The argument, summed up, goes as follows: Google potentially spends up to 35,000 man hours A DAY improving its search engine's crawling of the web. It already has (limited) 'top down' (I use that term loosely in its current keyword/keyphrase-based implementation by google) 'understanding' of what is on this page; more advanced top-down semantics is the obvious next step for google, understanding the relationships between the words on this page and mine. That's a huge challenge, but Google, with its awesome workforce and historical expertise, is best placed to do it, compared to everyone else in IT/search. For them, that's probably a brilliant barrier to entry and a great way for it to cement its dominance.
But a bottom-up approach puts the hard work in the hands of the page owner (and whatever software he wishes to use to microformat his own data). Anyone can then (relatively) easily pull out and mix+match that data. So in a bottom-up web, Google has no advantage over the next johnny-come-lately; for example Cuill ( http://www.crunchbase.com/company/cuill), if it can already index webpages better than Google, could also relatively easily read and do neat stuff with the microformats that you've been so kind as to wrap around all your information. I could be wrong on all this, but if not, don't expect Google to be at all keen to encourage bottom-up semantics, e.g. by implementing microformats on all its content as Yahoo did (and conversely, it's obvious to see yahoo's strategic motivation in doing so)
Top-down: Google supreme Bottom-up: Google challenged
Ian McAllister
· 1 year ago
Fred, have there been any derivative uses of the smartlinked content in your blog? Are those smartlinks clicked on more than other links?
Like you and many others I like the idea of microformats and other ways to embed metadata and structure, but am still waiting for a killer app that actually creates value from them. Do you have a position on what that killer app is, or at least what class of app? It seems like Google has to be part of the equation. I haven't heard much about Live Clipboard lately.
fredwilson
· 1 year ago
I think its something social. Nobody has built it yet to my knowledge
jackson
· 1 year ago
Careful now, haven't you seen Terminator? 1984? The Matrix?
I think I like the old dumb web better, it's safer.....
daryn
· 1 year ago
At Startup School on Saturday, someone asked Peter Norvig (An AI/CS legend and Director of Research at Google) a question relating to the semantic web. Peter's immediate, tongue in cheek response was:
"The semantic web is the future of the web... and it always will be".
It certainly made me chuckle, but I think it has a fair amount of truth to it. Structured formats and microformats will certainly enhance the ability to easily be machine-searched, but 1. we're a long way from seeing widespread adoption and 2. there will always need to be another layer of analysis beyond that (in the search context at least) to factor in spam/abuse and other relevancy weights and factors (pagerank).
Kelly
· 1 year ago
I get a tremendous amount of useful/helpful info from your blog, not the least of which is the head's up on the new Sun Kil Moon album!! Thanks! I love those guys. Did you notice they were the band in Shopgirl?
Pete
· 1 year ago
Please, please do not use AB Meta. This appears to be yet another proprietary fork.
I have already implemented both in the front end and am restructuring the backend to utilise their enormous potentai (look up sparql on google) on my website:
These formats have enormous potential, are open, and are standards. The same cannot be said for AB meta.
Alex Iskold
· 1 year ago
How do you declare that a page contains a book using these two standard formats? Can you show us an example?
Pete
· 1 year ago
Alex,
I don't know about microfromats they are not a de jure standard, just an easy way of embedding info into html.
But RDFa is a standard. It utilises xmlns just like rdf. There is bound to be a xmlns that deals with books out there(I don't know what it is and am not going to spend the time to find out though). And if there isn't you can create one, thats the beauty of rdf.
The important thing is to follow the rdf subject -> predicate -> object format. This can be embedded in html using xhtml+rdfa which is now a w3c standard.
Cool, I hope people build plugins (or whatever they're called) for sites like Wikipedia, which I often use to link to media.
ewan
· 1 year ago
Adaptive Blue's Smart Links seems a very clever system, but I do wonder if they need to go a step further in the linking and institute some kind of revenue sharing system for the affiliate links generated
That would seem to be one way to massively increase the uptake on smaller blogs which don't want the hassle of creating all their own affiliate links to various different sites.
Alex Iskold
· 1 year ago
Hi Ewan, we already support a whole bunch of affiliate programs.
ewan
· 1 year ago
Hi Alex, that's cool, you should probably tell people!
I can't find any mention of that on http://www.adaptiveblue.com/smartlinks.html - the only place I could find it was on the install page, which people are probably only going to find after they've decided to use your software.
That said, I think an option for you to do it all for people, handling the affiliate accounts, etc, and just dropping the payment into their paypal account for a 20% cut would be a popular option - who wants to create 6 affiliate accounts and payment details to 6 different companies, and if you create a new supported affiliate the user has to manually configure that too.
Alex Iskold
· 1 year ago
Hey Ewan,
As it turns out not a lot of people incentivised by these because you can't make a lot of money unless your blog is highly trafficed or unless you are on many blogs, like we are.
In terms of handling stuff via PayPal - that would be hard for us to do because then we need a lot of tracking infrastructure like LinkShare, etc. We are just a connector or a pipe right now.
jasoncrawford
· 1 year ago
The semantic web is an exciting idea. However, I'm not yet convinced that it will ever come about, primarily because I don't yet see what will motivate people to start structuring their content by marking it up with metadata.
The AB Meta FAQ you linked to answer a lot of "what" and "how", but doesn't address the "why". Why would anyone spend the time and effort required to do this markup? Or rather, what will motivate enough people to do it that the semantic web has a chance of coming about?
I think if the semantic web is going to happen, it's going to require some immediate payoff from doing the markup.
Why do you do it, Fred? (Other than interest in a portfolio company and the desire to play with a new technology.)
fredwilson
· 1 year ago
The smartlinks add something useful to the link. That's the 'immediate benefit' that I get
Fred
jasoncrawford
· 1 year ago
Interesting... I read AVC in Google Reader, and the smartlinks don't show up, so I never noticed them. And when I went directly to avc.blogs.com, it took a long time for the smartlink icon next to "April" to show up (I'm using FF2 on a MacBook).
Anyway, the smartlinks have clear value, but I still don't see where the immediate payoff from AB Meta is.
Leroy
· 1 year ago
why not just a browser based plug-in that allows users to summarize a pages content, CNN.com style with the bullets? Leave it open so people can continuously fine tune....but would save having to read entire articles....
All Comments
The argument, summed up, goes as follows: Google potentially spends up to 35,000 man hours A DAY improving its search engine's crawling of the web. It already has (limited) 'top down' (I use that term loosely in its current keyword/keyphrase-based implementation by google) 'understanding' of what is on this page; more advanced top-down semantics is the obvious next step for google, understanding the relationships between the words on this page and mine. That's a huge challenge, but Google, with its awesome workforce and historical expertise, is best placed to do it, compared to everyone else in IT/search. For them, that's probably a brilliant barrier to entry and a great way for it to cement its dominance.
But a bottom-up approach puts the hard work in the hands of the page owner (and whatever software he wishes to use to microformat his own data). Anyone can then (relatively) easily pull out and mix+match that data. So in a bottom-up web, Google has no advantage over the next johnny-come-lately; for example Cuill ( http://www.crunchbase.com/company/cuill), if it can already index webpages better than Google, could also relatively easily read and do neat stuff with the microformats that you've been so kind as to wrap around all your information.
I could be wrong on all this, but if not, don't expect Google to be at all keen to encourage bottom-up semantics, e.g. by implementing microformats on all its content as Yahoo did (and conversely, it's obvious to see yahoo's strategic motivation in doing so)
Top-down: Google supreme
Bottom-up: Google challenged
Like you and many others I like the idea of microformats and other ways to embed metadata and structure, but am still waiting for a killer app that actually creates value from them. Do you have a position on what that killer app is, or at least what class of app? It seems like Google has to be part of the equation. I haven't heard much about Live Clipboard lately.
I think I like the old dumb web better, it's safer.....
"The semantic web is the future of the web... and it always will be".
It certainly made me chuckle, but I think it has a fair amount of truth to it. Structured formats and microformats will certainly enhance the ability to easily be machine-searched, but 1. we're a long way from seeing widespread adoption and 2. there will always need to be another layer of analysis beyond that (in the search context at least) to factor in spam/abuse and other relevancy weights and factors (pagerank).
Please get behind the W3C standard RDFa:
http://rdfa.info/
And if you cant do that use Microformats:
http://microformats.org/
I have already implemented both in the front end and am restructuring the backend to utilise their enormous potentai (look up sparql on google) on my website:
http://localhero.biz/
These formats have enormous potential, are open, and are standards. The same cannot be said for AB meta.
I don't know about microfromats they are not a de jure standard, just an easy way of embedding info into html.
But RDFa is a standard. It utilises xmlns just like rdf. There is bound to be a xmlns that deals with books out there(I don't know what it is and am not going to spend the time to find out though). And if there isn't you can create one, thats the beauty of rdf.
The important thing is to follow the rdf subject -> predicate -> object format. This can be embedded in html using xhtml+rdfa which is now a w3c standard.
Pete
http://localhero.biz/
That would seem to be one way to massively increase the uptake on smaller blogs which don't want the hassle of creating all their own affiliate links to various different sites.
I can't find any mention of that on http://www.adaptiveblue.com/smartlinks.html - the only place I could find it was on the install page, which people are probably only going to find after they've decided to use your software.
That said, I think an option for you to do it all for people, handling the affiliate accounts, etc, and just dropping the payment into their paypal account for a 20% cut would be a popular option - who wants to create 6 affiliate accounts and payment details to 6 different companies, and if you create a new supported affiliate the user has to manually configure that too.
As it turns out not a lot of people incentivised by these because you can't make a lot of money unless your blog is highly trafficed or unless you are on many blogs, like we are.
In terms of handling stuff via PayPal - that would be hard for us to do because then we need a lot of tracking infrastructure like LinkShare, etc. We are just a connector or a pipe right now.
The AB Meta FAQ you linked to answer a lot of "what" and "how", but doesn't address the "why". Why would anyone spend the time and effort required to do this markup? Or rather, what will motivate enough people to do it that the semantic web has a chance of coming about?
I think if the semantic web is going to happen, it's going to require some immediate payoff from doing the markup.
Why do you do it, Fred? (Other than interest in a portfolio company and the desire to play with a new technology.)
Fred
Anyway, the smartlinks have clear value, but I still don't see where the immediate payoff from AB Meta is.