Comments on: A real Semantic Web browser, enabling trust on the Web http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/ Enabling a Reliable, Consistent and Trusted Experience Sat, 22 Nov 2008 10:02:57 +0000 http://wordpress.org/?v=2.6.2 By: Paul Walsh http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273545 Paul Walsh Sat, 05 Jan 2008 20:36:57 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273545 I forgot to mention that we're going to give our application away for free to any Trustmark Provider that currently exists - this will help bring current providers up to our level ;) I forgot to mention that we’re going to give our application away for free to any Trustmark Provider that currently exists - this will help bring current providers up to our level ;)

]]> By: Paul Walsh http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273539 Paul Walsh Sat, 05 Jan 2008 20:32:07 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273539 @Aidan - you click a few boxes and input a little text such as URL(s) and then click save. The application automatically creates the Content Label and places it on our server. It creates the visual certificate (if applicable for the assertion) and generates the relevant Trustmark. The user is immediately emailed the link tag to include in each template and a small piece of code should they wish to tweak the server instead of labeling each template. They also receive the code to display the Trustmark. Everything is automatically linked. The process of choosing the claims etc. takes just a few minutes. The time it takes to create the Content Label is almost instant. Yes, Content Labels are for unambiguously verifiable guidelines. It's not for stuff like 'My Web site is well designed'. This wouldn't be useful or practical. Take your mind away from 'Semantic Browsing'. We're providing a solution to a specific problem - that is, enabling more 'trust' on the Web – we want to provide more information about the suitability of the content on Web sites and not just the title and description. We've been working on the ecosystem for a couple of years now have other boxes to tick to help demonstrate why we think mass adoption won’t be a huge problem. Content Labels is going through the W3C as a ratified standard. This is called POWDER for political reasons I won’t bore you with unless you specifically want to know. So, Content Labels will replace PICS – the first W3C recommendation, which is used by IE today for filtering content. So, I don’t expect much of the longtail to be labeled but they’re not likely to be the sites you’d need to trust as we see it. What will make all of this easier, is if we can persuade people like you that it’s a good thing so you help by evangalising – just like the Microformats guys did :) We need to redesign and launch http://contentlabel.org and we’re having the logo/icons for Content Labels designed currently. Keep the questions coming! @Aidan - you click a few boxes and input a little text such as URL(s) and then click save. The application automatically creates the Content Label and places it on our server. It creates the visual certificate (if applicable for the assertion) and generates the relevant Trustmark. The user is immediately emailed the link tag to include in each template and a small piece of code should they wish to tweak the server instead of labeling each template. They also receive the code to display the Trustmark. Everything is automatically linked. The process of choosing the claims etc. takes just a few minutes. The time it takes to create the Content Label is almost instant.

Yes, Content Labels are for unambiguously verifiable guidelines. It’s not for stuff like ‘My Web site is well designed’. This wouldn’t be useful or practical.

Take your mind away from ‘Semantic Browsing’. We’re providing a solution to a specific problem - that is, enabling more ‘trust’ on the Web – we want to provide more information about the suitability of the content on Web sites and not just the title and description.

We’ve been working on the ecosystem for a couple of years now have other boxes to tick to help demonstrate why we think mass adoption won’t be a huge problem.

Content Labels is going through the W3C as a ratified standard. This is called POWDER for political reasons I won’t bore you with unless you specifically want to know. So, Content Labels will replace PICS – the first W3C recommendation, which is used by IE today for filtering content.

So, I don’t expect much of the longtail to be labeled but they’re not likely to be the sites you’d need to trust as we see it.

What will make all of this easier, is if we can persuade people like you that it’s a good thing so you help by evangalising – just like the Microformats guys did :)

We need to redesign and launch http://contentlabel.org and we’re having the logo/icons for Content Labels designed currently.

Keep the questions coming!

]]> By: Aidan Finn http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273433 Aidan Finn Sat, 05 Jan 2008 18:54:02 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-273433 Hi Paul, I'd be interested in hearing how your application for automating the creation a content label works. E.g. Is it fully automated or a tool for making a human more efficient in creating the content label? I can see how content labels are useful for each of the use-cases that you describe. My understanding of them is that they are suitable for making statements about a website that are unambiguously verifiable. Stuff like "this site X conforms to accessibility guidelines Y". This is fine for when you want to do search over a small well-defined subset of the web where sites in that part of the web have taken the time to add semantic annotations. But semantic browsing or searching is limited to sites that have semantic annotations and thats a tiny portion of the total web. And it's likely to be a tiny portion of the web for some time to come. This is not really a problem of the technology - it's a practical one of adoption. Anyway, I look forward to hearing more details of the technology and seeing applications of it over the next year. Hi Paul,

I’d be interested in hearing how your application for automating the creation a content label works. E.g. Is it fully automated or a tool for making a human more efficient in creating the content label?

I can see how content labels are useful for each of the use-cases that you describe. My understanding of them is that they are suitable for making statements about a website that are unambiguously verifiable. Stuff like “this site X conforms to accessibility guidelines Y”.

This is fine for when you want to do search over a small well-defined subset of the web where sites in that part of the web have taken the time to add semantic annotations.

But semantic browsing or searching is limited to sites that have semantic annotations and thats a tiny portion of the total web. And it’s likely to be a tiny portion of the web for some time to come. This is not really a problem of the technology - it’s a practical one of adoption. Anyway, I look forward to hearing more details of the technology and seeing applications of it over the next year.

]]> By: Paul Walsh http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-272859 Paul Walsh Sat, 05 Jan 2008 11:11:14 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-272859 Me neither James. I originally decided not to put the word 'Trust' in the name because we were unable to get a domain to mate. However, I don't think that's necessary anymore. Unfortunately I can't claim to have come up with Search Thresher, that was Ian Howard from Glaxstar. Me neither James. I originally decided not to put the word ‘Trust’ in the name because we were unable to get a domain to mate. However, I don’t think that’s necessary anymore. Unfortunately I can’t claim to have come up with Search Thresher, that was Ian Howard from Glaxstar.

]]> By: James Corbett http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-272751 James Corbett Sat, 05 Jan 2008 09:22:03 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-272751 Coming from an agricultural background myself I really like the thinking behind the name, Search Thresher. It seems like a perfectly descriptive label and better still it 'tells a story'. However.... I just can't say it without stumbling over it. Just like I can't say "sunshine and showers" quickly (I end up saying sHunshine and showers or sunshine and sowers). For me Search Thresher just doesn't trip off the tongue, but perhaps that's just me. Coming from an agricultural background myself I really like the thinking behind the name, Search Thresher. It seems like a perfectly descriptive label and better still it ‘tells a story’. However…. I just can’t say it without stumbling over it. Just like I can’t say “sunshine and showers” quickly (I end up saying sHunshine and showers or sunshine and sowers). For me Search Thresher just doesn’t trip off the tongue, but perhaps that’s just me.

]]> By: Paul Walsh http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-270034 Paul Walsh Thu, 03 Jan 2008 13:03:10 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-270034 @Aidan - you're absolutely right. My answer to your question will be in more detail when we tell our story in full, rather than bite size chunks where we talk about each part of the puzzle out of context. Until then, we're building a partner network of companies (similar to VeriSign) that design, build and test Web sites. We enable partners to audit and certify Web sites - each will be labelled using a Content Label (the Semantic stuff to which you refer). We have almost completed the build of an application that automatically generates Certificates and Content Labels - cutting the process time down from 4.5 hours to under 2 minutes. We intend to give this away for free to other Trustmark Providers to enable them to provide Content Labels. Use cases include - Child Protection - ICRA already label sites using Content Labels - Web Accessibility compliance - mobileOK compliance (new W3C standard) - Creative Commons (this think it's a great idea) - Privacy - Copyright - Medical - iPhone-ready (apple is looking into this) - lots more So, as you can see, we're building an entire ecosystem through which, we intend to demonstrate how it works using the extension. BTW, our extension is now endorsed by the W3C as one of four applications to demonstrate real implementations of the Semantic Web - as voted by people like Danny Ayers etc. Does that help to paint the picture? As you can see, 2008 is going to be a busy year. But we need people like you to believe in Content Labels and help by evangalising them :) @Aidan - you’re absolutely right. My answer to your question will be in more detail when we tell our story in full, rather than bite size chunks where we talk about each part of the puzzle out of context.

Until then, we’re building a partner network of companies (similar to VeriSign) that design, build and test Web sites. We enable partners to audit and certify Web sites - each will be labelled using a Content Label (the Semantic stuff to which you refer).

We have almost completed the build of an application that automatically generates Certificates and Content Labels - cutting the process time down from 4.5 hours to under 2 minutes. We intend to give this away for free to other Trustmark Providers to enable them to provide Content Labels.

Use cases include
- Child Protection - ICRA already label sites using Content Labels
- Web Accessibility compliance
- mobileOK compliance (new W3C standard)
- Creative Commons (this think it’s a great idea)
- Privacy
- Copyright
- Medical
- iPhone-ready (apple is looking into this)
- lots more

So, as you can see, we’re building an entire ecosystem through which, we intend to demonstrate how it works using the extension.

BTW, our extension is now endorsed by the W3C as one of four applications to demonstrate real implementations of the Semantic Web - as voted by people like Danny Ayers etc.

Does that help to paint the picture? As you can see, 2008 is going to be a busy year. But we need people like you to believe in Content Labels and help by evangalising them :)

]]> By: Aidan Finn http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-269956 Aidan Finn Thu, 03 Jan 2008 12:00:10 +0000 http://segala.com/blog/a-real-semantic-web-browser-enabling-trust-on-the-web/#comment-269956 The real problem with having a semantic web browser and the problem that has dogged the semantic web in general is that almost no web sites currently attach semantic annotations to their content. And for publishers there is very little benefit in adding semantic annotations to their content since there are no real applications that use them. So if you're going to use semantic annotations to filter content you're limited to a teeny tiny pool of sites from which to take content. How can you overcome this problem? Coming from a machine learning/NLP background I think that the solution is build more intelligent web crawlers that can automatically recognize different types of content and generate the annotations automatically. This scales much better as it doesn't depend on the presence of pre-existing annotations. The real problem with having a semantic web browser and the problem that has dogged the semantic web in general is that almost no web sites currently attach semantic annotations to their content. And for publishers there is very little benefit in adding semantic annotations to their content since there are no real applications that use them.

So if you’re going to use semantic annotations to filter content you’re limited to a teeny tiny pool of sites from which to take content.

How can you overcome this problem?

Coming from a machine learning/NLP background I think that the solution is build more intelligent web crawlers that can automatically recognize different types of content and generate the annotations automatically. This scales much better as it doesn’t depend on the presence of pre-existing annotations.

]]>