Linked by Eugenia Loli on Mon 27th Mar 2006 19:30 UTC
Google Phil Sim, a professional with technology editor journalist background, has written three interesting blog posts recently, discussing the much-rumored Google OS (1, 2, 3). He speculates that all user's data will be stored online on Google's servers and so one's desktop and files can be retrieved exactly as left by any other PC station, anywhere in the world, by simply using his Gmail credentials. It's like having your OS on a usb key with you at all times, only, without the usb key...
Thread beginning with comment 108846
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE: No Centralization
by afilloon on Tue 28th Mar 2006 18:28 UTC in reply to "No Centralization"
afilloon
Member since:
2006-03-28

"This goes against one of the most basic tenets of web architecture is decentralization."

The internet is based on decentralization of routing and infrastructure, not data.

"Better than a network OS would be to equip client operating systems with software agents that could act as your personal system admin. If you want access to your data from anywhere, you access it from your server, not some central Google or Microsoft server..."

Huh? This is still located on a centralized server. Are you advocating everyone own their own server and manage their own data? Just how would that work?

Reply Parent Score: 1

RE[2]: No Centralization
by alucinor on Tue 28th Mar 2006 19:08 in reply to "RE: No Centralization"
alucinor Member since:
2006-01-06

I was talking about the Web, not the Internet. And one of the architectural principles is in fact decentralization of data. Perhaps this idea wasn't emphasized in the WWW as we know it today, but it's central to the Semantic Web, as it is core to RDF that anyone can say anything about anything, meaning that Semantic Web agents have to know how to deal with potentially contradictory information.

I think in the future, personal data should be kept decentralized, and the only centralization should be something similar to DNS, so that if anyone needs to access my information that I choose (or my government forces me to) publish, then they can do it at Person://us.washington.seattle.smith.john or something similar, and get it in a vendor-agnostic format. This way, it takes advantage of the centralization afforded by DNS, but the actual location of the data could be on any server, including my own personal one, or just some space I'm renting from my ISP, whatever. I just don't want to see my data, which I rightfully own, under the ownership of an external entity.

To see how centralization is fragile, just research the attempts by the British NHS to centralize hospital data into the national Spine! Centralization is a weakness to the Web.

Hopefully a company or just a group of smart engineers will come along who will do for the server side and ownership of data what the PC movement did for the client side and ownership of programs (and I'm talking about the "golden age" of the PC movement, not the DRM crap being discussed today). As our culture becomes more wired, our data is going to become more valuable to us, and consumers will buy a software product that empowers them to easily store and manage their own information. I can't give you details on how such an automated administrator would work; it would probably be some kind of expert-rules-based system. But the potential for it is there. The pedulum of the software business is swinging back towards the network with SOA, but that doesn't mean we couldn't see swing back the other direction again with shrink-wrapped products one day.

Edited 2006-03-28 19:24

Reply Parent Score: 1

RE[3]: No Centralization
by afilloon on Tue 28th Mar 2006 22:41 in reply to "RE[2]: No Centralization"
afilloon Member since:
2006-03-28

As far as I know the changes to RDF, in regard to W3C Semantics, was first formally recommended in 2004 and the first working drafts were just published this month. This is hardly the core web yet. I donít agree that ďÖone of the architectural principles is in fact decentralization of data.Ē The core architecture of the Web protocols (which in fact is all the web is comprised of) is the linking and presentation of data and nothing more. Everything else is heaped on top.
What you are espousing is philosophy, and there is nothing wrong with that; just donít represent technical architecture as a philosophical foundation.
One of the inherent problems with decentralization of data is maintaining that data and accessibility. Your argument that we should each be responsible for our own personal data is an interesting though experiment, but where do you draw the line between your personal data and the data about you that companies need to keep doing business with you? How do you control the aggregation of that data? On a more mundane note, what happens when the data on rented server space is lost to a crash and the ISP doesnítí have a good backup? Are you going to be responsible to maintain and restore that data? Do you seriously believe that every citizen out there will?Iím skeptical of egalitarian, utopian plans to give everyone control over all of their personal data. Itís not that I believe that ďbig brotherĒ should control it, I just havenít seen a plausible plan to provide a workable solution to the general populace that would protect that data, both from a physical and from a privacy standpoint.

I look forward to hearing your thoughts on this.

Reply Parent Score: 1