Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.

Personal Operating Systems and a Subject-oriented Web

by ChrisCook Fri Mar 29th, 2013 at 11:55:11 AM EST

Saving data donkeys in quicksand is an interesting BBC article in relation to the phenomenon of data tagging.

A friend of mine worked out the importance of generic subjective tagging for messaging (of all types) about 10 years ago, but could never engage with anyone to develop the resulting applications he also developed. Among other things, these would essentially finish off what's left of the existing business model of advertisers, and in turn mean that the likes of Google and Facebook would have to move on from their existing business models.

But because of the voracious and ruthless nature of the corporate players involved and the pernicious regime of IP rights and law, the concepts were not implemented, although the elements which he had far-sightedly analysed are now beginning to emerge.

From that perspective and experience, I think that where tagging will lead is to a simple 'personal operating system' resident on personal devices, and which will connect - with the minimum of complex code - directly to decentralised/distributed data-bases.

The only central assets would be servers which resolve:

(a) basic personal identity to a market/enterprise/group identity ; and

(b) machine identity to market identity ie what I have for years termed a 'Dot Market' model with a market-specific domain such as Dot Oil, or Dot Gas.

This raises the Big Brother issue of who can be trusted with such servers, and maybe the former might be domiciled in (say) Iceland, and the latter in (say) Switzerland.

Individuals only have one basic personal ID, but they may potentially have thousands of market-specific, enterprise-specific and group-specific IDs.

Their basic personal ID is only physically located in one place at one time, which adds in the possibility of generic geo-authentication of transactions - ie mapping mobile device locations to static machine locations.  This brings another Big Brother issue of who can be trusted with that data.

Attempting to create a Semantic Web on an object-oriented machine-centric basis has always seemed to me to be a dead end, where increasingly sophisticated algorithms have attempted to derive meaning from data objects by reference to experience. But tagging is different and subjective or subject-oriented, because only you know what you mean, and will tag using language as you see fit.

I believe that we are in a transition from a complex, centralised, fragile, machine-centric Web 2.0 to a simple, decentralised, resilient, people-centric Web 3.0.


You can't be me, I'm taken
by Sven Triloqvist on Fri Mar 29th, 2013 at 03:39:08 PM EST
I have documented here some of the problems a failure to understand a more subject-orientated approach has created for the Irishtimes.com news website.

Index of Frank's Diaries
by Frank Schnittger (mail Frankschnittger at hot male dotty communists) on Fri Mar 29th, 2013 at 06:57:00 PM EST
Great post

"The future is already here -- it's just not very evenly distributed" William Gibson
by ChrisCook (cojockathotmaildotcom) on Fri Mar 29th, 2013 at 07:29:47 PM EST
[ Parent ]
Cutting a long post very short: tagging doesn't work outside of trivial applications.  Tags have a 1:Many Semiotic/Semantic relationship, e.g., the word "Set" has 424 definitions, and without being able to determine what the word "points-at," its Final Interpretant, there's no hope of intelligently digging through the hundreds (thousands?) of exabytes of text data out there.  

Google is a good example.  The basis of their search engine queries is keywords (aka Tags) plus the Informal Logical Fallacy of argumentum ad populum plus some epicycles and dongles which I concede allows it to "work" in a half-assed kinda way.  With text data growing at 75,000,000 books-worth (estimate) of text data a year it's a real good question how long it can continue to "work."  Google's quality of query response has been steadily declining since 2008.  

She believed in nothing; only her skepticism kept her from being an atheist. -- Jean-Paul Sartre

by ATinNM on Sat Mar 30th, 2013 at 02:36:20 PM EST
Seems to me that there is already a natural and automatic tagging mechanism in play in most situations: the progression of time. Junk piles up on your desk, and when looking for something in particular, you associate it with a time and then dig in the pile at the appropriate depth. If you don't try to organize your directories and folders, the easiest way to find stuff is to sort it by date. In Facebook, your timeline automatically ages content. Twitter perhaps the most obvious example.
by asdf on Sat Mar 30th, 2013 at 05:54:35 PM EST
[ Parent ]
Knowing the time stamp is useful for locating where the data landed in Past-Present-Future.  Doesn't help for extracting data out of a database.  

She believed in nothing; only her skepticism kept her from being an atheist. -- Jean-Paul Sartre
by ATinNM on Sat Mar 30th, 2013 at 10:14:09 PM EST
[ Parent ]
Nope, it doesn't help a database search. Automatic tagging would work--if there were a way to do it. It would require some magic formula that provided more information than what is already commonly in the records.

If users have to enter metadata for every record, tagging is hopeless. Just getting them to use helpful file names is almost impossible...

by asdf on Sun Mar 31st, 2013 at 12:27:02 AM EST
[ Parent ]
Tagging is hopeless but it keeps coming back.

Like malaria.  

She believed in nothing; only her skepticism kept her from being an atheist. -- Jean-Paul Sartre

by ATinNM on Sun Mar 31st, 2013 at 12:48:37 AM EST
[ Parent ]

Go to: [ European Tribune Homepage : Top of page : Top of comments ]