Web 2.0: Data, Metadata and Interface

One key takeaway from the Web 2.0 panel was that data, interface and metadata no longer need to go hand in hand. When working on an application/website, one thinks of the overall picture including the data, the metadata, and the interface. With Web 2.0 apps, the data might be from one place, the metadata from another, and the interface from a third party or a remix. The diagram below shows the move towards Web 2.0 along with examples.

the move towards Web 2.0

Amazon, in its early days, owned the data (the catalogs), the metadata (the associated keywords, where in the structure the item fit in), and the interface. Amazon has gradually evolved to the point where most of the metadata (e.g., ratings/reviews) comes from its users, developers can use API’s to build their own interfaces, and it also shows data from other vendors (e.g., the second hand sellers).

Flickr is the quintessential Web 2.0 application. Its data and metadata is contributed by its users; while the interface is its own. Its API’s are used by developers who tend to use its data, but not the interface (such as Mappr, Color Pickr).

HousingMaps marries data from two streams (Google Maps and Craigslist). Its interface is mostly that of Google Maps, but it adds in the housing information on top of the Google Maps interface.

Del.icio.us does not have any data of its own. It exists as a metadata aggregator, for data on various sites, tagged by users.

This has many implications, which I will talk about in a follow-up post.

Panel Roundup and Podcast
For those interested, the podcast (audio file: 82.8MB, 02:00:27)of the event is now available.

Several people posted pictures of the event. Watch the BayCHI tag on Flickr. Interestingly, BayCHI has an official “No pictures except by BayCHI” policy – a policy that in increasingly hard to work with in these days of camera phones and Flickr. (Yes, we are rethinking that policy).

Bill Scott wrote a synopsis of the event. Ashley Richards wrote about the internet lost in translation. Bokardo comments on Jon’s post on paying for access to remixing data.

PeterMe noted that the panel was too techy – too much talk about API’s. (Comment: No doubt, the word API has never been uttered at BayCHI so many times! But interestingly the focus on API’s was at the behest of the audience – during the question-answer session. In my opinion, the logic (philosophy if you will) of Web 2.0 reflects its technological underpinnings. A good example is the open source movement. Now, we even have open source beer. But initially, to understand the philosophy of open-source you had to understand developer speak. As Stewart Butterfield noted at the panel, Flickr wanted rich interactivity (refreshing parts of the page at a time) so they had API hooks – they kind of went with it, rather than fighting it. The API’s facilitated the openness. Currently, the logic behind Web 2.0 is baked into API’s, RSS etc. Also, I question whether any business will move to this approach because it is a compelling philosophy. They will shift because it is an attractive business proposition, or because technically it makes sense/is unavoidable, or a mix of both.)

8 thoughts on “Web 2.0: Data, Metadata and Interface

  1. Some other companies examined in the same manner:

    Odeo leverages data produced by others, adds in an an interface and metadata.

    Craigslist takes people’s data, adds metadata (fits it into a hierarchy; note that Craig does not feel a need to move away from the hierarchical model); and adan interface for it. In turn services like HOusingMaps use Craigslist feeds and remix with other data streams (such as Google Maps). Watch out for more remixing of Craigslist data.

  2. I was there at the panel. It ruled … I really don’t know how you would talk about Web 2.0 without mentioning API’s.

    All this talk about openness is a little hypocritical: it’s unclear how open the proponents of Web 2.0 themselves are. Everyone’s real open ’till the time its developers building stuff not for profit. The moment there is commerical usage, its back to ye old business table. No one is even publishing some sort of rates for commerical usage. So these topics do need to be addressed from a business point of view as well.

  3. In some ways, I see your arrow diagram as exactly backwards. Historically, APIs have been used by Microsoft and many others as tools to lock-in developers to particular platforms or ways of doing things. Andale’s integration with, for example, eBay is classic system integration: they purchased liscensed access to ebay’s data and are using it in exactly the way eBay permits.. There’s absolutely nothing “open” or “reuse/remix” about it. From that point of view, Flickr’s no more “open” than Microsoft: a Flickr color picker application or creative visualization tool depends on Flickr’s API, period. “Using data in new contexts or interfaces” is not itself a new phenomenon or unique characteristic of Web 2.0. It’s certainly not unique to the web.

    When you say “Flickr wanted rich interactivity…so they had API hooks. The API’s facilitated the openness”, you confuse one kind of “openness”–straightforward access to data via APIs–with another–an attitude towards sharing and community.

    That second “openness” is about what you’re allowed to do with data you’ve put into an application. Flickr allows you to get your pictures out, eBay will never let you delete pricing history for your sales.

    And I’d dispute what you say about del.icio.us: they most certainly do have their own data. They have huge amounts of information about preference, popularity, and relevance of sites.

  4. Andrew,

    Web 2.0 is associated with the phrase “Web as platform”. What do platform vendors do? Platform vendors try to get you hooked on their APIs, so that you will be locked in. As you point out, software companies have done this for years. But web sites have only now started doing this. That is the difference. When web sites start acting like software companies, then you have your Web 2.0.

    You say systems integration, I say remixing … ;->

  5. Andrew,

    As a note, the diagram does not attempt to capture all aspects of the move towards Web 2.0. Its my way of understanding the relative independence between data, metadata and interface and the move towards deeper interactions between sites.

    Interesting point about Andale. One thing the panel highlighted: by the time we are talking about commerical usage of API’s, whether its for the eBay API, or Technorati, its all worked out at the business table. This was something that emerged at the panel – that the business issues are far from worked out. Till the time you want to use the data for noncommercial use, its ok. But the moment, its commercial usage, its back to the business table. I heard from at least one person who gave up the idea of taking a remixed app a business once they understood this aspect.

    As regards, my note about Flickr. I am mostly reflecting what Stewart said. When I initially asked him to be on the panel, and asked him to talk about the Flickr open API’s (among other issues) – he brought up interactivity, AJAX, refreshing parts of the page. I was myself surprised he made the link. But he reiterated it at the panel. I directly quote from him below: (this is why I love podcasts!) (between minutes 11-15 of the podcast, right at the beginning of his talk, he started with this).

    My take: If you talk to technology people, and discuss Web 2.0 – they will talk API’s. It does not mean its not about openness / attitude / philosophy. Thats the language they speak.

    From Stewarts’s talk (podcast of talk:

    “…The next step after that, and I think this is one of the defining characteristics of Web 2.0, is that we use http to address little bits of information (not whole pages) and render them in the browser.
    … So there’s a movement from just rendering a document in the browser … to calling lots of tiny little applications that would render parts of a page. One of the interesting things about this kind of UI (which has been popularized under the AJAX moniker): once you are committed to doing it (and it doesn’t really matter if it’s AJAX in the sense people normally talk about it, which is javascript driven, or something like this, which is Flash) … you’ve automatically made a public API. Just because of the nature of the web, the code is public, the traffic over the web server is public, what your client is calling. If you’re geeky enough to run an application to look at your tcp/ip traffic you can see exactly what is going on, and you can build on top of it. So lots of applications that didn’t publish APIs (google maps is a good example, at least initially) had open APIs that the public was able to remix. So you can take that it a step further: If you’re going to be having the API anyway, publish it, document it, and let people use it.”

  6. On the issue of using free APIs or even commercial ones.

    Suppose I want to set up a (hypothetical) site to do the following:

    1. Download job listing from CraigsList and put them in my db

    2. When a user comes to my site and enters keywords in the search box, a window opens to display the CraigsList listing

    3. If the user then clicks on that listing, he lands on CraigsList where the full listing is displayed.

    Consider two scenarios:

    1. I don’t make any money –I’m just doing this for fun

    2. I put a Google AdSense link hoping someone will click

    Question: The answer to #1 is obvious. But does #2 make “commercial use” of CraigsList’s data? or for that matter anyone’s data?

Comments are closed.