Re: embed RDFa --> embed coolness into Yahoo search results

Ian Hickson wrote:
> I think this illustrates a misunderstanding of how many people author 
> HTML, CSS, and so forth.

I know how people author HTML, CSS, etc.

As I've explained before, in my opinion, you have a double standard. You
expect that people will screw up RDFa, but somehow they'll get
SQL-in-the-browser right.

In fact, as we continue to improve the user-feedback mechanisms for RDFa
(SearchMonkey, bookmarklets, etc...), I think it's going to be a *lot*
easier to get RDFa right than to get SQL-in-the-browser right.

> The markup will get changed, just like Digg's did, just like the example 
> Yahoo wrote did, and in the process, if the language isn't something the 
> author understands, it will likely get broken.

Can't wait to see how that works out with SQL in the browser.

We have some more work to do on providing users with rapid feedback on
how their RDFa changes affect the output. We'll keep working on that.

> We need to invent languages that aren't brittle under copy and paste.

>From everything we've seen at Creative Commons, RDFa is fairly robust
under copy-and-paste. That's exactly why it was designed with prefix
declarations on any element.

[...]

> My point isn't that they made a mistake. My point is that the language 
> makes the mistake easy to make, and hard to spot.

No, the problem is that we don't have a complete toolset for feedback to
the user. Imagine writing HTML, even simple HTML, without having a
browser to render it. Now imagine <canvas>. The bar you're setting is
unreasonable even for basic HTML. One needs tool-based feedback to see
how markup is interpreted.

We're working on those tools.

As it turns out, I contacted Digg after your bug report (thanks for
that), and they already knew and were in the process of fixing it. How
did they know? Because SearchMonkey wasn't giving them the expected
result.... feedback = improved markup over time. We're getting there.

>> I wish you spent half as much time trying to understand the useful 
>> aspects of RDFa as you do trying to break it :)
> 
> I don't understand the distinction. How else can one evaluate a technology 
> than by trying to work out what is wrong with it?

You've claimed multiple times on the mailing list that you don't have
time to read our existing documents, our use cases, our mailing list
discussions on how decisions were made. But you found the time to write
up a number of deviously prepared examples and to mock the Yahoo
implementation on IRC. I don't think that's an attempt at objectively
analyzing a technology.

-Ben

Received on Sunday, 22 March 2009 20:24:04 UTC