Sergey Brin is telling employees to stop making old products and start improving new ones. "For example, said Chief Executive Eric Schmidt, Google plans to combine its spreadsheet, calendar and word-processing programs into one suite of Web-based applications."

Read more of this story at Slashdot.

Intuitively, a document should be important (regardless of its content) if it is highly cited by other documents. Not all citations, however, are necessarily of equal significance. A citation from an important document is more important than a citation from a relatively unimportant document. Thus, the importance of a page, and hence the rank assigned to it, should depend not just on the number of citations it has, but on the importance of the citing documents as well. This implies a recursive definition of rank: the rank of a document is a function of the ranks of the documents which cite it. The ranks of documents may be calculated by an iterative procedure on a linked database.

A natural question is: what is Google's current approach to link buying? Of course our link-weighting algorithms are the first line of defense, but it's difficult to catch every problem case in adversarial information retrieval, so we also look for problems and leaks in different semi-automatic ways. Reputable sites that sell links won't have their search engine rankings or PageRank penalized – a search for [daily cal] would still return dailycal.org. However, link-selling sites can lose their ability to give reputation (e.g. PageRank and anchortext).

What if a site wants to buy links purely for visitor click traffic, to build buzz, or to support another site? In that situation, I would use the rel="nofollow" attribute. The nofollow tag allows a site to add a link that abstains from being an editorial vote. Using nofollow is a safe way to buy links, because it's a machine-readable way to specify that a link doesn't have to be counted as a vote by a search engine.

Google has made their web-based email client Gmail more accessible in their recent update. Now, the URL on top will change with your actions – meaning that if you switch to a certain conversation thread, or you switch to a label, or you perform a search, you can then actually copy the address on top as a kind of permalink (oldtimers among us may still remember this kind of feature from websites of the 1990s!).

Here’s why this can be useful:

• Sometimes a conversation thread contains details for a certain event, an event which you now decide to add to e.g. Google Calendar. What’s easier than copying & pasting the conversation information into your event? Right, just linking to the Gmail thread from your calendar event is. (Your thread is password-protected with your Google account credentials of course, so you won’t create an additional privacy risk either – though I would probably not post such permalinks in public places, who knows what the permalink ID can be abused for with some hack!)
• You can now share searches with a friend, e.g. tell them “Check out the previous mail I sent you at http://mail.google.com/mail/#search/grab+coffee+at+starbucks ...”.
• You can now bookmark certain labels. In Firefox, you can also then create a shortcut for such a label bookmark. Say you bookmarked “http://mail.google.com/mail/#label/job” – right-click the bookmark and select properties; in the keyword field, enter “job”. Now, you can enter “job” in the Firefox address bar at any time to be taken straight to your Gmail job label/ folder and the conversations contained within it.

[Search sharing tip via official Gmail blog.]

[By Philipp Lenssen | Original post | Comments]

TechCrunch reports that, according to the PayPerPost CEO Ted Murphy, bloggers taking part in review system PayPerPost are currently being punished with lowered PageRank. Part of PayPerPost’s sold blog reviews are using normal, non-nofollowed links, which can also have positive SEO effects (traditionally positive, at least – nowadays, such things are more likely to get one penalized perhaps).

Remember, in January 2007 Google’s head of web spam team Matt Cutts made Google’s stance on paid reviews quasi-official by writing:

<<It should be clear from Google’s stance on paid text links, but if you are blogging and being paid by services like Pay Per Post, ReviewMe, or SponsoredReviews, links in those paid-for posts should be made in a way that doesn’t affect search engines. The rel="nofollow” attribute is one way, but there are numerous other ways to do paid links that won’t affect search engines, e.g. doing an internal redirect through a url that is forbidden from crawling by robots.txt.>>

Ted Murphy, playing devil’s advocate, also alleges that TechCrunch themselves use non-nofollowed paid links, which is against the Google guidelines. Ted may be technically right: sometimes (usually monthly, as TechCrunch’s Duncan Riley says) TechCrunch is posting a “thank you” note to sponsors. In these posts, TechCrunch is linking to the sponsors without using nofollow. Assuming this thank you note is not part of the official contract with advertisers (I don’t know) then it’s a matter of judgment if you consider them an implied part of the ad deal anyway, in which case these would be paid links. Remember, Google is against paid links whether or not they have a disclosure headline or footer (like TechCrunch has)... what Google requires is what Matt Cutts calls a “machine readable” disclosure, e.g. a nofollow attribute.

Then again, we might be able to expand nofollow to even more places of “paid links,” which goes to show how witch-huntish this issue tends to become if applied broadly and strictly. For instance, what if a blogger decides to add linked disclosures when they are reporting on a company which advertised with them before – is that disclosure, as it’s linked, now an indirectly paid link? And what happens if in my blog I review a book, including a normal link to the author’s homepage, when the book has been sent to me as review copy... I may disclose that fact in my post, but isn’t this now a link paid by goods (the book)? What happens if I have a blog archive going back to, say, 2002, when there was no nofollow attribute around, and I linked to my sponsors below posts... is my blog now getting penalized for having done so because I don’t change my existing 50 posts from 2002-2004, even if my HTML used was state of the art back then? What if I’m being paid to blog and one of the blogging guidelines is to link to other parts of the blog network – are those links paid links now?

A text link advertiser adjusts methodologies

In related news, at least one paid links service seems to be implementing counter-measurements to the recent paid links penalizing by Google (if that was what it was indeed – we can’t really look into the Google black box). In a mail sent out to users of Text-Link-Ads.com which was forwarded to us, recipients were asked to update their ad code [disclosure: I consulted Text-Link-Ads.com’s Patrick Gavin on projects like SketchCast.com]. Users were also urged to remove any link back to Text-Link-Ads.com, as well as avoid adding ad disclosures such as “Sponsored Links” or “Advertisements” (or as Plan B at least using images of such disclosures) – I would think that’s because this could otherwise be a give-away for Google to trigger automated ranking penalties.

So it looks like some text link advertising systems are not giving up yet faced with Google’s moves, but rather adjusting their methodologies. Whether any of this will help, time will tell, though removing any disclosure at all (while perhaps helping against bot-downranking) seems to be deceptive to the human readers of a blog.

What’s important to note in the continuing “expanding” of nofollow uses – if you look back to the original nofollow introduction by Google in 2005, it was not meant to be applied to e.g. hand-picked ads, but only to stuff like blog spam where you didn’t see the link before (“anywhere that users can add links by themselves, including within comments, trackbacks, and referrer lists”) – is that Google has a conflict of interest here... because they sell their AdSense system to people. I can’t tell if this conflict of interest actually clouds Google’s judgment, but it seems to be clear that those bloggers which decided paid reviews or paid links are becoming too risky may now decide to switch to competitor AdSense as one option, making more money for Google.

[By Philipp Lenssen | Original post | Comments]

Thanksgiving is a major holiday for my family. With the first changes of color in the foliage, I begin the counting the weeks until the family gathers at my cousin’s for this event. So, with Thanksgiving less than a week away, I decided to share my mounting enthusiasm with one of my siblings by sending a note with a picture of what I hoped would be a gorgeous roasted turkey. I fully expected that finding such an image would be as simple as firing up my browser turning to either Google or Yahoo image search and selecting a choice roasted turkey. To my surprise this was not to be.

Here is what I found when I went turkey tracking. An image search for “turkey” at both Google and Yahoo! yielded differing, but unsatisfactory results. Google’s included live turkeys, some turkey humor, maps and images of the country Turkey, but no succulent roast turkeys on page 1. On page 1 of the image search results Yahoo! had a number of live gobblers, no maps, some images of Turkey and Turkish people, and lone image of a roast turkey – not quite what I had in mind. Both search engines suggested that I narrow my search to include “Thanksgiving turkey.” I took the suggestion, and the results were somewhat better. Google offered up more turkey clip art and a few images of roasted turkeys. Yahoo! had similar results.

As I reviewed the results, I got to thinking – a dangerous activity – where are the large purveyors of turkeys? Why are there no images from Butterball or Perdue, two large poultry producers or even the food magazines that every year have many gorgeous images of roasted turkeys within their covers? I thought initially that my mistake was not searching correctly so I gave image search another try using “butterball turkey” and “perdue turkey.”

Now we’re cooking, I thought. The results in Yahoo! for “butterball turkey” were short on images of cooked turkeys, long on displays of the logo. There were also disturbing images of maimed and dead (not in grocery settings) turkeys and images chronicling a turkey “offal to oil” initiative. With Yahoo! it wasn’t until the second page of results that a search for “butterball” turkey yielded a beautiful cooked turkey.

A search for “perdue turkey netted on Google netted a wide variety of images including some appetizing pictures of turkey sandwiches. On Yahoo! the results were startling, for I could have filled a grocery bag with raw turkey parts in neat packages, but no beauty shots of roast turkey.

Still not satisfied, I made one last try using “roast turkey” as my image search term. Google offered up lots of images of roasted turkeys, but curiously not a single one on the first page was from a turkey producer or packer. On Yahoo! there were lots of roasted turkeys some with a more homey touch since they were from Flickr, but again not a single one on the first page from a turkey producer.

Just to get a rounded picture I searched for “thanksgiving turkey” in the main “web results” pages for Yahoo! and Google. Lo and behold, Google offered a lovely cooked turkey, from a family advice site, not a turkey producer. Yahoo! had turkey clip art but no roasted turkey.

Is this a missed opportunity? Surely, I am not the only person searching for a beauty shot of a roasted turkey? Were my searches that defective? My persistence at least should have been more easily rewarded. I even took the help that the search engines offered. I’m left with the conclusion that image search has been neglected. A visit to the turkey sites reveals very sophisticated marketing programs for assisting the consumer in preparing the ritual bird. There are hotlines, chat rooms and podcasts, but alas, lowly image search appears to be overlooked. With the advent of universal search, businesses must be found in all the right places. (Note: this is the title of the workshop that I will be giving at SES Chicago along with Greg Jarboe.)

Oh! When all was said and done, I decided not to send a beauty shot, but instead to go with the image of the roasted turkey in a bikini – an effect that can be achieved with creative use of foil and the tanning effects of oven roasting.

Earlier this week, Google Blogscoped reported the change on clickable areas on Adsense units, and it is only now that the Google Adsense Blog posted an official explanation about the changes.

Whereas in the past, the whole Adsense block units including the white spaces are clickable areas, Adsense has changed it to make only the ad title and URL as clickable. This according to the Adsense Blog post was implemented to reduce the incidents of invalid and accidental clicks on those ads and to encourage site visitors to stay on the site, interact with the site content and if they deemed it necessary, click on the ads.

The post further explained that the new clickable format aligns with the text ad formats that is being implemented on Google.com. The new format would also benefit Google advertisers by making them pay for ad clicks that were intentional and would contribute to the success of their ad campaigns.

Just before this post was made in the Inside Adsense Blog, it has already elicited various reactions from both bloggers and advertisers. For Adwords publishers, these will certainly increase advertisers campaign value and satisfaction. But for Adsense publishers, this may not be good news after all.

A study conducted by some Penn State University researchers revealed that webmasters who use robot.txt files to determine which part of their sites are open or close to web crawlers,  favors Google’s spiders and bots over other search engines. Using a new search engine which the researchers themselves have created specifically for the study, it was found out that some webmasters may have been writing  robot.txt files which does not uniformly blocks or accepts search engines but allows Google and some Yahoo and MSN to crawl almost all of their pages. 

The study entitled “Determining Bias to Search Engines from Robot.txt” was given during the recent 2007 IEEE/WIC/ACM International Conference on Web Intelligence in Silicon Valley. Authors of the paper were named as C. Lee Giles, Yang Sun and Ziming Zhuang  who are all students of Penn State’s IST Department. 

Does this study explain why Google gets to give more results as compared to the other search engines? More so, does this study explain why Google is currently the top search engine today? I don’t think so. 

What interests me though, and which was not explained by the study was why those webmasters would consciously write robot.txt files that favors Google’s web crawlers to get into their pages? To gain higher page ranks for their sites? To drive more traffic?

Let’s lighten things up with a gadget post. You may have seen that Everex launched a $200 computer that runs Linux. It looks like Wal-Mart sold out of them, but not to worry: more are on the way.

Why should you be interested? Well, instead of Windows, it comes installed with gOS, which is a version of Ubuntu that is customized to work well with web-based tools from Google, Flickr Facebook, and Skype. When I heard that, I had to order one of these PCs to check it out for myself.

In this post, I’m just covering the unboxing. I’ll use the PC for a while and let you know what I think in a later post.

First, the box. The “gPC” stands for Green PC, not Google PC as some people have thought:

When you open the box, you’ll find a pretty flyer on top of the PC:

Here’s what the flyer looks like in more detail. For people that have never seen Linux, the flyer is a great little introduction:

For $199.00, I was expecting a barebones machine, but it comes with a PS/2 keyboard, mouse, and even USB-powered speakers:

The front of the machine looks like a standard computer. You can see the DVD-ROM drive, speaker and microphone jacks, and two USB ports. There are also two silver buttons for power and reset:

The back of the PC is pretty standard:

Directly connected to the motherboard, you see four USB connectors and an ethernet port. The expansion card is a modem, but I believe the machine only supports broadband connections, not dial-up.

The machine itself is light but sturdy. I jostled it quite a bit and didn’t hear anything loose or rattling around in the machine. Okay, that’s it for tonight. Tune in later to see what I think of it.