Google Top Results have "added extras"

I agree with you Adam! So, let's help put a stop to it. 

This is an epic response because I need to explain the future thinking of various people in order to substantiate my claims.

Unfortunately many adult sites 'do' make out to be something else - this is why we have non-profilt organisations such as ICRA who's board members include AOL, Microsoft, BT and other such names who care about protecting children from potentially harmful content. ICRA now use RDF-CL for content labelling. 

I just read another post which I'll reply to with a link to this one - Google TrustRank. I don't know anything about this, but what I can say is that Google are co-editors with Segala and ICRA on the W3C Mobile Web Initiative conformance document. 'The' Trustmark is the RDF-CL (content label) that I talk about, the visual trustmark will be an added benefit. Also, RDF-CL is being proposed by Yahoo! also...

Child protection filters in browsers and search engines are no longer efficient enough in my humble opinion. All they do is filter content based on stuff like 'if the terms 'free' and 'breasts' appear within 4 words of each other, this means it must be adult content'. This isn't effective enough and doesn't avail of the features of the semantic web method; RDF. The system currently used by IE is called PICS - this is an old W3C recommendation but a few of us are proposing that this is replaced by RDF-CL.

Let's pretend
Let's pretent that all web masters of adult content marked up their site as 'adult content' in the title and description so users can decide whether to enter or not, do we still have a problem? The answer is yes, so let me provide you with another use case:

The most widely used website by children in the US is called american girl. As you can imagin, there are a number of sites that will come up on various search engines that don't lead you to www.americagirl.com when you search for american girl. I'm sure you can imagin a few sites containing american girls that you don't want your kids to see. How do you know if they're going to select the right site from the search results? The only full proof is to have a search annotation such as an icon for you to click, so you can see what the site is trusted for. Or, better again, add the trusted annotations to your preferences so they're either highlighted or filtered out altogether.

I'm not saying every site should be independently checked. But perhaps you could add real SEO benefit if you do, as it could ensure your site is annotated for trust. Trusted for say security (you do this already if selling anything), all it would mean is that using metadata to display this trust is equally important if not, more important than the digital siguature. If you don't believe me, check out a cut down version of what I'm talking about at www.trustwatch.com. Here you will find some of the biggest security trustmarks on the Internet. They've partnered with Ask Jeeves to provide a search engine that does annotate everything. Although this is very 'closed' using propietary technology and there's no automation behind it - I won't bore you with the detail here.

What would happen if I'm not declaring trust for something?
As with trustwatch, your site could have a ? beside it in a search results to represent 'hey, we're not sure if it's to be trusted for anything in particular as it doesn't say anything'. Or, it 'could' be filtered out if users specific a particular type of trust in their preferences.

Does any of this make sense? I'm only writing the copy to explain all of this for our new site, so forgive me if I'm being vague and long winded :)

'Today', search engines place information in search results by presenting users with the title and desction of a page. As you know, this is done using title and description tags in the header. The most common method of doing this is Dublin Core Metadata. But there is another form of meta tag that enables search engines to display more information about a site. This is called RDF-CL and is used to do the stuff I mentioned above.

The meta tag (to over simplify it) might be new to most, but it's not new to the tecnologiest or those who are proposing it as an official method to the W3C. By the way, there are a number of European Associations that are now going to use this form of content rating and quality labelling but I can't say more than that.

So, if web masters used this new type of meta tag that linked to the trusted metadata describing their site in more detail, search engines and browsers would have more information to provide users in search results. It could also increase its chances of being found by users. 

Another use case
Web Accessibility is a great reason to add this information to a website as it almost revolutionises how commercial businesses see accessibility (more on this another time). Forget about the WAI categories for a second. Let’s say the only best practice design consideration you’ve implemented is the inclusion of text resizing.

Isn’t it possible that users who need to resize text on a site will soon be provided with preferences in search engines and browsers that ensure all search results contain websites with this design feature?

Now we can demonstrate how WAI guidelines can represent actual users who may never find sites that don’t consider their preferences. So, the more guidelines you include, the more preferences you support, and the more preferences you support, the more potential customers you attract.

Let me provide you with a use case that I don't think is widely appreciated on the Web. How about, a declaration on a site to say 'there's nothing offensive to community 'x' on this site'. Then, they could decide not to visit that site before even selecting the link. Or, if you do have an adult site, you might want to make it prominent in search results to 'attract' users.

By the way, the process of making claims, adding link tags and creating trusted data so search engines can find it, is all be automated. All web masters do is check boxes on a website that belongs to the trustmark or independent verification authority. Our own site will provide this for web accessibility and mobile web standards and we're working with ICRA to provide a Segala Verified trustmark for child protection. We are working with other trustmarks so they can offer the same service. The new site will provide more user friendly information and less copy - probably because it'll be revised by my new CTO!

Cheers
Paul

http://www.icra.org/projects/content_label_incubator_0.9.htm 
http://www.quatro-project.org 
http://www.icra.org 
http://www.w3.org/2005/11/mwi-ukevent-presskit.html 
http://www.segala.com

On 12:20:10 18 January 2006 AdamCrawford wrote:

 

Hi Paul

A few thoughts on your post above:

With reference to your statement "We know that many search results contain porn sites with titles and descriptions say otherwise" - I believe that this is the worst kind of search engine spam particularly - if the fabled Disney/child-targeting terms are used - but in all honesty can not recall ever seeing any.  Why anyone would do it is beyond me; it makes no commercial sense.

As for systems for search engines to categorise websites, I see that as the responsibility of the engines themselves and not the site publishers.  After all, search engines survive off indexing websites, analysing their content and returning the (percieved) most appropriate sites nearest the top.  It is in their interest to develop their algorithyms to facilitate this.  This is what they DO.

i.e. solve the problem from the root rather than attempting to polish all the leaves and branches.

Back to your tie-in with the original thread, this also raises the question of how much the search engines can squeeze onto their SERPs before they start to look too cluttered.


As I say, just a few observations.

All the best
Adam

 

 

Hi Adam,

Google will also be co-editors of the W3C Mobile Web Initiative Trustmark document, namely 'mobileOK'. The Trustmark will be in the form of a machine-readable content label containing meta data. There will be a visual logo but that's secondary. Segala and ICRA are also co-editors with Segala producing the first draft.

I'm not saying they will use content labelling for their own search engine annotation as that would attract a few bullets in my direction. What I can say, is that @Semantics, ICRA, Institute of Informatics & Telecommunications (IIT), NCSR, Segala, T-Online, Vodafone AND Yahoo!!!! are proposing RDF-CL as an official form of content labelling to enable Trust. It's the content label that allows search annotation and filtering...

Or, as I like to call it - Content discovery and selection based on Trust.

Cheers :)
Paul

On 12:25:51 18 January 2006 AdamCrawford wrote:

 

Quick addition re: trust.

"For those of you interested in a new SEO driver 

At present, IE uses a system called PICS; a W3C recommendation for Internet filtering for child protection. ICRA have stopped using PICS as it’s outdated and doesn’t offer the flexibility of the Semantic method; RDF-CL (Resource Description Framework-Content Label). RDF-CL is currently going through what’s called an incubator activity within the W3C, as a proposed replacement for PICS. This is the *only* method that enables search engines and browsers to annotate search results based on trust."

Google is widely believed to be using "trust" in its current link analysis algo's.

Look up TrustRank for discussion over the last year or so

http://www.google.com/search?q=TrustRank

or Grid Eigen Trust for some white-papers and older discussion

http://www.google.com/search?q=Grid%20Eigen%20Trust