Does anyone know about Trustmarks? Options

[cre8pc]

I haven't heard of this before and wondered if any of you are following trustmarks. It seems to me that this could be easily manipulated and abused. What's in it for search engines?

Trustmarks and accessibility - the Paul Walsh interview

Some snips:

Can you sum up the Segala trustmark scheme for our readers?

Although trustmarks exist around the world, Segala is the first company to apply machine-readable labels to web accessibility and mobile content by providing a trustmark that is detectable by Internet search engines and browsers.

How much will it cost?

The cost of self-regulated certification is under review. Suffice to say, it’ll be very low cost so even a freelance developer in their bedroom could easily afford to label their own content.

Do you think that the search engines will start actively looking for trustmarks? Might this impact on search results, rendering trustmarks obligatory?

They will actively look for trustmarks. As I said earlier, browsers are falling over themselves to be the first to support content rating and quality labelling.

We’re also meeting with the leading search engines in the US this week.

Do you envisage a time when search engines will penalise rankings, for companies that have been reprimanded or kicked off the trustmark scheme?  

If companies have been reprimanded or kicked off the trustmark scheme they will be rated as ‘not to be trusted’. Users can then make an informed decision as to whether they want to display those sites in their search results. The bottom line is, be trusted or people won’t want to use your site.

Is the end goal to allow users to manage their ‘trust levels’ via the search engines, or via their browser?

[organicseo]

Kim, I haven't heard about this. Thanks for sharing. Trustmarks for trustworthiness of a websites. Gr8.

[Scratch]

I've implemented ICRA tags on this site: http://www.hmrc.gov.uk/

It's pretty straightforward. You apply, someone checks the site is what you say, and you get codes to hide in metacontent on your site.

[PaulWalsh]

Ah good question about trust. The good news is that it’s not possible to abuse the system as the machine (e.g. search engine) reading the content label (trustmark) is required to cross reference the trustmarks database to ensure the URI is certified. It then checks the trustmarks database to see what claims are being made. Claims aren’t stored on a claimant’s site.

Phil Archer, CTO ICRA, Daniel Appelguist, Senior Technology Strategist, Vodafone and myself, had the pleasure of Tim Berners-Lee walking past us at our meeting in the w3c nerve center at MIT last Friday. Our meeting was with Steve Bratt, COO of the w3c to discuss making our method of content labelling official – this will be open for public feedback in a few weeks.

We’re using the Semantic Web method called RDF… let me know if you want to know more!

The difference between ICRA and Segala tags is that Segala’s will be verified and monitored – new technology on the way that will enable self-regulation through user rating! Phil and I are working very closely on all of this to help ensure we end up with a more trustworthy Web. Oh, forgot to mention that Yahoo! and T-Online are on the project team.

-- Paul

[randfish]

Paul - Great to see you here.

Could you explain the basic intent behind this? What spurred the decision to create such a thing and how will you make sure that adoption is widespread enough to make it practical for search engines to use this. After all, if only 50% of "trustable" sites carry the data, it's still almost useless... Only at very high adoption rates could this truly become valuable to search engines, directories, consumers, etc.

Or, perhaps you could show me why I might be erroneous in thinking this way...

This post has been edited by randfish: Dec 13 2005, 01:37 AM

[rmccarley]

So it's basically third-party meta tags? How do they ensure you don't alter your content after getting approved?

[sufyaaan]

How do they ensure you don't alter your content after getting approved?

The difference between ICRA and Segala tags is that Segala’s will be verified and monitored ...

[PaulWalsh]

Guys, I’m so sorry for taking so long to get back to you. To tell you the truth, I was doing a little research on the Web for a presentation I’m giving in London later this week and I came across this site again – it’s bookmarked now though!!

Paul, you’re right to say that content labelling is pretty useless unless it’s widespread… IMHO it will be widespread and sooner than you might think. Machine readable trustmarks are relatively new to end users, but aren’t that new to the technologists. Also, Yahoo! has been building a community for a while now and Google has started to realize the benefit of taking users details so they can ‘try’ personalise search results.

I've cut my message into a few sections so it's a less harsh on the reading.

Paul

[PaulWalsh]

Where did the idea of providing Trust for search annotation come from?

Well, Microsoft currently use a W3C Recommendation called PICS in IE for filtering content based on appropriate content for children. We’re proposing that this ‘old’ method is replaced by the cool Semantic method RDF – we feel like we’re actually delivering a slice of the Semantic Web that people can relate to as it’s not talking about your toaster communicating with your personal diary (sorry TimBL). --> Yahoo! is one of the companies proposing this method http://www.icra.org/projects/content_label_incubator_0.9.htm.

Since writing my last post, the co-editors of the W3C Mobile Web Initiative’s trustmark; (mobileOK) conformance specification have been chosen – Segala, ICRA and --> Google. ‘The’ mobileOK Trustmark will be the RDF-Content Label, visual logos will be secondary.
The WAI has a working group called Evaluation and Reporting Language (EARL) – the purpose of this group is to create a machine readable language that represents test assertions that can be interpreted by machines – also based on RDF.

Are you seeing a trend?

Paul

[PaulWalsh]

Why I think content labelling will be widespread soon:

It'll be an offical method of content labelling in the W3C soon - probably February 2006 when we meet for a face to face in Spain.

I gave a keynote chat behind closed doors a few weeks ago to talk about content labelling and how Segala has adopted RDF-CL for search annotation and filtering capabilities that’s platform agnostic and device independent. I spoke to around 25 heads of policy for companies such as BT, AOL, MSN, Vodafone, Yahoo! and more – hosted by ICRA at the GSMA HQ and chaired by Jonathan Zittrain one of the original founders of Berkeley. Since then, we’ve been asked if we can help setup other trustmark schemes in Europe – I’ll tell you more about this when I can.

So although we provide a trustmark for web accessibility and mobile web standards, we also provide the certificate system for other trustmarks – this appears to be receiving a lot of attention nowadays and we haven’t even launched the new online system that does all this automatically yes. In fact, we’ll be providing the tools for others to include on their site, so they can provide ‘Semantic Certificates’ (as I call the content label).

Paul

[PaulWalsh]

How does it actually work?

Site owner wants to make a claim about something so engages with an independent trustmark provider. Upon the successful completion of an evaluation/audit, the site owner is given a link tag to place in the header of each certified page (or they can make a very small tweak to the server so it covers the entire domain). Site owner is also given a visual logo or small piece of text to place in the footer (logo on www.o2.com) – basically anything that a user can click to invoke a certificate as it includes a ‘report misuse of trustmark’ feature. The certificate is stored on the trustmark providers server and includes information such as expiry date, validator, site owner, domain, claim details etc.

Then you have what’s called a content label, this is a file that resides on the trustmark providers server and contains all of the information that’s included on the visual certificate but in metadata format. A tool that can read RDF-CL will spider the site, see that it contains a link tag pointing to the content label on the providers server. It checks that the RDF-CL is a legitimate RDF-CL by checking the W3C name space and then checks to see if it’s a legitimate label from the provider by checking the provider’s namespace. You can trust the claims because they reside on the trustmark provider’s server and not the site making the claims.

Take a look at www.trustwatch.com – this demonstrates search annotation to provide end users with more trust on the web for security. However, this is a closed shop and doesn’t offer the Semantic stuff that comes with RDF-CL – we’ve agreed not to patent this and leave it open within the W3C.

We’re going to offer an automated system that generates all of this stuff automatically so the cost is reduced significantly. We’ve yet to come up with a cost model but it’ll be low enough for developers to want to do this on a mass scale – needs to be accessible to everyone!

We’ve got leading agencies in the UK screaming out for this as it means they can ‘self-certify’ their own work and we just spot check. If users comes across a site that doesn’t meet its claims since the last time it was verified, they can report it directly to us from the certificate. We can then send a gentle reminder to the agency.

We’ve also been asked to provide a Segala Certified Partner Programme to differenciate accessibility savvy suppliers in different fields, so this will be online in a few months.

Make sense?

Cheers
Paul

[PaulWalsh]

The difference between ICRA and Segala tags is that Segala’s will be verified and monitored ...

sufyaaan,

(just figured out the quote thing, I'll get the hang of this web thing soon )

You’re right – although I would add that Segala will soon provide a ‘ICRA Website. Some organisations don’t fully trust the ICRA ‘self-label’, so we will help by providing added trust to their brand.

There’s lots more to all this stuff but I don’t want to bore you with all the detail unless you ask

Actually some more bed time reading can be found on an EU QUATRO - EU Safer Internet Plus that's using our Trustmark as a primary case study.

Cheers
Paul

[PaulWalsh]

I've implemented ICRA tags on this site: http://www.hmrc.gov.uk/
It's pretty straightforward. You apply, someone checks the site is what you say, and you get codes to hide in metacontent on your site.

Good stuff!

The site carries a label that uses ICRA's original vocabulary and a technology known as PICS. Although it is valid, ICRA no longer issues this kind of label. They've now moved onto RDF-CL which is what I'm referring to downstairs in one of my epics.

Paul

[Scratch]

Cool, thanks Paul.

Should I get the department to apply to ICRA for tags in the new syntax then?

Interesting to see Yahoo! keen to collaborate in this area. I'm starting to talk to them about my concept for a new community infrastructure that will intrinsically deliver trust of sites, but after a model more like web-of-trust. See thread here.

[PaulWalsh]

Cool, thanks Paul.

Should I get the department to apply to ICRA for tags in the new syntax then?

I'm pretty sure you don't have to update your current system for the new vocabulary. But think ICRA would be MORE than delighted for you to make the transition. Do you have the authority to do this? Tell Phil Archer (CTO) that I sent you BTW, they’ve invited me to sit on the board which is nice.

Paul

[PaulWalsh]

Interesting to see Yahoo! keen to collaborate in this area. I'm starting to talk to them about my concept for a new community infrastructure that will intrinsically deliver trust of sites, but after a model more like web-of-trust. See thread here.

Wow your post was even longer than mine - very interesting indeed as it's all very semantic like stuff. Perhaps we'll get together to discuss this soon as we're building a system that will automatically generate the content labels containing metadata.

You can delivery this kinda stuff by starting with RDF-CL as it’s based on the Semantic method RDF (resource description framework). Trust needs to start somewhere and where better to start than Trustmarks that are independent. For example, medical advice should be independently checked by the medical profession…

I didn't want to expand on my post any further as it was getting too deep, so thanks for opening the door again

We're releasing a new Firefox extension soon to demonstrate annotated search results for trusted 'content'. Soon after this we plan to release a full browser with technology that's not yet available (Glaxstar OwnArea), which will enable end users to rate the trust they have in claims made by a site. The weight of the trustmark is then given by end users.

Only then should we move onto relevance - the same technology and dynamically stored user profile records bookmarks anonymously so if a lot of people bookmark a particular site, we can say that it may be popular, trusted, reliable, or something like that...

As I said in a previous post, Google are co-editors of the W3C Mobile Web Initiative's mobileOK trustmark with Segala and ICRA. The actual trustmark is machine readable in the form of RDF-CL - oh and by the way, there will be a visual logo too in case someone wants to display it on a site.

[rmccarley]

This almost sounds like ISO standards for web sites...

Would that be in the right direction?

[PaulWalsh]

This almost sounds like ISO standards for web sites...

Would that be in the right direction?

Only if clients mandate suppliers to be independently certified by trustmark providers. Then again, that’s more about the process than the actual content label.

Search engines already provide the title and description in search results - this doesn't actually tell end users much about the site other than what its publisher wants you to know. Using additional metadata to provide end users with trust by way of annotation or filtering, is a good way forward IMHO. I mean, you would agree that this is good for child protection to ensure they’re not subjected to inappropriate content.

I like the idea for accessibility because I see it almost revolutionising how commercial businesses implement it (aim high!). So, if the only best practice technique currently implemented on a site is text resizing, then why not make that claim so people who require this feature, can specify the requirement in search preferences. By doing this, you demonstrate how individual guidelines can be mapped to real users who may discriminate against your site if you don’t support their needs. At the same time, you don’t discriminate against the site owner just because they don’t meet all checkpoints in a specific level of guidelines.

This will become mainstream more quickly when the creation of metadata is much easier.

Content labelling needs to be cheap as chips to ensure developers will want to do it without worrying about the extra expenditure.

[rmccarley]

Ok, I got it and it does sound cool. I'm wondering if an open source version will appear that works more like social bookmarks. Or if a social bookmarking community will run with the concept and implement it in their own structure. Reddit just added a "not for the office" section that does that in reverse.

This actually seems liek a perfect SB niche waiting to happen.

[PaulWalsh]

Ok, I got it and it does sound cool. I'm wondering if an open source version will appear that works more like social bookmarks.

This actually seems liek a perfect SB niche waiting to happen.

We've decided that RDF-CL will be open source as it needs to be for the entire industry to want it. Also, the most widely used standards and methods are open source. I'm attending a few W3C face to face meetings in February - one of them will involve RDF-CL. If you're interested, I'll report back here with an update on progress - hoping to have this official by February!

Paul