect them with your emails, contacts, and Google Calendar events), automatically add tasks for starred messages or specific labels, and much more!" So you can add events to Calendar directly from Gmail, you can flag messages for follow-up and see which tasks are connected with a certain contact. You can also create tasks every time you star or label a message and let RTM to auto-complete tasks using names from your contact list or Google Calendar events.


"It happens to be the product that inspired us when we started building RTM back in 2004 -- that's right, the one and only Gmail. We've always thought it would be incredibly cool if you could manage your tasks alongside your mail -- and have your tasks know what's on your calendar and who your contacts are too," explained RTM's blog.

And Remember The Milk fits great in the landscape by using visual elements from Gmail Chat, Google Reader, by using even small features from Gmail (like undo) and creating a section for tasks in the settings. This is definitely the service that has the best integration with Google services and it looks so good in Gmail that it should be a part of Google's mail application.

Orkut OneBox for People Search

Most of the data from Google's social network orkut wasn't included in search results because orkut requires authentications to see profile pages. Now Google intends to index the information from user profiles and show it in search results as an OneBox.

"Some orkut profiles will appear in Google search results as an orkut OneBox. A OneBox is a summary of your orkut profile, including key details like your name, photo and location. A OneBox appears only when someone who is logged in to orkut performs a search on Google.com for another orkut user. Keep in mind that this project is still in its experimental stage, so for now, only a small percentage of orkut users will see the OneBox in their search results," mentions Google in orkut's help center.

In the privacy sections of the settings page there's an option that lets you remove your profile from the orkut OneBox, but the description doesn't correspond to the one from orkut help center: "show my orkut information when my orkut friends search for me on Google.com".

As previously mentioned, orkut could be the base of a new people search service that integrates data from different sources.

Let's Test Powerset

Powerset, the search engine that shows results for natural language queries, started to let its testers enter any query.

From the about page: "Our unique innovations in search are rooted in breakthrough technologies that take advantage of the structure and nuances of natural language. Using these advanced techniques, Powerset is building a large-scale search engine that breaks the confines of keyword search." For now, Powerset only indexes pages from the English Wikipedia and it's not publicly available (but you can request an invite).

For example, Google shows almost the same search results for [Pyra Labs acquired by Google] and for [When was Pyra Labs acquired by Google?]. Even if the answer can be found in one of the first snippets, Google doesn't highlight it or display it prominently, like you can see in Powerset.


Do you know a query that shows irrelevant results at Google? Post it in the comments and I'll upload a screenshot of Powerset's results. Obviously, to compare Powerset's results with Google results, you need to restrict Google to en.wikipedia.org, the only site currently indexed by Powerset.

Updated.

Query #1 (from Matt Cutts): How many states are in the United States?
Conclusion: the first 10 Powerset results are terrible. On the other hand, Google shows the answer in a OneBox, but also a strange book search OneBox at the top: "How Many Doctors Do We Need?" by Duncan Yaggy, Patricia Hodgson.



Query #2: What Nobel Prize winners were born in Russia?
Conclusion: Google's results are better even if you restrict them to Wikipedia. The top result from Wikipedia (the third Google result) is a page titled Noble laureates by country. Only few of the people mentioned in Powerset's results are Russian who won the Nobel prize and there's no complete list.



Query #3: Who was the last president of United States?
Conclusion: The results #3, #4, #6 mention George W. Bush, but there are other names of former presidents. Google's fourth results has this title: "George W Bush: Last President Of The United States Of America?".



Query #4: Who are the founders of Yahoo?
Conclusion: the second result includes the answer, but it's only partially highlighted in the snippet. Google's top results has a good snippet: "The two founders of Yahoo!, David Filo and Jerry Yang, Ph.D. candidates in Electrical Engineering at Stanford University, started their guide in a campus."

Post Blogger Comments Using Your Own Domain

Blogger has recently added support for posting comments by authenticating with OpenID. The integration is now live for all Blogger blogs and that means you no longer have to show your identity by using a Google account. You already have an OpenID if you have an account at AOL, WordPress, LiveJournal or at one of the other OpenID providers. You can be your own OpenID provider, but this is a little more complicated.

A simple way to use your own domain as an OpenID is to delegate it to another OpenID provider. That means you should get an OpenID from a site like MyOpenID and add these two lines in the head section of your homepage (you should replace http://NAME.myopenid.com with the real OpenID):




Now you can use your domain or subdomain everywhere you're allowed to enter an OpenID. To post a Blogger comment, select "Any OpenID" from the drop-down and enter your URL. You'll be redirected to the OpenID provider to enter your password.



Google removed the option to enter your URL when you don't sign up with a Google account, but you can still show the URL if you use OpenID.

Update: Blogger brought back the option to enter a URL when post unauthenticated comments.

Google Gears Wants to Upgrade the Web

http://2.bp.blogspot.com/_ZaGO7GjCqAI/R2quYimvc8I/AAAAAAAAHCc/Qx9eztPh_ik/s400/google-gears-beta.png

"A gear is a component within a transmission device that transmits rotational force to another gear or device." (Wikipedia)


Most people think that Google Gears is a way to make web applications work offline, but it's much more than. Google Gears addresses many limitations from today's browsers and wants to make the browsers more powerful.

Here's a list of features that could be included in the next releases, but you can find more in Google Gears wiki:

* image manipulation - "a module to give Javascript a way to resize, crop and compose images together on the client side. This will allow, for example, images to be resized into a web-friendly format before being uploaded to a photo album. Another use is for composition of images together as an efficient alternative to server-side composition or CSS layering. Yet another use is for basic photo editing - a user can edit a photo with instantly applied changes before uploading it to the server."

* location - "an abstraction for the various LBS APIs that currently exist on mobile platforms (GPS-based, network/cellid-based). The API consists of the Location class, which encapsulates various location attributes (latitude, longitude, etc), and also provides the means to query the platform for a location fix." This suggests that Google Gears will be available for some mobile phones.

* desktop shortcuts for web applications that lets you click on an icon from your desktop and open the "application" in your browser. This could make the transition from the desktop to the browser easier, even if some people could find it confusing. Mozilla Prism is a similar initiative.

Google's Dion Almaer describes Gears in a very plastic way. "We get to drive a few makes of cars (browsers) on the (information) highway. When we want new features, we have to wait for a new model to come out, and recently it feels like Cuba. The top selling car is a 1950’s Chevy. As drivers that are passionate about the driving experience, the Gears team is trying give everyone a foundation to replace the engine, even as you drive."

Google Gears is still in an early phase of development and it could include many other features, but it will be interesting to see how Google intends to push its adoption. The Gears-enabled services could ask the users if they want offline access or other fancy features, Google could also include it in the toolbar or distribute it with popular applications from third-parties. For now, Google Reader, Remember the Milk and Zoho Writer are the most important applications that use Gears, but next year most Google services will use Gears.

Who Are My Gmail Contacts?



Google has an awkward way of dealing with contacts in its communication apps. In Gmail, your contact list includes all the people you've ever replied, but you can also add other contacts manually.

"Email addresses are automatically added to your Contacts list each time you use the Reply, Reply to all, or Forward functions to send messages to addresses not previously stored in your Contacts list," according to the help center. While this should save you some time and effort, your contact list will include a lot of people you wouldn't have normally added. For example, if somebody sends me a tip for this blog and I reply to thank him, my contact list includes that person. Gmail doesn't have an option to turn off this feature, so all you can do is to either ignore your contact list or create a group that contains only your real contacts.

By default, if your conversation with someone includes more than 2-3 replies, that person is automatically added to your list of Google Talk friends. To chat with someone you normally need to ask for permission, but this feature bypasses the annoying question because Google assumes you really know that person. Fortunately, you can disable it in Gmail's settings and Google Talk, but not many people will do this. "If there are other Gmail users whom you frequently email, you'll be able to chat and see each other online without having to send an invitation. Gmail automatically determines which contacts you'll be able to talk to without having to invite each other." (Gmail help center)

So the rules are simple:

Rule #1:
if you reply to someone's email, that person is added to your Gmail contact list.
(Error #1: you may not know that person)

Rule #2 (opt-out):
if you reply to someone's messages more than 2-3 times, that person is added to the list of Google Talk friends.
(Error #2: see error #1. Also that person may not be your friend.)

So it's quite likely that your Gmail contact list and Google Talk friends list include people you don't know. Now that we have these two lists (obviously, Google Talk friends are also Gmail contacts), you may wonder where you could use them.

In Google Shared Stuff, a rather obscure social bookmarking site, all the web pages you bookmark are public, but there's also a page with "stuff from people you know". That page shows the most recent bookmarks from your Gmail contacts, but many of these contacts are people I don't know.

Google Reader added a feature that shows shared items from your Google Talk friends. Here's how it was introduced:

One of my favorite uses for Reader is to share interesting stuff with my friends. I click "Share" whenever I find an interesting item, be it hilarious or serious. This way, all my friends can subscribe to my shared items (and I to theirs), and we can easily see if a friend has found something interesting. This can be inconvenient, as I have to distribute my shared items link to my friends and vice-versa. So, we've linked up Reader with Google Talk (also known as chat in Gmail) to make your shared items visible to your friends from Google Talk.

Except that, according to the rule #2, my Google Talk friends aren't necessarily my friends.

Google will probably continue to use your contact list for other services, so at some point your Gmail contacts or Google Talk friends might see your public documents, photo albums, notebooks, personalized maps, blog posts. All of these public actions dynamically generate a news feed (the way you know it from Facebook) and your contacts should be entitled to find things about you. The main problem is that your contact list has been generated automatically and has little to do with you. Those people aren't necessarily your friends, your family, your co-workers, they're just some people you happened to email at some point.

Before using theses random lists of people to broadcast information about you, Google should clearly define their purpose and let you manage them. The problem with creating a social layer over Google's web apps is that Google is not a social network and your contacts are not your friends and not even people you know.

Holiday Village, the New Winter Theme for iGoogle

There's a new winter theme in iGoogle: Holiday Village. This mysterious theme has appeared and disappeared for almost a week and many people wondered why Google doesn't release it for everyone. Apparently, Google realized that Christmas isn't far away and the theme should be available in your iGoogle pages.

Here's the description: "watch the snowflakes fall and the lights twinkle in this cozy and quaint alpine village."

If you can't see the theme, you might try to copy this code in your address bar while visiting iGoogle and then press Enter. It seems that not all computers can access the theme's files, so this code could reset the theme to the default one.

javascript:_dlsetp('preview_skin=skins/holidayvillage.xml');


Translation Service for Google Talk

Google launched translation bots for a lot of language pairs, even more than the ones available at Google Translate. To use them in Google Talk or in any other IM client that supports Jabber, you need to add one of the bots as a friend, start a conversation with the bot and enter the text you want to translate. Google's bots have a simple format for addresses:

[original_language]2[destination_language]@bot.talk.google.com

For example, if you want to translate text from English to French, you need to add en2fr@bot.talk.google.com. To translate Chinese text in English, add zh2en@bot.talk.google.com. Here's the list of languages pairs and you'll notice many new pairs, displayed in bold below:

ar2en, en2ar (Arabic<->English)
bg2en (Bulgarian->English)
de2en, en2de (German<->English)
de2fr, fr2de (German<->French)
el2en, en2el (Greek<->English)
es2en, en2es (Spanish<->English)
fi2en (Finnish->English)
fr2en, en2fr (French<->English)
hi2en (Hindi->English)
hr2en (Croatian->English)
it2en, en2it (Italian<->English)
ja2en, en2ja (Japanese<->English)
ko2en, en2ko (Korean<->English)
nl2en, en2nl (Dutch<->English)
ru2en, en2ru (Russian<->English)
uk2en (Ukrainian->English)
ur2en (Urdu->English)
zh2en, en2zh (Chinese<->English)


While this is a great interface for translating short texts (you can access Google Talk from Gmail, from google.com/talk or in many other ways), you can also use the bots to translate conversations in real-time. You need to invite the two corresponding bots in a group chat, so this doesn't work in the desktop client, which still does not support group chats. For some reason, when I invited one of the bots, it was offline, but it still translated my messages.

It would be great if Google adds this as a standard feature of Google Talk and you only need to enter your native language, but the translation quality is not that great and it could cause problems in some situations. The good news is that Google's statistical translation system advances really quickly and you'll see more and more languages pairs in the future.

Slowly Transitioning to Online Software

New York Times has a long article about the differences between Google and Microsoft in terms of vision. "The growing confrontation between Google and Microsoft promises to be an epic business battle. It is likely to shape the prosperity and progress of both companies, and also inform how consumers and corporations work, shop, communicate and go about their digital lives. Google sees all of this happening on remote servers in faraway data centers, accessible over the Web by an array of wired and wireless devices — a setup known as cloud computing. Microsoft sees a Web future as well, but one whose center of gravity remains firmly tethered to its desktop PC software."

Eric Schmidt, Google's CEO, envisions that 90% of today's computing tasks can be moved online. "To explain, Mr. Schmidt steps up to a white board. He draws a rectangle and rattles off a list of things that can be done in the Web-based cloud, and he notes that this list is expanding as Internet connection speeds become faster and Internet software improves. In a sliver of the rectangle, about 10 percent, he marks off what can't be done in the cloud, like high-end graphics processing." (my emphasis)

Google also thinks that people don't use all the features that are available in many desktop applications. "If you're creating a complex document like an annual report, you want Word, and if you're making a sophisticated financial model, you want Excel. That's what the Microsoft products are great at. But less and less work is like that," said Google's Dave Girouard.

And for Google, things are going in the right direction: more people have access to fast Internet connection, users don't want to keep their data on a single computer as they found the advantages of sharing and collaborating online. There's also the advantage of a much lower price for storage and computing. Google's "vast data centers are designed by Google engineers for efficiency, speed and low cost, giving the company an edge in computing firepower and allowing it to add offerings inexpensively."

For now, 2.000 companies start to use Google Apps every day (most try the free version), Google Docs had 1.6 million US users last month (source:Compete.com), Gmail doubled its US users to 20.1 million in November (source:comScore).

Google Is All About Large Amounts of Data


In a very interesting interview from October, Google's VP Marissa Mayer confessed that having access to large amounts of data is in many instances more important than creating great algorithms.

Right now Google is really good with keywords, and that's a limitation we think the search engine should be able to overcome with time. People should be able to ask questions, and we should understand their meaning, or they should be able to talk about things at a conceptual level. We see a lot of concept-based questions -- not about what words will appear on the page but more like "what is this about?" A lot of people will turn to things like the semantic Web as a possible answer to that. But what we're seeing actually is that with a lot of data, you ultimately see things that seem intelligent even though they're done through brute force.

When you type in "GM" into Google, we know it's "General Motors." If you type in "GM foods" we answer with "genetically modified foods." Because we're processing so much data, we have a lot of context around things like acronyms. Suddenly, the search engine seems smart like it achieved that semantic understanding, but it hasn't really. It has to do with brute force. That said, I think the best algorithm for search is a mix of both brute-force computation and sheer comprehensiveness and also the qualitative human component.

Marissa Mayer admitted that the main reason why Google launched the free 411 service is to get a lot of data necessary for training speech recognition algorithms.
You may have heard about our [directory assistance] 1-800-GOOG-411 service. Whether or not free-411 is a profitable business unto itself is yet to be seen. I myself am somewhat skeptical. The reason we really did it is because we need to build a great speech-to-text model ... that we can use for all kinds of different things, including video search.

The speech recognition experts that we have say: If you want us to build a really robust speech model, we need a lot of phonemes, which is a syllable as spoken by a particular voice with a particular intonation. So we need a lot of people talking, saying things so that we can ultimately train off of that. ... So 1-800-GOOG-411 is about that: Getting a bunch of different speech samples so that when you call up or we're trying to get the voice out of video, we can do it with high accuracy.

Peter Norvig, director of research at Google, seems to agree. "I have always believed (well, at least for the past 15 years) that the way to get better understanding of text is through statistics rather than through hand-crafted grammars and lexicons. The statistical approach is cheaper, faster, more robust, easier to internationalize, and so far more effective." Google uses statistics for machine translation, question answering, spell checking and more, as you can see in this video. The same video explains that the more data you have, the better your AI algorithm will perform, even if it isn't the best.

Peter Norvig says that Google developed its own speech recognition technology. "We wanted speech technology that could serve as an interface for phones and also index audio text. After looking at the existing technology, we decided to build our own. We thought that, having the data and computational resources that we do, we could help advance the field. Currently, we are up to state-of-the-art with what we built on our own, and we have the computational infrastructure to improve further. As we get more data from more interaction with users and from uploaded videos, our systems will improve because the data trains the algorithms over time."

Google is in the privileged position to gain access to large amounts of data that could be used to improve other services.
Google Is All About Large Amounts of Data by Ionut Alex C

Find Your Gmail Contacts in orkut

It's not clear whether the fact that Gmail automatically adds to the address book all the people you've ever sent messages makes it irrelevant or comprehensive. It's not clear whether orkut will ever become a respected social network in other places than Brazil and India. But one thing is for sure: orkut really wants to become a serious alternative to Facebook. After adding the news feed, third-party apps via OpenSocial (not yet live), more privacy controls, orkut lets you find your Gmail contacts on orkut and add some of them to yout friends list.

"Your Gmail address is already filled in, and since you've already logged into orkut, we don't have to ask for your Gmail password once again. We'll show you the contacts who are already on orkut, and let you choose which ones to add as friends. For contacts who aren't on orkut yet, we make it easy to choose the ones you want to invite to orkut. You can even add a personal message to send with the invite. Soon, we'll be adding support to import your contacts from other email accounts," promises Jude Britto from orkut.

Even if you don't use orkut, some of your Gmail contacts might have orkut profiles with juicy details.


I wonder if Google would have integrated orkut's data in search results if the service was more popular. orkut lets you find people using advanced filters, but unlike Facebook, you can visit any profile if you are logged in with a Google account. In Brazil and India, orkut is the fourth link in the navigation bar and it sends you to a universal search page with results grouped in three categories: users, communities and topics. orkut, the personals and people profiles categories from Google Base could be the foundation of a people search service.

Google Profiles


They're coming and it will be difficult to get away without having one. Google Profiles will be integrated in most Google services so you have a coherent identity and a simple way to manage your contacts.

"A Google Profile is simply how you represent yourself on Google products — it lets you tell others a bit more about who you are and what you're all about. You control what goes into your Google Profile, sharing as much (or as little) as you'd like."

Until now, you could create profiles in Blogger, orkut, Google Groups, Google Co-op and all of them could contain different information. You could also add photos in Gmail, Google Talk and orkut, so the situation started to become confusing.

The new Google profiles are already available in Shared Stuff, Google Maps, Google Reader and will be added to other web applications. For example, in Google Maps you'll find the link to your profile at the top of the page.

Profiles are public and contain basic information about yourself: a nickname (the real name is displayed only to your contacts), your occupation, your location, a list of links, a photo and a short description. They are embedded as iframes in pages that showcase user-generated content (personalized maps, shared bookmarks).


It's not a stretch to see that these profiles are the perfect host for your activity streams and your public activities could become a part of the profile (uploading photos to a public album, bookmarking web pages, posting a new blog post). It's basically FrindFeed's widget that can be contemplated at Paul Buchheit's blog.

A side-effect of the public availability of your profile is that people can find it. "Can people do a Google search for my name and find this profile? It depends. If you put your full name in the Nickname field, pages on which your profile appears may be returned as results by Google." You can already find more than 100 profiles attached to Google Maps pages. Maybe Google will even create a directory for profiles and start to suggest friends based on personal descriptions, location and activity streams.

Features that Just Work

The best features are those that just work. They don't require special settings, answers to all kinds of questions, advanced interfaces and reading the manual to understand how they work.

Even if Google's search algorithms have evolved a lot in the last 10 years, Google works without having to type anything else than a query. It doesn't ask if you want personalized results, recent web pages or if your query has anything to do with celebrities. You can get away with spelling errors because Google automatically detects them, you can also type ambiguous queries without seeing a dialog that asks you to be more explicit. Now you don't even have to specify if you want images, news or videos because Google adds them to the list of search results.

The "add subscription" box from Google Reader is smart enough to take care of all the possible situations. You can enter a feed, but you can also enter the address of a site. Unlike My Yahoo, Google Reader detects if the site has feeds and picks the first one. But what if the user types New York Times? Google Reader shows the feeds that match this query and lets you choose the one you like. The feature could be improved by automatically subscribing to the top result for navigational queries like TechCrunch, where there's a single best result (at least in English).


Another feature that just works without human intervention is auto-save. You'll find it in Gmail, Blogger, Google Docs and it basically saves your text frequently so you don't lose what you type if your browser crashes or your Internet connection is down. You don't have to setup this option or mention how often you want to save your text.

But things aren't that great in Google Calendar, where you have to choose between 5 options if you want to add a calendar:


... or when you constantly need to choose between iGoogle and Google Reader when you subscribe to feeds, even if you only use one of the two products:


... or when iGoogle asks you location after you select a theme even if you've already added your location in Google Maps.

The features that just work are most of the times barely visible and that's a good thing. They're a part of a system that delivers what you want without constant interruptions and annoying workarounds.

Google Reader Shows Shared Items from Your Friends

Google Reader finally becomes social and automatically subscribes to the list of shared items from your Google Talk friends (which is a subset of your Gmail contacts). Your friends will also see your shared items, but you can remove those you don't want to read you favorite posts.


Google uses the profiles from Shared Stuff and Google Maps.


"If any of your friends on Google Talk are using Reader and sharing items, they'll automatically show up in the Google Reader sidebar under Friends' shared items. You can read these items in a combined list, or click the "+" icon to expand the list and see the shared items from each of your friends," according to the help center.


This feature was long overdue and Google made the right decision to not allow every Gmail contact see your shared items, but you should also remember that people can automatically become your Google Talk friends if this option is enabled in Gmail: "Automatically allow people I communicate with often to chat with me and see when I'm online".

A lot of people complain on Google Groups that this is a privacy issue and you may not want to share those items with your Google Talk contacts. "I think the basic mistake here (...) is that the people on my contact list are not necessarily my "friends". I have business contacts, school contacts, family contacts, etc., and not only do I not really have any interest in seeing all of their feed information, I don't want them seeing mine either. This is a major privacy problem." Another user has an interesting way to use the shared items: "I share items with a very specific audience. I know the people who have subscribed to my shared items and that allows me to share items I think would be of interest to them. I'm not trying to be Boing Boing. I'm just trying to be me. Now, when I know that any of my "friends", from any possible social circle, from any level of familiarity, can see what I share, I really can't be me at all." The problem is that Google Reader didn't manage to explain what "share" really means and didn't consider that Google Talk contacts aren't necessarily friends.

Robert Scoble, who has a very popular shared items feed, noticed that Google Reader doesn't filter duplicates (a certain post can be from one of your subscriptions, but also in your contacts' shared items). "Google Reader now is bringing me TONS of duplicates from people. This clutters my all items feed and keeps me from finding new, original items."

Filtered feeds are a great way to deal with information overload: if your friends have a special interest in a domain, they could offer you a summary of the most interesting things that happened in that particular field. Of course, your friends could also share funny things from the web or share everything they read. "If you don't want to see shared items from any particular friend, you can hide their items from your list; just click on the friend's name and click the Hide button."

In the next updates, Google Reader should let you comment on shared items or chat with your online friends about certain posts, separate your shared items with tags and recommend shared items from people that have similar interests with you.

Notes:
* this only works in the US English interface (you can change the language in Google Reader's settings).
* if you don't see all your Google Talk friends that use Google Reader, don't worry. You'll see them after they open Google Reader and find out about the new feature.
* my shared items (and the the feed)

Google Video's Redesigned Homepage

Google Video's homepage has been redesigned and now focuses on trends and rankings. There's a section for hot videos that includes the most blogged, most shared and most viewed videos. Movers & shakers continue to include videos with a growing popularity, while the recommended videos continue to be uninteresting or already watched.

If you expand some of the sections, you can see a daily/weekly/monthly archive of the most popular videos. Unfortunately, I couldn't find the feeds for this useful data.

Google Video doesn't intend to remove the video hosting feature. The homepage shows videos recently uploaded to Google Video and each search result page includes a link to the upload form. There's also an option to restrict your search to videos hosted by Google.

The homepage still doesn't communicate that Google Video is actually a search engine. It's surprising to see a search engine promoting videos from AOL on the homepage, showing enhanced features like the inline player only for YouTube and Google Video and being so biased towards YouTube and Google Video when ranking search results.

Subscribe to Custom Search Results

I reported in October that Google added Subscribed Links to the preferences page and renamed them as search add-ons. Google removed that feature a couple of days after I wrote that post, but now it's back for good, at least in the US English version of Google.

You'll notice at the bottom of the preferences page a list of subscriptions and a link to a directory of subscriptions. Each Subscribed Link matches some of your queries (for example, queries about weather, nutrition, traffic) and shows information extracted from a database. It's like a Google OneBox created by third-parties, but you can decide if you want to see it and there's always an option to unsubscribe.


According to the FAQ, "using Subscribed Links, you can add information created by providers you trust to your Google search results pages. Whenever you search on Google in an area of their expertise, you'll see a custom result from those providers in your search results. (...) Your Subscribed Links will appear in the fourth search result position", even if they aren't counted as a real search result.

Here's what happens if I subscribe to the Nutrition Facts knowledge source and search for [calories in cream of celery] or even [cream of celery]:


Even if other results offer similar information, I get the benefit to find what I want without having to click on a search result. In some cases, it's not necessary to enter the context of my search because my subscriptions already provide that context ([cream of celery] shows information about calories because I'm interested in nutrition).

The Subscribed Links are some special search results with custom snippets and a limited range of queries that are displayed based on your preferences. They allow you to perform specialized searches on a general-purpose search site.

And if you're not satisfied with the limited number of Subscribed Links that are available, you can create one and promote it on your site.

Annotating the Web with Google Toolbar

The latest version of Google Toolbar integrated Google Notebook, the service that allows you to save interesting parts from a web page for later. Google also started to highlight the fragments from a web page that were previously saved and show your comments in a tooltip.

As you probably know, notebooks are by default private, but you can invite some of your contacts to collaborate. This way you can create a shared list of bookmarks with annotations that allow you find the most important parts of a web page and read your collaborators' opinions. Your annotations are highlighted with a different nuance of yellow so you can distinguish them from those added by your contacts.

For now, you can't export the annotations or subscribe to someone's public annotations, but it's not hard to see these features added in the next iterations of Google Toolbar. Trailfire already lets you find trails, "collections of web pages, assembled and annotated by any Trailfire member". The collections of notes could be used to guide people to the most important parts of a Google search result or they could become a part of a dynamically-built encyclopedia page.

The current implementation from Google Toolbar provides you with a simple way to share clips from web pages with your friends and also add comments.

Google Knol, an Encyclopedia Written by Experts

Udi Manber from Google writes about a new service for sharing knowledge called knol.

"Earlier this week, we started inviting a selected group of people to try a new, free tool that we are calling knol, which stands for a unit of knowledge. Our goal is to encourage people who know a particular subject to write an authoritative article about it. The tool is still in development and this is just the first phase of testing."

Unlike Wikipedia, Knol wants article written by people who are an authority on a subject. The articles written in Knol are more like scientific papers because they have clearly defined authors, references, even if they don't necessarily include original research. "We believe that knowing who wrote what will significantly help users make better use of web content," explains Google.

Knol will be open to anyone and it will be interesting to see how Google verifies your identity. If you claim to be Nelson Mandela, how can Google know that this claim is real?

Google provides tools for editing the text, hosts the article and allows you to monetize it (but that's not required). Obviously, Google can't guarantee that an article is accurate or complete and that's the small role of a community: rate the articles, write reviews and suggest edits.

"Once testing is completed, participation in knols will be completely open, and we cannot expect that all of them will be of high quality. Our job in Search Quality will be to rank the knols appropriately when they appear in Google search results. We are quite experienced with ranking web pages, and we feel confident that we will be up to the challenge."

Google did a similar thing when it allowed people in the news to comment on news articles. The number of comments is very small (around 150 for the last 30 days), but they're interesting and add a lot to a story.

Wikipedia managed to become one of the most important sites on the web even if it allowed anyone to edit an article. According to the online encyclopedia, "The English Wikipedia edition passed the 2,000,000 article mark on September 9, 2007, and as of December 13 it had over 2,125,453 articles consisting of over 921,000,000 words. Wikipedia's articles have been written collaboratively by volunteers around the world and the vast majority of them can be edited by anyone with access to the Internet. Having steadily risen in popularity since its inception, it currently ranks among the top ten most-visited websites worldwide." But one of the most important problem of Wikipedia is that articles lack credibility and it's difficult to tell if they contain accurate information. Assuming Google manages to verify people's identity, Knol could solve this problem.

Udi Manber, who heads the project, told Danny Sullivan that the main goal is "to help people put knowledge on the web that doesn't currently exist, which in turn should make search better, since there will be better information out there." Google certainly hopes to attract important authors and that's probably the reason why Udi Manber talked about the project on Google's blog. But how will the project scale when it becomes available to the public?

YouTube Visualization for Discovering Related Videos

YouTube has a very cool visualization that lets you discover related videos. Just go to a YouTube video, click on the full-screen button and then click on the small button that shows a network. A lot of video ballons will pop up and the configuration will change once you hover over a button. It's an interesting way to discover videos and it reminds me of visualization sites like Liveplasma and Musicovery.


Google Answers to Relaunch as Google Q&A

Google intends to relaunch the question-answering service Google Answers, which was closed last year. In Google Answers, "users could post a question (...) and specify how much they were willing to pay for an answer. A researcher then searched for the information they wanted and posted it to Google Answers." Some of the former Google Answers researchers built a similar service at Uclue.com.

Google Q&A, code-named Confucius, no longer has paid experts and works in a similar way with Yahoo Answers. Google Q&A was launched in Russia in June and in China, two months later.

Here's a message from Google's translation console:

"Q&A - This message is a name of successor for Google Answers. We will use it in OneGoogle toolbar, which you see on top of google.com page in the more.. section. Also, please use full name to translate it. That is, Questions and Answers. Abbreviation should be used only for English. URL showing this message: toolbar on top of http://www.google.com."


It's interesting to note that Google Q&A is also the name of a feature that displays answers to simple questions at the top of Google's search results page. Maybe Google will combine the facts automatically extracted from web pages with the explicit answers from the new service.

Will Google Give Up on Universal Search?

"With universal search, we're attempting to break down the walls that traditionally separated our various search properties and integrate the vast amounts of information available into one simple set of search results." (Marissa Mayer, May 2007)

Google already integrates in Universal Search video results, books, news, images, local search results and now subscribed links. The next ingredients could be blogs, products, scholar papers and some Google Base verticals.

Even if Google says it ranks results from specialized search engines along with standard web results, this is more like a figure of speech since Google displays groups of search results from Google News, Image Search and Google Maps, not individual results. More likely, Google determines how relevant some image results are for a query and finds the proper placement. Instead of actually mixing results from many specialized search engines, Google includes some of the previous Oneboxes in a controlled fashion. Google adopted a conservative approach in order to not clutter the search results and because it's difficult to compare completely different entities. How to compare some search results for "Christmas" from Digg with a news about Christmas ornaments?


Ask.com decided to separate the specialized results and display them in a column generally used for advertising (like in this search for Christmas). Now Valleywag has a screenshot of a Google experiment that displays the universal results in a separate column.


Does this mean Universal Search didn't work as expected and Google tries alternative interfaces?

Google Calendar Sync for BlackBerry


Synchronization is probably the most requested feature for Google Calendar and Google has finally done something about it. It's not quite what you would've expected, but it's still a promising start. Google Sync is an application that lets you synchronize your Blackberry's calendar with Google Calendar.

"Using your BlackBerry smartphone's native calendar, you can now access your Google calendar even when you don't have network coverage and be alerted for upcoming appointments with sound or vibration. Your Google Calendar stays synchronized whether you access it from your computer or your phone. You can add or edit entries right on your BlackBerry smartphone or on your Google Calendar on the web."

The application should be available at http://m.google.com/sync if you visit the page from your Blackberry's browser. According to Google Mobile Blog, it supports Google Apps accounts.

This week, Google launched a mobile package of applications for Blackberry and an unified interface for iPhone. Google really tries to build solid mobile offerings, but only Google Maps is available for the most important mobile platforms. Hopefully, Google Sync will be released for other platforms and it will synchronize your contacts as well.

Google Toolbar 5 for IE

There's a new version of Google Toolbar for Internet Explorer that brings a lot of updates and new features.

Google Toolbar's buttons can display gadgets when you click on the little arrow. Here's the YouTube gadget in a custom button:


You'll find a lot of built-in buttons in the Settings: almost every Google service has a custom button. In addition to gadget support, buttons can now display alerts, show status text next to the icon and create actions based on text selection.

Bookmarks and notebooks are now completely integrated: each note is a bookmark with some clipped content and notebooks are a way to organize and share your bookmarks. You no longer need to install a separate plug-in for Google Notebook.



When you add a note, the text will be highlighted every time you open the page. If you also add some comments, they're displayed in a tooltip when you hover over the text (you're annotating the web):


You can now create profiles for AutoFill, there's an inline find bar similar to the one from Firefox and Google can replace the "Page not found" errors with useful suggestions.


Probably the most interesting new feature is that you can choose to save your settings on Google's servers so you can access your buttons, settings and AutoFill data even if you don't use the same computer.

This beta version of Google Toolbar 5 doesn't integrate with Google Docs and doesn't synchronize your local documents with the documents stored on Google's servers.


Ratings and Comments in Google's Community Maps

Google LatLong Blog reports that Google Maps added a new community layer for the My Maps feature: comments and ratings. That means you can create personalized maps, transform them into wikis, import content from KML files and search results, add photos and videos, embed the map into your site. Even though these community maps are included in search results, I think they're hardly noticeable unless a popular site links to them or embeds them. Ratings and statistics will help Google create charts and show things like "popular community maps in your area" or maybe connect your search profile with some maps.

Maybe Google should also create a default community map anyone can edit. This way, you'll be able to add placemarks directly on the main map without having to create a personalized map.

Gmail Mobile Application, Now Available for Google Apps


The mobile Gmail application for Google Apps that used to only work if you had a Blackberry is now available for all mobile phones. Just go to m.google.com/a on your mobile phone and download "Mail by Google". The application looks exactly like the recently released Gmail Mobile 1.5, except that it has a blue icon and supports Google Apps accounts. You'll also notice that you can login with standard Gmail accounts.

It's unclear why Google didn't add support for Google Apps in the standard Gmail application (available at gmail.com/app), but it may have something to do with Google's branding and the Gmail name.

Tuesday, December 11, 2007

iGoogle's Winter Theme

For those who like winter, for those who understand its beauty and playfulness, there's an iGoogle theme called Winter Scape. It was available back in March, when Google launched themes, but it was replaced by a theme for spring, now known as Seasonal Scape.

Description: "Brrrrr! Wintertime is here and there's snow on the ground. This friendly snowman will keep you company on your iGoogle homepage as the day goes by."

How it looks like:





How to get it: the theme is no longer available from iGoogle's interface, but you can still add it to one of your tabs. Copy this JavaScript code:

javascript:_dlsetp('preview_skin=skins/winterscape.xml');void(0);

go to iGoogle, paste the code in the address bar, press Enter and click on the Save button in "Select a theme for this tab" to keep the new theme.

Picasa 2.7 for Linux


Google has a preview release of Picasa 2.7 for Linux. The previous version launched last year didn't include the features from Picasa 2.5+ for Windows (like the Picasa Web integration), so this is a good opportunity to add these features to Linux. Here's what is new in Picasa 2.7 for Linux:

* Upload to Picasa Web Albums
Use the new "Web Album" button to post your best photos online to share with friends and family.

* Save edits to disk
Save edits, undo saves, and revert to the original file with ease. We've got batch saving too! Picasa will even match the jpeg quality of the original. Right-click on your saved files to try the new "locate original" feature.

* Folder hierarchy views
Browse through folders Explorer-style. Use the button at the top of your Albums List to try them out.

* Improvements to Import
Import into an existing folder. (...) We've made importing photos from your camera faster too.

* Better RAW support
Now you can work with RAW files from the Canon 30D, the Nikon D200, Adobe DNG files, and more.

* Many other enhancements
Larger thumbnails, better caption editing, ability to configure the row of buttons, special "Starred Photos" album, search by ISO and focal length.

The release still doesn't support videos, full-screen slideshows and it's still using WINE.

More about Picasa for Linux at Google Groups.

Google Starts to Index Images Uploaded to Blogger

Even if this sounds hard to believe, Google Image Search started to index images uploaded to Blogger in December 2007. Until this month, all the images were prevented from being indexed by search engines for unknown reasons. This move is closely related to the fact that the images hosted at Picasa Web Albums started to be indexed by Google.

The chart from Google Analytics shows the number of referrals from images.google.com for this blog:


... and here are some results from Google Image Search:


The images uploaded from Blogger's interface are hosted at Picasa Web Albums, but they're also available at subdomains like bpX.blogger.com, where X is a digit. Another Blogger oddity, inherited from Picasa Web, is that you can't directly link to an image (if you click on a link to one of the two images uploaded above, you'll see a dialog that asks you to download the image). Blogger even has a workaround for this silly restriction: it automatically creates web pages that include the pictures (here's a link to the same image, but this time the image is included in a web page).


Images uploaded before August last year, when Blogger launched the latest major upgrade, are still not crawlable.

Google Finds Less Search Results

Everybody should know that when you use a search the engine, the number of search results is just an estimate. You only look at the first 10 or 20 results anyway and, in some cases, the search engine doesn't let you access more than a certain number of results. For example, Google only lets you see the top 1000 results, mostly for efficiency reasons.

"When you perform a search, the results are often displayed with the information: Results 1 - 10 of about XXXX. Google's calculation of the total number of search results is an estimate. We understand that a ballpark figure is valuable, and by providing an estimate rather than an exact account, we can return quality search results faster." (Google help center)

But recently something has changed in Google's algorithm that estimates the number of results. Here's a comparison between the number of results for [Moby] in May (notice the recently-launched bar that used gradients) and today:

Searching for Moby (May 17, 2007)

Searching for Moby (December 10, 2007)

From 15 million results to only 2 million results, there's a long way. For the same query, Yahoo estimates 18,900,000 results, Microsoft finds 7,730,000 results, while Ask only finds 4,089,000 results. Notice that all the other three major search engines show bigger numbers than Google. You might think that this query is just an exception, but that's not the case. Almost every query shows much less results in Google than in other search engines.

A search for [Google] shows (the numbers may vary across different data centers):
* 132,000,000 results - Google (screenshot)
* 1,610,000,000 results - Yahoo
* 244,000,000 results - Windows Live
* 281,620,000 results - Ask.com

And even if this estimate has never been reliable, it's strange to see a such an obvious inaccuracy. If you use complicated queries (more than 3-4 keywords), the estimates become more accurate and Google starts to show more results than other search engines.

In other related news, Google started to treat subdomains the same as directories for some queries. "For several years Google has used something called host crowding, which means that Google will show up to two results from each hostname/subdomain of a domain name. That approach works very well to show 1-2 results from a subdomain, but we did hear complaints that for some types of searches (e.g. esoteric or long-tail searches), Google could return a search page with lots of results all from one domain. In the last few weeks we changed our algorithms to make that less likely to happen in the future," explains Matt Cutts.

Related:
Google - Yahoo Comparison
Persistent queries (Greasemonkey script)
Index size and estimation (given two search engines, what are the relative sizes of their indexes?)