I think the distinction between data and HTML grows less and less everyday.
Trying to impose a lot of superficial structure onto data that relates to
itself in uncountable linguistic and metaphysical ways limits the
perspective of the people making the systems, as well as those using the end
product. And in the end, if Google shows me a webpage with the graph I want,
and Wolfram shows me a custom graph using that data, it's all the same to
me.

Also, I think the use of collaborative expertise is great, but it's most
powerful when used like Wikipedia uses it - with very little structure
imposed on the article pages. That way, the pages maintain the ability to
change as the people working on them gain better understanding. Wikipedia is
sustainable - it gets bigger and better each year it seems.

Now, compare that to Wolfram Alpha. Hundreds of experts sifting through
data, trying to stuff it inside Wolfram's UI. I imagine this system will
suffer the same rot that similar systems face. Unlike Google, Wolfram's data
1) isn't being refreshed by constant analysis of all human dialogue and 2)
probably can't change/adapt much over time.

It seems much better to have an adaptive computer system analyze the
internet and create data, than to have a bunch of wizards in white robes
compiling spellbooks of questionable worth... even if you can access them
all through one search box.

It's like the cool kids are all making search engines, and Wolfram is trying
to revive the encyclopedia industry.

Maybe I should hush up until it comes out though.

Andrew
Co-founder, TrailBehind.com




On Wed, May 6, 2009 at 4:36 PM, R E Sieber <[email protected]> wrote:

> I work with humanities people and No one wants to create metadata. RDF,
> they don't understand. Believe me, I've tried (triplets, why is that so
> hard?). If they do, they don't understand (a) the rigidity imposed by the
> structuration and (b) the level of abstraction needed to describe the
> domain. That's why some kind of structured, hierarchical metadata-on-the-fly
> from crowdsourcing is so interesting. What's needed is structures (data
> gathering plus hierarchy plus reputation system) put in place where domain
> experts and other interested parties can check the results. By results, I
> mean not only getting the concepts (most likely a controlled vocabulary)
> right but also the relationships among concepts right.
>
> Renee
>
>
> SteveC wrote:
>
>>
>> On 6 May 2009, at 15:56, Sean Gillies wrote:
>>
>>> On May 6, 2009, at 4:41 PM, P Kishor wrote:
>>>
>>>  On Wed, May 6, 2009 at 5:18 PM, SteveC <[email protected]> wrote:
>>>>
>>>>>
>>>>> On 4 May 2009, at 08:44, Ed Parsons wrote:
>>>>>
>>>>>  The general scepticism here I think is well placed, semantic based
>>>>>> systems
>>>>>> always demo well, the key to more widespread adoption is the
>>>>>> automation of
>>>>>> the still largely manual creation of ontological relationships. But
>>>>>> one day
>>>>>> I'm sure this will work, after-all TBL is usually right.
>>>>>>
>>>>>
>>>>> Come on, nobody believes the web of data stuff anymore surely. If they
>>>>> did
>>>>> someone would do something like RDF but actually usable and easily
>>>>> implementable in actual HTML.
>>>>>
>>>>>
>>>> That would be RDFa, or GRDDL.
>>>>
>>>
>> Yeah so I just looked through the wikipedia pages on both, they feel a bit
>> unloved. They don't look like a whole lot of fun, and the chicken and egg of
>> no data to crawl and no crawler for the data must be frustrating.
>>
>> Has anyone spent any time on what would make people actually mark this
>> stuff up?
>>
>>  It seems HTML5 has told RDFa to FOAD.
>>>
>>
>> Nice.
>>
>> Best
>>
>> Steve
>>
>>
>> _______________________________________________
>> Geowanking mailing list
>> [email protected]
>> http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
>>
>>
>
> _______________________________________________
> Geowanking mailing list
> [email protected]
> http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
>
_______________________________________________
Geowanking mailing list
[email protected]
http://geowanking.org/mailman/listinfo/geowanking_geowanking.org

Reply via email to