impl.emented» search

Microsoft Plummets on Search - Android in Trouble [Best of July '08 #3]

Jonas Bolinder — Mon, 21 Jul 2008 15:48:39 +0000

This week featured some troubling news for Microsoft and Google (July 14-20 2008):

Recent data from Hitwise shows that Microsoft has lost almost 50 percent of its market share in search from June 2007 to June 2008. In the U.S. the drop is from 9.8 % market share to 5.5 %. In the U.K. the drop is from 5.7 % to 3.7 %, and in Australia a dramatic fall from 14.7 % to 6.7 %. Google is the great winner, while Yahoo shows a minor drop and Ask a minor gain. Ask is now almost as big as Microsoft on search in the U.S. As it seems, to remain a major player in the search field, Microsoft desperately needs to get hold of Yahoo's search service.
Tags: Microsoft, search
Google's supposedly open mobile platform Android, loses in credibility among developers, as it was revealed this week that a few select top contestants of the Android Developer Challenge were given access, under the counter, to an updated version of the SDK. Additionally, the code was provided under a non-disclosure clause, rhyming poorly with the promise of open, though technically Google has done nothing wrong.
There seems to be a lack of communication from the part of Google. As suggested by AndroidGuys, more frequent updates to the official Android blog would certainly calm down some hard feelings among developers. It will be interesting to see in the coming year what impact Android will make on the mobile phone market, which currently is overwhelmed, in terms of buzz at least, by the new iPhone.
Tags: Android
Gnip, the notification proxy service, finally hooks up to the XMPP firehose of Twitter. Does this mark the beginning of the end of the troubles of Twitter?
Tags: Gnip, Twitter
OStatic points to an "executive summary of Open Source", a white paper challenging 10 Open Source Myths. Some examples of myths covered: "Open source is free", "Open source equals open standards", "Open source is not for mission-critical functions", "Open source is for non-conformists" and "Open source software is lower quality".
Tags: open source
5 tips on using Twitter to improve your online-reputation. Some examples: "Start conversations with notable peers", "Share valuable industry news" and "Monitor your Twitter reputation".
Tags: online reputation, Twitter
IPv6, the next generation internet protocol, about to replace the current IPv4 in a couple of years, might pose a security risk, writes the Wired. The problem is that it is enabled by default in many operating systems, and lots of applications may not yet be up-to-date to handle the new protocol. Affected operating systems include Windows Vista and Mac OS X. Windows XP is however safe, i.e. IPv6 is not enabled by default. You can test here which IP protocols are supported by your computer.
Tags: IPv6, security

From Yahoo Search BOSS to Rocking Fractals [Best of July '08 #2]

Jonas Bolinder — Mon, 14 Jul 2008 20:57:06 +0000

Lots of interesting stuff the past week (July 7-13 2008):

Yahoo continues their open strategy with the launch of Search BOSS (Build your Own Search Service). This is a different offering than SearchMonkey, which just allowed you to enhance the presentation of the search results. The new service has quite liberal terms. It allows you to re-order the search results and mix in other results as you see fit. These are desired options I wrote about in my first post on this blog. The terms also have a no-attribution requirement, which expressively requires you not to mention Yahoo in your search offering. Still you are not allowed to use the search index data in any way you might want to. You must provide a search service, with a search box, and only query the index in response to a search performed by a real user of your service. Also, considering the current uncertainty of the future of Yahoo, building a business on top of BOSS seems a bit risky.
Tags: search, Yahoo
From the security department, there was much talk about a potential, yet unexploited flaw in the Domain Name System, DNS, the "address book" of the internet, which translates human friendly URLs to IP addresses. Dan Kaminsky, who discovered the flaw, secretly informed the major vendors without talking to the security community, which has spawned some speculation about the significance of the discovery. The details of the vulnerability will not be publicly disclosed until August 7. Though, it is likely related to insufficient randomness provided by the 16 bit session ID of DNS. In that respect, the vulnerability is rooted in the same problem domain as the OpenSLL flaw I wrote about in May.
Tags: security
Two news from the virtual world(s): Google launches a customizable and embeddable virtual world called Lively. Second Life, the hitherto largest virtual world, reported on the first successful teleportation of an avatar between two (experimental) virtual worlds. They also recorded a video of the event.
Tags: virtual worlds
Ars Technica writes about recent speculations about a new non-Windows operating system in development at Microsoft, code name Midori. To be released some 9 years from now, probably beyond Windows 8.
Tags: Windows
The new URL shortening service bit.ly, has a host of innovative features, including an API, as reported by Marshall Kirkpatrick. Dave Winer is one of the brains behind the service.
Tags: URLs, tools
Soon we might be able to program Flash using C, meaning a speed boost that is useful in gaming applications, for example.
Tags: Flash
Moopz, a FriendFeed conversation aggregator reviewed by Sarah Perez. Among the features are item threading, noise reduction and automatic summary and tag generation. Only items containing a link are considered, thus reducing the noise. First impression is that it's a useful service. (Btw. I'm jobol on FriendFeed.)
Tags: aggregation, FriendFeed
Josh Catone is back to blogging. Now at SitePoint, a resource site for web developers and designers, where he is running the News & Trends blog. Glad to have him back.
Tags: Josh Catone
Finally, a bonus video on the math theme featuring fractals, found via Brad Feld.

Twitter is Cooling Off, but Noise is Heating Up [Best of June '08 #1]

Jonas Bolinder — Tue, 10 Jun 2008 13:46:07 +0000

Some of the interesting posts this week (June 2-8 2008):

The much hyped messaging/micro-publishing service Twitter is losing in interest among its users, partly due to the frequent and prolonged outages of the service. Dave Winer, for example, writes that Twitter is a ghost town:

[...] the upward momentum is gone, the new idea every 24 hours that so inspired us is a distant memory. Now we're going the other way. When I log onto Twitter these days, it's empty, quiet, a ghost town.

Though a recent announcement from Twitter infused Dave with some hope for the service.
Tags: Twitter
Marshall Kirkpatrick has an interesting article about noise in news, and why it might be a good idea not to filter the flow of information coming at you. For example, a study by Sanda Erdelez shows that "the more total information our minds are exposed to, the more particular items we'll be able to recall in the future." Marshall also refers to a post by Hutch Carpenter, who defines people as "signalists" or "discoverers" based on whether they "filter in" or "filter out" information with a certain content. For example, a signalist would read only content about Apple, whereas a discoverer would read all content except about Apple, for example.
Tags: news filtering
Marshall at ReadWriteWeb listened to my request last week for a comment policy, and wrote a post about how to comment without being spammy. In short: Be transparent about who you are, be super humble, and add value to the discussion. Then it's OK to provide a link to your own business.
Regarding the question of comment ownership, Dave Winer weighs in, and suggests that "I own the collection of comments on my blog, and you own the comments you've placed on my blog and all others," which is in line with my opinion on the subject.
Tags: comments
Facebook released part of their platform code as open source this week, and hardly nobody notices. A reason could be the unusual choice of license, CPAL (Common Public Attribution License), which Mike Gunderloy at the OStatic blog gives a good description of. CPAL is based on the more familiar Mozilla Public License, with a couple of modifications. One being an attribution requirement, which acts as a "poison pill" according to Mike.
Tags: open source
Mike Gunderloy in another post reports on Wikia Search, which promises to be a truly open search engine, with not only the underlying software being open source, but also the search results themselves being open to editing. A paradise for black hat SEOs and spammers as it sounds, but hopefully that could be fought back with a strong community. After all, Wikipedia manages to be fairly spam free, despite the open editing features.
Tags: open source, search
Alex Iskold writes about Zemanta, a "lazy man's" semantic blogging tool that automatically suggests related content – images, articles, videos and links as you type. Fred Wilson currently tries Zemanta, and I decided to try the provided WordPress plugin too. I had some issues with the service updating too frequently, which hangs the UI temporarily and makes the experience frustrating. It would be really nice to be able to disable the automatic updating.
Tags: blogging, Semantic Web

News and Noise [Best of May '08 #3]

Jonas Bolinder — Mon, 19 May 2008 21:15:23 +0000

This week turned out interestingly too (May 12-18 2008):

Robert Scoble has been blogging actively recently, and has an interesting post about news and noise. As he says: "The news is in the noise," which makes places like Twitter and FriendFeed interesting for the makers of news. The news items appearing on Google News and Techmeme are filtered out of the noise, which makes these news sites essentially noise free, and in some sense less interesting. Google News only tracks large well-known sites with multiple editors, which further reduces the noise, and also blocks individual bloggers from being covered. Getting on Techmeme is still feasible for an unestablished individual blogger, but it requires several "votes" by elite bloggers, as Scoble explains. On Techmeme's item selection process, Robert writes:

TechMeme: watches signaling from key members on Twitter and Google Reader. If enough people who are on the TechMeme Leaderboard Twitter and share an item on Google Reader you’ll see the item pulled onto the page.

Tags: news, Techmeme
Calais, the semantic markup service, is out with an improved version that recognizes terms from areas other than business, like media, music, entertainment, sports, medicine and healthcare. Calais also announced the release of WordPress and Drupal plugins.
Tags: Semantic Web
Robert Scoble also has a recent post stating 9 reasons why FriendFeed won't go mainstream, a few of them being: "Normal people don’t like noise", "FriendFeed doesn’t work well on mobile phones", and "Comments get fragmented, even inside FriendFeed." Honestly, Robert in a companion post states 9 reasons why FriendFeed will go main stream.
Tags: FriendFeed
The OStatic blog, one of my recent subscribes, tips about the Google Doctype project, an "open encyclopedia and reference library for developers of web applications", with resources on Web security, DOM manipulation, CSS and more.
Tags: web development
Yahoo! Search's open platform SearchMonkey is released to developers and site owners. Developers are allowed to create small applications that enhance the info about the URLs returned from a Yahoo search (but are not allowed to reorder the search results). The applications trigger on the returned search URL, e.g. I could build an application that enhance the search results from this domain (http://impl.emented.com). The structured data available to developers for enhancing the results come from core and semantic data gathered by the Yahoo crawler (called Slurp), special data feeds submitted by site owners to Yahoo and custom data services that you could create yourself.
Once developed, the applications are to be used by people searching the web with Yahoo Search. Obviously, an application that only triggers for a site which seldom show up the search results, will have difficulties to attract users.

Site owners are allowed to contribute with structured data about their site in a few different ways: By submitting Atom feeds to Yahoo, by using semantic markup or by creating custom web services.

Tags: search, Yahoo
Google Friend Connect was released this week, and caused some controversy about what the service really is. For example, Marc Andreessen thinks it is not a competitor to Ning, but "a mechanism that provides the embedding capability for Open Social gadgets to be used all throughout the web -- with the added benefit that with a Friend Connect-enabled Open Social gadget, the user gets her social context anywhere she goes." Robert Scoble on the other hand thought that Friend Connect could be a competitor of Ning. The Friend Connect site says that the service provides "a core set of social gadgets such as member management, message board, reviews, and picture-sharing." The member gadget is said to provide core social features for the site, like sign-in, invite and activity showing to other social networks, member browsing across social networks, and on-site friend connect. A site owner can also easily add OpenSocial applications to the site.
Thus, an existing site can easily be socially enabled, and as such potentially compete with a Ning social network aimed at the same audience. Also, someone considering to start a new social network now has the choice of building an own site and using Friend Connect, instead of using Ning. Building an own site is of course somewhat more complicated, but it gives a much greater degree of control. So, in conclusion I think that Friend Connect could compete with some of Ning's audience, and also provide an alternative for some social network builders. Note that for really serious social network builders interesting in implementing OpenSocial, there is still the option of using Shindig.

Tags: OpenSocial, social networks
Jeff Atwood criticizes the frequent use of XML, due to its verbosity and poor human readability. Alternatives could be YAML or JSON.
Tags: programming
RSSmeme, an aggregator for Google Reader shared items, has released an API, which makes it possible to get all kinds of specialized feeds out of the service.
Tags: aggregation

Yahoo Semantic Search and YouTube Platform [Best of March '08 #2]

Jonas Bolinder — Sun, 16 Mar 2008 16:39:55 +0000

John Musser at ProgrammableWeb scored three of four interesting posts this week (March 10-16 2008):

Yahoo Search and the Semantic Web – John Musser
Yahoo search embraces semantic technologies, with support for microformats and the more expressive RDFa and eRDF, which are methods to embed RDF in XHTML documents.
Tags: Semantic Web, search
New YouTube API Means YouTube Anywhere – John Musser
YouTube expands its API, becoming more of a platform.
Tags: YouTube, API
Latest OpenSocial News from GSP West – John Musser
Kevin Marks talks about Caja, an open source JavaScript sanitizer for OpenSocial containers, and Shindig, open source reference implementations of the OpenSocial container and server.
Tags: OpenSocial
Twine Disappoints After Semantic Web Hype – Marshall Kirkpatrick
Will Yahoo succeed where Twine fails?
Tags: Semantic Web

February Highlights I: The Social Graph API, Web 3.0, MySpace Platform and more

Jonas Bolinder — Tue, 12 Feb 2008 20:25:31 +0000

Let's start out this blogging task by listing a few interesting and thought-provoking posts I've come over the last week. Hopefully, it will be a recurring theme of this blog. The inspiration for this kind of link post comes from blogger Steve Rubel, who frequently writes "links for [date]" posts, with a short description and tags.

Lessons from the O'Reilly Social Graph FOO Camp – Dare Obasanjo
He concludes that social graph discovery, as delivered by the Google Social Graph API, is more important than social graph portability. He also suggests that people can be identified using hash values of their email address, which would keep privacy, but still enable discoverability.
Tags: Social Graph, portability, identity
A Radical Option For Yahoo: Out-Open Google – Erick Schonfeld
Yahoo should fully open up its search API to compete with Google in the long run. Currently all major search engines (Google, Microsoft, Yahoo, Technorati) have commercial and other restrictions on their search APIs, which make them hard for entrepreneurs to build on.
Tags: search, API, vision
Web 3.0: Is It About Personalization? – Josh Catone
The Web 3.0 debate continues. Personalization and recommendation are promises of the Semantic Web.
Tags: Web 3.0
Silicon Valley after a Microsoft/Yahoo merger: a contrarian view – Marc Andreessen
"Build something of value, and the world is yours," and: "The best way to get bought is to not be for sale."
Tags: entrepreneurial
MySpace Platform Launches – John Musser
The world's largest social network opens up to developers. The API builds on OpenSocial, a JavaScript API. Applications available to MySpace users in March.
Tags: MySpace, OpenSocial
Is XMPP the 'Next Big Thing' – Slashdot
The eXtensible Messaging and Presence Protocol seems to finally take off as a glue to build distributed web applications.
Tags: XMPP, protocols
Google Gathers Social Graph Information From The Web, Launches API – Michael Arrington
The Google Social Graph API can be used e.g. to add friends when joining a new social application. It uses publicly declared connections (links) between people as expressed by XFN and FOAF markup.
Tags: Social Graph, API, XFN
How YOU Can Make the Web More Structured – Alex Iskold
He suggests simple annotations to make a web page accessible to semantic "top-down" analysis: Meta tags, tagging using rel-tag markup, and using standardized class names.
Tags: Semantic Web, tagging, microformats