Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
New tagging standard released by Yahoo, Freebase, AdaptiveBlue, Zemanta... (commontag.org)
52 points by vuknje on June 11, 2009 | hide | past | favorite | 28 comments


Why do new standards have to be so heavy? Why can't it just be:

    <p ctag="wikipedia/The_Beatles">We're talking about The Beatles here</p>
Why can't they use Wikipedia? Everybody knows how to search Wikipedia, but dbpedia search (called "Navigator") isn't even linked from dbpedia.org (I had to use search engine to find out how to search dbpedia)

It even seems like anything that could be tagged should be well-formed XML document (well, XHTML).


This is a Bad Idea. You're mixing up more structure with the content. This reduces the usefulness of the tags because you and I may not agree on the tags.

What is needed are overlays that sit on top of the content. This way, you can specify tags for whatever sections you want, and I can do the same while the content remains untouched. The advantage is you can share your overlays with your friends, or you could even setup a centralized service that aggregates the overlays and uses the most common sets of tags.


I don't understand what you mean by "the content remains untouched." Referring to the original Beatles markup above, isn't it useful content to say this sentence is about the Beatles -- the topic described by that Wikipedia page. Presumably, you and I would agree on that mapping and reject tying that sentence to the Wikipedia page on the organism Beatles.

The overlaid, personal tags could be through some other mechanism that wouldn't even be part of the page. It could be some browser plugin that maps my tags for the band Beatles.


Why should the webpage server/provider have a monopoly on how you consume their content?

They should be using overlaid personal tags as well, it's just when the webpage is served, their tags would be sent over as well. Then you can choose if you want to use their tags or your own personal ones.


Like topic maps?

Or more like Ted Nelson's idea, "Embedded Markup Considered Harmful"?

http://www.xml.com/pub/a/w3j/s3.nelson.html


You want to tag somebody's pages, I'm talking about tagging my own pages (which is what this "standard" is about, as I understand). They are already mixing structure with content, they're just doing it one of the hardest possible ways (from webmaster standpoint).

Just out of curiosity - why would you want to tag somebody's content?


Just out of curiosity - why would you want to tag somebody's content?

You lack imagination! :P Tagging = annotation ;-) why wouldn't you want to be able to add some form of notes/annotations to some document??


like http://parselets.org ?


I agree. I think that you can put the tags into CSS classes and avoid most of the structure. My response got a bit lengthy, so I made it a post: http://news.ycombinator.com/item?id=654730


Horrible idea, just horrible. I hate how everyone tries to force useful ideas into HTML/XML or HTTP. You know you can create new filetypes, structures and protocols for your ideas if they don't fit nicely into the current ones??


I discussed the need for a tagging standard over 2 years ago in a blog post - check out the comments on the article, they have a lot of insight to offer.

http://neosmart.net/blog/2007/the-need-for-creating-tag-stan...

There are a couple of comments by startups dedicated to managing/organizing tags across different sites with some of the various ideas they've used to tackle this issue.


I hope it takes off, but I really doubt it will. It's too complex, and the benefits are not so clear. I understand the long term benefits/goal, but it's hard to motivate content producers to adapt something new if they can't see quick returns.


The benefits are clear, but the web browsers will have to adapt to them. Also, the implementation is fucking awful as I mentioned in other comments :/


That's not actually tagging. Maybe it's a namespace standard.


It is tagging, using RDFa. The "new standard" in the headline has a new semantic namespace, but the cool part is a new XML namespace, with a nicer way of writing RDFa tags. Example:

  <body xmlns:ctag="http://commontag.org/ns# rel="ctag:tagged">
    <span typeof="ctag:Tag"
             rel="ctag:means"
        resource="http://dbpedia/resource/Twitter />
    <span typeof="ctag:Tag"
             rel="ctag:means"
        resource="http://rdf.freebase.com/ns/en/web_2_0 />
  </body>
This relates your page semantically to the concepts "Web 2.0" and "Twitter".

Better explanation: http://faviki.wordpress.com/2009/06/11/common-tag-is-release...


I understand the spec. I understand what RDFa is.

It's still not tagging.

Tagging is a UI and organizational conceit. A minimal way to attach a nickname for a class of things to a new thing to classify it, such that users are actually able to use it, and to feedback the list of used nicknames so as to encourage convergence.

Given that I made up the term in the first place, I'm pretty sure I'm right here.


Then there'd be two ways to view the word "tagging." One is the UI and organization style you popularized through del.icio.us. The other is the common English gerund of "to tag", which is the way Common Tag is using it. Maybe it would have been better for them to call it Common Label...

Re: common "tags" through a shared namespace -- seems like it'd be a good thing to do and would increase the relevancy of our stuff to search engines. Would Google actively support Common Tag or does Google support it naturally through RDF indexing?

This approach also reminds me to review the debates on Ontology.

http://www.peterme.com/archives/000558.html

I would like to implement some kind of semantic marking for the data on my web site. If not Common Tag, what's the best way to say "this is an Author who has written that" or "this is regarding the Terminator movie (1984) directed by Cameron..."?


RDF or RDFa plus any of the myriad schemas that are available. You would use Dublin Core for the first, and you could reference the IMDB URI space for the second.

Commontag does not appear to do these things.


Common Tag uses RDFa so I'm confused (bear with me) on your objection to the scheme. The use of Freebase and DBPedia makes sense since all of their data is under relatively open license, unlike IMDB.


Some tagging sites are using it, Zigtag for example. What is the relationship between Common Tag and tagging? Or are they unrelated?


"This relates your page semantically to the concepts "Web 2.0" and "Twitter"."

I'm sorry, but no, it doesn't do anything resembling "semantics". Some background for why I say this is here:

http://www.reddit.com/r/web_design/comments/8ly99/ask_web_de...


It just looks horrible. I think all this is too complicated, which is also the reason that semantic web failed.

How is it suppose in the UI? Instead of just typing my tags, am I supposed to search for the appropriate tags first?

It is an interesting problem how to get people to tag correctly, but I don't see how a standard would help?

Also, I think modern search engines are capable of discovering that "NY" and "New York" mean the same thing.


No, it is semantic tagging based on RDF, stored on semantic databases like Freebase and DBpedia.

Practically your tags will be part of an universally shared, searchable database ... that's the real time data web baby! :D


An oldy but a goodie: http://sylloge.com/personal/2004/09/flickr-and-del.html


I was surprised to not see Delicious in the list, since they're part of Yahoo now.


We the people strongly oppose such aberration and condemn those involved in heresy to be burned at the stake!

Unless they accept going back to the drawing table...


How about this:

  <tag type="author|reader|auto" rel="label|means|etc" value="label1,label2"/>
which can be included anywhere in the doc and can be parsed/ignored according to some browser setting, leaving the whole document intact?


XML is the thing where you have angle brackets, right?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: