Nova Spivack: Making Sense of the Semantic Web

The following video was taken of Nova Spivack at The Next Web 2008 Conference held April 3 and 4 in Amsterdam, The Netherlands. Other commentators have blogged on Nova's slideshow presentation given at the conference. See, e.g., Is Keyword Search About To Hit Its Breaking Point? (25 April 2008) by Erick Schonfeld, and Nova Spivack: "The Semantic Web as an open and less evil web" (3 April 2008) by Anne Helmond. I, myself, recently blogged on a statement of prior art filed by Nova Spivack and Kristinn Thorisson in a US patent application that I called Categorizing the Internet's Serious Problems (11 June 2008). For many of you, this latter blog may be a good starting point for what is to follow.

Because the video only became available in early June, I have been provided with an opportunity to blog on both the slideshow and the video at the same time. I have also included reference to substantial quoted text and, furthermore, you will find here the inclusion of two 'missing' slides. However, I would like to add that I am not blogging this as an opinion piece but as an addition to this blog's Reference Library. I'll blog my opinions in later entries.

I have stacked the video (42m07s) and accompanying slideshow (44 slides) one on top of the other, below, for those of you who have the time to listen and watch the video, and may want to jump back and forth between the two. The video irregularly shows where Nova is in his slide presentation, and often a slide to which he is referring is difficult to read or only partially shown. So having the slideshow available for quick reference should be useful.

I have included a substantial transcription of Nova's presentation beginning at Slide 17 (the 15m mark) through to the end of the 42m video and finishing with Slide 31. I have paid special attention to the time frame covering Slides 17 - 31 because Nova is then particularly focused on explaining the emerging mechanics and standards of the Semantic Web. Nova is a fast, fast talker (literally). I have taken care to accurately qoute Nova but if you find that I have mis-quoted him, please let me know.

Be aware that Nova does not orally present slides 32 through 44 which cover his product, Twine.

In addition to the embedded slideshow, below, I have obtained an updated version (dated 19 May 2008) of the same slideshow from Nova himself (Thanks, Nova!). I'm glad he shared it with me because there are two slides presented in the video that do not appear in the embedded slide show but do appear in the updated slideshow version. These 'missing' slides are presented by Nova after the 15m mark and so I have separately embedded them further down the page of this blog with accompanying quoted text. See the slides below entitled The Growing Linked Data Universe and The Growing Semantic Web, respectively. The updated version of the slideshow is in PPT format. You can download it from among the source references to this blog.

So, again, I have stacked the embedded video and slideshow below. 

Nova Spivack at The Next Web Conference 2008 from Boris Veldhuijzen van Zanten on Vimeo. | View | Upload your own

 
Below you will find my transcription (i.e., numerous quotes from Nova's oral presentation) beginning at 14m56s, the missing two slides, and a few of my bracketed editorial comments. All quoted language is that of Nova Spivack's unless otherwise noted. All slide titles are emboldened and italicized.

14:56. Slide 17: Two Paths to Adding Semantics.

14:56. "There are two paths [to the Semantic Web]. One is called the bottom up approach, or classic approach, and that is where everybody is going to go learn RDF and OWL and manually create all of this semantic web metadata. It's not going to happen. It's really, really too hard ... The top down approach ... is to do this automatically [so that] the RDF and OWL code [is] embedded into the data - into the content - automatically [which] turns out to be more practical."

15:35. Slide 18: In Practice: Hybrid Approach Works Best.

15:55. Slide 5: The Higher Resolution Web.

[Nova jumps back to this slide in the embedded presentation. This slide was orally skipped over in the initial part of his presentation. In the updated version of the presentation (see above) this slide is numbered 16.]

15:55. "What we are doing is creating a higher resolution web. So it's like digital photography ... The semantic web is like saying we are going to give you ten times the number of megapixels in your data. So we are going to make a higher resolution web because each piece of data is actually going to carry more meaning ...."

16:47. "The way that Google sees the world is very flat. There are just basically pages and links. And that's all it really knows. In a semantic graph you know what types of pages or what types of things these data records are, and what types of links the connections really mean. Its a richer web meaning and we call this a semantic graph.

17:17. Slide 6: The Web Is The Database.

[Nova jumps back to this slide in the embedded presentation. This slide was orally skipped over in the initial part of his presentation. In the updated version of the presentation (see above) this slide is numbered 17.]

17:17. "Using the open standards the web becomes a database ... The [Semantic] Web is the database."

18:08. Slide 19: Smart Data.

18:12. "So we call [the data within the semantic web] 'smart data'. So smart data is data that carries what is needed to make sense of it .... So that you don't have to refer to some other application .... All that you need to understand to use that data is carried by the data itself ... You can make really, really dumb software but yet it can do really smart things ... because the intelligence is in the data."

18:53. "Ultimately we might have this piece of general use software that when you point it at semantic web data about health, suddenly it can give you medical advice. But then if you point it at semantic web data about data about the stock market, then it can give you investment advice .... That is the dream of the semantic web. All human knowledge will be on the web in a machine understandable fashion and then all software will be able to use all of this knowledge."

20:26. Slide 20: The Semantic Web Is A Key Enabler.

20:30. "Another concept is ... just in time data. The semantic web, because the data is self-describing enables and application to pull in data it's never seen before and use it right away. The challenge in the old-fashioned way of doing this is "what's the schema?" ... With the semantic web we do away with all of that annoying communication and simply the [smart] data tells you all of that. So it does that using an ontology. An ontology is like a schema. Basically, the data points to a document that describes its structure and the rules for using the data. And so when you see the data you go to the ontology [that is] written in the same language as the data and you can actually use that to make sense of the data without having ... to go read a schema document and then type that into the program.

21:45. Slide 21: The Semantic Web = Open Database Layer For The Web.

21:45. "[Another] way of thinking about the semantic web is that it is an open database layer for the web. So if we are making the web ultimately into an operating system ... [we can call the] Web OS ... it's going to have a file system ... and I think the semantic web is a candidate for that file system.... You should be able to write an application, point it at the web, get data, publish data and not care where it is ... You should be able to do this the way you do [this] on a desktop computer when you write an application ....

22:26. The semantic web standards provide a way of representing the schemas with ontologies, ways of representing rules, the data itself can be represented, mappings that say that this is the same as that can be represented, and there is also a query interface for doing searches in an open way across this data. So this is really a stack that creates a database layer for the web at large.

22:44. "So, this is really a stack that creates a database layer for the web at large. [Nova reveals Slide 22: Semantic Web Open Standards]. And there are several [semantic web] standards that are important. RDF is the main standard and that is really the way the data is represented with things called that are called 'triples' .... OWL is built on RDF ... [and] is just RDF with some more statements in it ... some more expressive power for defining schemas. SPARKL is a query language ... like SQL [but] for RDF. There's a rules language called SWRL ... that hasn't been standardized yet but there is a lot of talk now around the rules. And GRDDL which is for transforming data so you can say here is how to take this XML data and turn it into RDF on the fly. And you can make these GRDDL profiles for websites that enable anybody who wants to see [a] website in RDF to get the RDF [enabled version of the website] immediately.

23:56. Slide 23: RDF "Triples".

23:41. "Let's talk about how the data is represented .... The basic unit of data in the semantic web is called a 'triple' and that's because it has three parts. It has a subject, a predicate and an object. So, for example, "Susan works for IBM" ... [where] Susan - who is actually a URI that represents a data record that describes Susan - works for - which has a URI that defines what you mean by 'works for' somewhere in an ontology - IBM - which has a URI that points to a representation of IBM. Now these three things could be in different [databases] ... So, it's [like] a giant mashup on a very, very atomic level of data.”

25:10. Slide 24: Semantic Web Is Self-Describing Linked Data.

[Comment: Slide 24 represents a picture of a Data Record with an ID and fields connected in one direction to ontological definitions in another direction to other similarly constructed data records with there own fields connected in one direction to ontological definitions, etc. These data records - or semantic web data - are nothing less than self-describing, structured data objects that are atomicly (i.e., granularly) connected by URIs.]

25:14. "This [Slide 24] is illustrating how these data records are all connected together whether it is within one application or across applications. It becomes an open database just like the web but for data .... Is there a better term than semantic web? Yeah, it's data web. That's a better term."

25:35. Slide 25: RDBMS vs. Triplestores.

25:35. "The traditional way of storing data in a relational database would be [by] using [tables] and tables are annoying because they are not really the way we think .... You have to do all of these little tricks to make [data] point to other [data]. Now in the semantic version ... you just make a big list. [You create a] list of triples, each triple is a statement and has URI's in it. And so there's a challenge here that these lists of triples get really, really long. You could easily have a billion, or ten billion, or a hundred billion rows in one of these lists. And if you ... stick [such a] list into a relational database ... you get really bad performance because relational databases were not designed for data that has [billions] of rows and not many columns. Relational databases were [designed] for lots of columns and not as many rows. The optimization in the relational world was for a different shape of database."

26:28. "So to solve this we've created things called triple stores. These are new kinds of databases that are designed for these lists. These lists have a lot of benefits over the relational model. They're much easier to maintain, and they can actually live on top of relational database ...."

26:44. Slide 26: Merging Databases In RDF Is Easy.

26:44. "So one of the nice things about [the lists in the triple stores] is that merging data is extremely easy .... You don't have to do any fancy [relational] database refactoring .... [It is easier to use triple stores because] the way the data is integrated is through URIs.... If you have a URI for IBM [in one data record] and you have [the same] URI for IBM [in another data record], now we know they are referring to the same [data record for IBM]. And so the matching is done at the URI level rather than having [a] human sitting there and having to refactor the database. [Using URIs rather than humans] scales a lot better to the web ...."

27:41. “Now there is this universe of linked data that is emerging .... and there are a number of different ontologies that cover different domains ... [see The Growing Linked Data Universe slide at 27:53, below] .... So there are a bunch of different ontologies and applications that all connect to each other and are sharing data in this growing web of connected data.”

27:53. Missing Slide: The Growing Linked Data Universe.

 

28:34. Missing Slide: The Growing Semantic Web.

 

28:54. "You can see [referring to The Growing Semantic Web slide] a lot of activity around consumers right now and online services for developers and consumers. The applications side is starting to emerge in the enterprise space and that's kind of how it looks today."

29:04. “So where are we and where [is the Semantic Web] going .... [Nova reveals Slide 31: Future Outlook] Right now we are still in the early adoption period of this technology but there's a tremendous amount of momentum and a lot of adoption taking place among developers and also some early applications .... So I believe that this period of 2007 to 2009 is really the first wave .... [and during this period there will come to be a couple of million users or more of the semantic web] and then it will [be considered as] mainstream. So when we get into Web 3.0 [in] 2010 that's when real mainstream adoption happens. That's why I believe that semantics will be baked into a lot of mainstream applications from companies whether it is Google or Microsoft. Adobe already does it. Yahoo! already does it."

30:28. "Where the semantic web and data portability [project] meet is that the semantic web provides some open standards for making your data even more portable."

30:40. "[In conclusion the semantic web will] do for data what the web did for documents .... It's very hard to do this today because [while the standards are there in many respects] the tools are not [yet developed] and so if you have a company and you are thinking [that you want] to use the semantic web, it isn't easy. The place to start is [with] a few simple standards. One is called FOAF, it's friend of a friend, and that is for describing user profiles. Another one is SIOC, that is for sharing data about forums, discussions, [and] user accounts .... As you get deeper into the technology, and as more API's come out, and more services become open ... it will get much easier."

[Q & A]

32:48. "My personal opinion is that the semantic web does not introduce any new business models. I think that it just makes the existing business models better...."

38:06. Attendee's Question: "As far as I understood the semantic web it depends on definitions of different [things] .... If I understand correctly these definitions are done by people .... but before you also mentioned that people are really inconsistent so how do you [reconcile this]? Also how additionally do you handle the [definitions] of fast changing things?"

Nova: "Good questions. So there's a big misconception which is that the Semantic Web demands some kind of agreement. And you sometimes see that when people criticize the semantic web they say, "Well nobody's ever going to agree on a definition of ... all these things". In fact the Semantic Web was designed for disagreement. So, anyone can make their own ontology [and therefore] describe the world however they want. So that's a good thing but then it creates this problem that there might be many definitions. You know, here's my way of describing a car. Now over here this is how Toyota, Mercedes Benz, BMW describes a car. And it's different. So one of the things that they built into the standards was a way to map definitions to each other. So you can say [that] this definition is equivalent to that definition. You can also say [that] this piece of data is the same as [that] piece of data. So, anybody can make those mappings. Not just the people who made the data but anybody can make those mappings. And so in a community-driven, bottom up process, when mappings are created you can then begin to infer the equivalence or connections and in fact there's a lot of research going on that just from a few mappings you can make inferences that connect a lot of things together. So I think we will see something like the Wikipedia where lots of different definitions are composed and the winners will be the ones who have the services, who have the content, that uses that definition ... that get the most users ...."

[end]

Categorizing the Internet's Serious Problems

The following quoted text is taken from the statement of prior art in US patent application entitled Methods and systems for managing entities in a computing device using semantic objects filed in 2003 by Nova Spivack and Kristinn Thorisson, and assigned to Radar Networks, Inc., the provider of Twine. I've also blogged about this patent application in US Patent App 20040158455:  Methods and systems for managing entities in a computing device using semantic objects (Radar Networks) Φ. And I previously blogged about Spivack and Twine in The Funding of the Emerging Semantic Web.

This statement of prior art categorizes the Internet's serious problems into the following categories:

1. Information overload
2. Information complexity
3. Dis-integration
4. Spam
5. Lack of targeting
6. Lack of personalization
7. Lack of privacy control
8. Information deficit

Here's the quoted text:

• "Knowledge workers, teams and organizations routinely work with a large and complex array of information. This includes e-mail messages, instant messages, chats, discussion postings, calendars, contact and to-do lists, documents, photos, maps, and database records. This information also includes tacit knowledge and expertise that resides only in people's heads. The average knowledge worker interacts with several dozen information types, hundreds of Web sites, and dozens of different applications. Existing information systems are focused mainly on data, rather than on relationships between data. There is a growing need to enable applications and users to see how various types of information are related across different information systems and locations. However, there is no tool for connecting, managing and sharing this information in a unified way.
  
  The growth of the Internet, as well as the increasing amount of information it contains, are leading to serious problems for many computer users. In particular, they are leading to a problem referred to as "information overload" in which parties are overwhelmed by more information than they can effectively process, navigate, search, track, respond to, utilize, cope with, or manage given limited time and resources.
  
  A related problem is "information complexity" in which, due to the sheer volume of information choices on the Internet, and its disassociated nature, is making it overly difficult to locate particular desired information when it is needed. Another related problem is "dis-integration" that arises due to incompatible or nonstandard information and services, which leads to software and service incompatibilities, as well as obstacles to processing and managing information effectively. Another problem is "spam" that arises when Internet participants receive unsolicited, unwanted, or irrelevant information from other parties on the Internet. An additional problem that is related to spam is "lack of targeting" which arises because information providers such as publishers, advertisers, and marketers are unable to effectively distribute their information to appropriate, interested parties, due to lack of information about the interests and policies of those parties.
  
  Another related problem that is also related to spam is called "lack of personalization" which arises when parties on the Internet are unable to effectively subscribe to, filter or control the information they get from others. Another problem is "lack of privacy control" which results because Internet participants are unable to effectively control what information about themselves is shared with or by other parties on the Internet. Yet another drawback is "information deficit" that results when parties are unable to find, or do not receive, the information they need or are relevant to, even though it is available somewhere on the Internet or even on their own computers.
  
  These problems, and related problems, are becoming serious obstacles to knowledge work, commerce, collaboration, publishing, marketing, advertising, search, communications and communities. In particular these problems are reducing the productivity of Internet participants. Parties must spend increasing amounts of time and resources searching for information they seek, trying to ensure that they receive information they want from others, trying to block or delete unwanted information received from others, responding to information they receive from others, managing and organizing information they want, tracking changes to information of interest to them, trying to distribute relevant information to others appropriately and trying not to mistakenly distribute unwanted or irrelevant information to others. With the expanding and pervasive use of the Internet and its increasingly central role in relationships, interactions and transactions of all kinds, those entities that provide content and/or Internet software tools and services are searching for and implementing ways to solve the above problems. However, attempts to solve these problems face numerous obstacles. Presently the Internet is comprised of many separate infrastructures and software tools that are used for different modes of communication. For example, e-mail communication takes place via e-mail servers and client software applications that communicate via specialized e-mail messaging protocols, yet Web searching for example takes place using search engines and databases that are accessed via Web browser software and Web transaction protocols. Thus, even if one were to solve the problem of information overload for e-mail it would not necessarily solve this same problem for Web searching.
  
  A principal problem stems from present systems' inability to store, route and use meta-data about the data resources that they manipulate. It is therefore a goal of the present invention to provide a comprehensive solution to these limitations, in the areas of information overload, search, sharing, collaboration, communication, transactions, knowledge management, information distribution, and automated and manual manipulation of computer-stored data and resources, allowing information to be connected in meaningful ways.
  
  Using traditional search systems, parties seeking something enter queries that are tested against databases of information that are provided by one or more parties with things to offer. If matches are found, the seekers are notified with links to the appropriate provider. One problem with such systems, however, is that they do not work in reverse; there is no way for providers to locate seekers who want what they offer. Instead, providers must wait passively to be found by seekers. Seekers on the other hand, must do all the work. Another problem is that it offers only search by keyword; there are no mechanisms that support higher-level organization of the information.
  
  Providers who want to be found may resort to marketing in order to reach seekers. For example, many search engines provide an option to buy keyword advertising, enabling providers to market what they offer to seekers who enter relevant queries. Although they do this, they do not enable providers to search for seekers who want what they offer, nor do they help them locate seekers who are not presently searching but are still interested. Thus providers must use external marketing channels such as direct email, banner advertising, paper-based direct mail and other forms of advertising to locate interested seekers. These inefficiencies result in increased transaction costs for seekers and for providers.
  
  The present invention provides a single universal underlying infrastructure for managing information overload, distributing, locating and filtering information between information providers and recipients that works equally well across all types of Internet relationships, interactions and transactions. This single solution can be used to route and filter e-mail and instant messages, search the Internet, share files, publish and subscribe to information, market and advertise, coordinate and collaborate with others, personalize services, engage in online communities, and improve the efficiency of on-line commerce between buyers, sellers and intermediaries." (emphasis added)
  

Information Design Patterns

As a follow-up to Moritz Stefaner's 2007 master's thesis, Visual tools for the socio–semantic web, as highlighted in Elastic Tag Mapping and Data Ownership, I am blogging here another noteworthy thesis coming out of the same university.

The Form of Facts and Figures is an unpublished master's thesis presented in early May, 2008 by Christian Behrens at Potsdam University of Applied Sciences, Department of Design.

"The topic of my Master thesis project is the development of a design taxonomy for data visualization and information design. In its core, the project consists of a collection of 55 design patterns that describe the functional aspects of graphic components for the display, behavior and user interaction fo complex infographics. The thesis [when made available will be] in the form of a 200-page book that additionally includes a profound historical records of information design as well as an introduction into the research field of design patterns."

There is a slide-show presentation of parts of his unpublished book at niceone.org.

In the meantime Behrens has currently posted 26 visualization examples from his thesis coupled with descriptions, layouts, implementations and real-world examples to a separate, well-designed website called Information Design Patterns. These very interesting data visualizations run the gamut from Thread Arcs to Data Tips to Stacked Area Charts to Facet Browsing to the following Bubble Chart ....

Information Design Patterns website

Again, but this time essentially in Behrens own words, this website is a design pattern browser consisting of a set of modules that reflect the characteristics of the pattern systematics described in his thesis, and providing the user with a set of useful tools to navigate and explore the collection.

Efficient monitoring of objects in object-oriented database system which interacts cooperatively with client programs

With the rise of the Semantic Web and Cloud Computing, it is a premise of this blog that the most efficient methods - for the authoring and sharing of trustworthy information within social networks and along complex product supply chains - are far and away object-oriented methods.

The following quoted text is taken from a Hewlett-Packard patent filed in 1988 and issued in 1992. It is entitled Method of monitoring changes in attribute values of object in an object-oriented database and still reads fresh even today.

• "Monitoring [of data] imposes a heavy computational and input/output overhead on a database system, especially if the system is large and a number of values are being monitored at the same time for several different clients. Various methods have been proposed to minimize this overhead.
  
  For example, in one such system an 'alerter' is called if specified boolean conditions are satisfied [see Efficiently Monitoring Relational Databases in references]. A 'retrieve always' mechanism in another system causes queries to be re-executed upon each update to specified relations [see Triggers and inference in database systems in references].
  
  Systems of 'triggers' have been proposed for relational database systems; such triggers typically invoke a database procedure upon updates of user-specified base relations [see System R: A Relational Approach to Database Management in references].
  
  A technique which is somewhat similar to the trigger system is the use of a 'declarative integrity constraint,' in which a proffered update to the database is rejected if specified boolean conditions are not satisfied at commit time [see Implementation of integrity constraints and views by query modification in references].
  
  Another technique, access-oriented programming, is implemented in some object-oriented languages such as 'LOOPS'. A message to set values of instance variables is intercepted by means of a user-provided trigger procedure which may in turn set or display some other value [see Integrating Access-Oriented Programming Into a Multiparadigm Environment in references]. The trigger procedures are dynamically added and removed from running systems to avoid interfering with other system logic [see Active Objects: An Access Oriented Framework for Object-Oriented Languages in references].
  
  Finally, expert systems such as "SYNTEL" and "OPS5" provide a method of monitoring virtual memory data retrieved from persistent data [see Syntel: Knowledge Programming Using Functional Representations in references].
  
  Each of these proposed methods offers certain benefits, primarily in the context of the particular environment for which it was designed. However, there remains a need for an efficient way to monitor objects in an object-oriented database system which interacts cooperatively with client programs."

Computing in the Cloud: Possession and ownership of data

The following video is provided by UChannel, a collection of public affairs lectures, panels and events from academic institutions all over the world. This video was taken at a conference held at Princeton University's Center for Information Technology Policy on January 14, 2008. The conference was sponsored by Microsoft.

What you will see is a panel and discussion format moderated by Ed Felten, Director of the CITP. The panel members are:

1. Joel Reidenberg, Professor of Law, Fordham University
2. Timothy B. Lee, blogger at Technology Liberation Front and adjunct scholar, Cato Institute
   
3. Marc Rotenberg, Executive Director, Electronic Privacy Information Center
   

Here is a paragraph descriptive of the questions addressed by the panel. 

"In cloud computing, a provider's data center holds information that would more traditionally have been stored on the end user's computer. How does this impact user privacy? To what extent do users "own" this data, and what obligations do the service providers have? What obligations should they have? Does moving the data to the provider's data center improve security or endanger it?"

The video, entitled "Computing in the Cloud: Possession and ownership of data", is useful and timely. And the panel is well constructed.

Tim Lee, who readily states that he is not a lawyer, very much serves as an apologist for the online companies who believe that "total, one-hundred percent online privacy would mean ... that there wouldn't be any online [sharing] services at all" (Video Time ~ 2:07).

The online services Lee briefly touches upon by way of example are the ubiquitous use of Web cookies for collecting a wide variety of information about usage of the Internet by online users (~5:30), Google's Gmail which employs a business model of examining contents of users' e-mail and tailoring advertising presented to users (~8:05), Facebook's News Feed service which permits users to keep track of changes to their 'friends' accounts, and Facebook's Beacon service which sends data from external websites to Facebook accounts for the purpose of allowing targeted advertisements (~10:54).

Joel Reidenberg, a professor of law, believes that the distinction between government and the private sector loses its meaning when we think of computing in the cloud (~ 15:10), but that the prospect of cloud computing also reinforces the need for fair information practice standards (~16:00). He is of the opinion that as computing moves into the cloud it will be easier to regulate centralized gate-keepers by law and/or by technical standards (~23:50).

• This is my first 'take away' from this video clip in that Professor Reidenberg reinforces the need for centralized standard setting bodies as I concluded in Portability, Traceability and Data Ownership - Part IV.
  

Marc Rotenberg, also a law professor, emphasizes that without user anonymity, and without transparency provided by the online companies, there will be no privacy for users in the cloud (~29:47 - 37:20). And in doing so Rotenberg challenges Tim Lee for his statement that there cannot be complete user privacy for online companies to provide the services they provide (~33.30). This actually makes for the most interesting exchanges of the video from the 38:00 minute mark to the 44:00 minute mark. 

There is also an interesting dialogue regarding the application of the Fourth Amendment. One of the conference attendees asked the panel why there had been no mention of the Fourth Amendment in any of their presentations. Here is the response from Reidenberg at the 53:30 mark:

"Cloud computing is threatening the vitality of the Fourth Amendment ... [because] the more we see centralization of data [by private, online companies], and the more that data is used for secondary purposes, the easier it is for the government to gain access outside the kind of restraints we put on states in the Fourth Amendment."

In other words, why should the government worry about overcoming Fourth Amendment hurdles about confiscating a person's data when it can sit back and relatively easily purchase or otherwise obtain the same personal data from the big online companies? And do so even in real-time? Why, indeed.

For me, the second 'take away' from this video is found in another cogent comment by Professor Reidenberg at the 88:53 mark:

"The [online] company that ... figures out ways of ... building into [its] compliance systems ... [privacy] compliance mechanisms ... will be putting itself at a tremendous competitive advantage for attracting the services to operate in [the cloud computing environment]."

The technological data ownership discussed and described in Portability, Traceability and Data Ownership - Part IV, supra, is a privacy compliance mechanism.

For those who are interested in the legalities and government policies revolving around burgeoning data ownership issues related to software as a service, the Semantic Web and Cloud Computing, and who are motivated to sit through a 90 minute presentation, here is the video clip ....

Update on Thursday, June 5, 2008 at 12:09PM by Steve Holcombe

In the United States, and notwithstanding the impotency of the Fourth Amendment protections against government search and seizure, it is an irony that the growing centralization of the Cloud may well render it more amenable to government regulation and lawsuit liability.

"Cloud computing opens doors for privacy enhancements [driven by regulation]. It's easier to target for regulation by law or by technical configuration [the] gatekeepers. So to the extent there is a central management, they're easier to find, they're easier to regulate, and they're easier to hold liable than distributed systems."

Quoting Reidenberg at the 23:50m mark.