The Downward Spiral of Microdata | Hacker News Reader

The Downward Spiral of Microdata

manu.sporny.org

30 pointsAliCollins20 comments

We've all been talking about "The Semantic Web" for a very long time. I would love to be proven wrong but it all reminds me of EDI[1] which for more then 30 years has tried to develop standards around data formats. I think all of these efforts so far have been doomed to die a horrible death because they don't have an obvious benefit for those involved in the massive investment required to get it right. I think theoretically it's wonderful to imagine a world of infinitely parseable machine consumable information. If anything the robots will appreciate it when they take over the world. Meanwhile I'm not convinced that Humans can attach enough value to the effort to make it worthwhile. But I could be wrong, this wouldn't be the first time and I hope not the last time. :-)

[1] http://en.wikipedia.org/wiki/Electronic_data_interchange

I believe any "universal ontology" like this is theoretically impossible in the same way that finding an isomorphism between integers and reals is impossible. I am also struck by how appealing everyone finds such classifications.

But then i am something of an existentialist and doubt the transcendent reality of logic and reason.

While having nothing to do with "existentialism", you might appreciate this. It's the introduction to a book originally published in France in the late 70s, but that might just serve as a world wide web manifesto...

http://interconnected.org/home/more/2005/06/1000Plateaus00Rh...

(You seem philosophically-minded)

The Semantic Web should remind you of every schema and interchange format you have ever seen. That's its purpose. The graph is a very versatile data structure conceptually.

The Semantic Web cannot die because it is multiple concepts and no single concept at all. The Semantic Web is ideas not (just) implementations, and ideas never die.

Worst of all, for most businesses there is no benefit in sharing their data. And if there is, only consumers of data care about standardized schemas and methods of access. For producers, standardization means customers can more easily switch to competitors. Also, deviation from standards should be expected because producers of data want to have unique selling points. The schema.org stuff is useful, it lets companies use their web site as a decentralized phone book by adding metadata for Google (Maps), e.g. phone numbers, addresses and the like. It's the Semantic Web of opening hours.

From a product standpoint, schemas are small technical details with relatively low influence on the outcome: it just needs to work. Unless you're operating at extreme scale, you should put much more work into whatever you actually do with the data.

Semantic Web people talk about the Semantic Web because it's what they do. They don't just build X, they create an ontology to build X, which somehow makes X better. And, most likely, slower.

For anyone interested in the alternative, check out RDFa[0]. They seem to be functionally equivalent and are supported by all major search providers. [1] (if down, see web archive version [2])

[0]: http://rdfa.info/

[1]: http://manu.sporny.org/2012/mythical-differences/

[2]: http://web.archive.org/web/20140719161748/http://manu.sporny...

To be clear, is this stating the format as outlined on Schema (http://schema.org/) is dead?

This article is saying the Microdata spec for semantic markup is dying (schema.org shows Microdata, RDFa and JSON-LD). RDFa/RDFa Lite is still active.

Having worked with both Microdata and RDFa formats, I'm glad Microdata is dying-RDFa Lite is much simpler to implement. The Microdata spec is unnecessarily verbose and, as the article said, RDFa Lite can do everything Microdata can do & more.

Hopefully having a single spec in the future can cut down the confusion and increase adoption of semantic markup.

"Google suggests using microdata" https://support.google.com/webmasters/answer/99170?hl=en&ref...

I have no horse in this race, but isn't Google the search engine kinda the deciding factor here?

Not necessarily. Google search is one application for metadata, definitely, but it is by no means the only one and nor is it the most interesting imho.

The same page also points out that microformats and RDFa are supported by Google too. I wonder whether the recommended status of microdata on that page is because it was proposed and edited at the W3C by a Google employee (that's pure speculation on my part btw).

Alternative site: http://archive.today/v9fQO

Site is down for me. Anyone have a mirror?

Down for me too.

Found this: http://web.archive.org/web/20140719161748/http://manu.sporny...

Thanks.

@pg do you think having automatic archived mirrors linked on all submission would be detrimental to linked articles ?

The slashdot[1] effect has been around for a long time.

Auto mirroring would be nice but at the same time it leaves a lot of open questions about content ownership, ability of the content owner to edit, monetize etc. The New York Times would not be happy if HN links were automatically mirrored and I imagine they'd be quick to protect their rights on that front.

[1] http://en.wikipedia.org/wiki/Slashdotted

If not auto-mirror then at least links to Google Cache and Webarchive in the interface would probably be welcome.

You could install Resurrect Pages https://addons.mozilla.org/en-US/firefox/addon/resurrect-pag... which gives you links to IA and Google caches with a right-click.

Even just a summary? We've got nothing. The site is down and even the archive.org link above isn't responding.

My guess just based on title is something along the lines of silos are bad so lets make a better silo. But, our silo will be a really nice silo. Forgetting that's what everyone has been doing forever. The wheel of IT rotates forever and this is a very old idea that fits the title.

Nope, it actually says that this silo is empty, let's go to that other one that's more popular and about to be standardized instead of making 2 standards.

http://pastebin.com/upiEHcJT

Reply on news.ycombinator.com