BOLD Analysis
Melvin Carvalho
====
Here's my take on the analysis of Linked Data. I'm a big believer in web technology, in particular, over the years, I've reached the opinion that Linked Data / Sem Web is deeply aligned to the end game of the web. At the beginning of this decade, I made the decision to leave Wall Street, where I had started my career as a graduate, and moved to teaching myself about Web Technologies. The reason being that I felt that investment banking was a declining industry, and that the web was rising. Time will tell if I was right on that call!
The reason I mention this, is that I became fairly well acquainted with current business analysis methodologies. However, I would say that almost all modern popular analysis is highly inappropriate for describing the Business of Linked Data. The reason being that analysts have a quarter by quarter prediction time frame, up to say, 18 months, and it's well proven that their estimates outside of this range are generally poor. Indeed many are not allowed to predict anything further than two years out. This timeframe simply is not appropriate for the analysis web technologies.
Therefore, I will use a more long sighted methodology, as used by my old bridge adversary, Warren Buffett, and by value investors. For those not familiar it comes down to having a good technical product (which is well discussed), but also protecting your business by what's called a 'moat'. The four classical moats are : Network Effect, Intangibles, Switching Costs and Economies of Scale. The fifth moat has emerged in recent decades is called 'Ecosystem'. I'll discuss the landscape of these 5 with a view to where BOLD can fit in.
Network Effect
============
A topic in itself, as Kingsley emphasised in another thread. Many 2.0 startups have made a play on network efffect, where the basic principle is that new additions to a large network become worth more and more, as the number of participants grow, much to the detriment of all other competitors. This has probably been the classic 2.0 strategy, with facebbook / GYM / Ebay and a plethora of others achieving big audiences, and large notional values due to a smooth interface and the ability to attract or buy users. However in the last few years, the walls in the garden appear to be coming down, with the web turning into a kind of giant platform, and one huge network. Therefore, I think with time BOLD, though at present is confronted with several elephants in the room, will be well placed to take advantage of this trend which will surely continue.
Economies of Scale
===============
I think Amazon Web Services and Google and the like have done a brilliant job taking market share in the data area. However this has driven down the cost for applications, which is the lifesblood of BOLD. So, as long as you're not trying to compete with amazon on hosting, you should be able to benefit from an improving environment to build your apps. Though, It may still be possible to take on the big guys if you broaden the cloud to include end user PCs.
Switching Costs
============
This really is an area where the 2.0 look to have a very strong moat. Both habit and the hassle of changing providers encourage users to stick with the services they know. It's a hard moat to beat, but it appears to be eroding nonetheless, with interop becoming more common, however it appears that business using BOLD will have to really up the stakes in terms of product and user experience, in order to attract customers. The bar is quite high, but we can get there and higher.
Intangibles
=========
This includes things like brand, IP and other items that are not physical. I think MySpace is a classic example of how fickle a brand can be on the net, so the scene is set for the next set of giants to emerge. Re: IP I think things are lookng quite good, because, although many firms hold a nuclear arsenal of patents, non aggression seems to be the standard position. The W3C also shows incredible leadership and has an unparalleled track record of keeping good ideas open and therefore allowing them to thrive.
Ecosystem
=========
The newest moat, I believe, is the play that BOLD business should be making, while, at the same time, waiting for the other moats to erode. Classic examples of firms that have benefited from an ecosystem of apps are Apple (iphone), Facebook and Microsoft itself. However the master / slave approach offered in almost all cases, can quickly become dwarfed by the power of apps that can talk to each other in the broader web OS. IMHO, this is where the innovation lies, and it's definitely a case of the sum of the parts being much greater than the whole. In collaboration BOLD businesses can each claim a section of which can become the biggest pie, on the net, or perhaps the planet. The great advantage of Linked Data is that its open schemaless nature allows every peer to benefit from a single subsystem. A classic example might be secure microcredits that can be sematnically passed from app to app to generate an economy, another might be a reputation system across all WebIDs (we're close already), yet another is FOAF+SSL providing secure login for all apps that wish to you it. The list goes on and on. As BOLD business start thinking and collaborating in this manner, we're going to start to see some explosive synergies, and consequently, growth. Will this be enough to overcome the other moats? Well I, for one, am a believer!
Conclusion
=========
BOLD business appear to be in a holding pattern right now, while other dominant web firms inevitably watch their moats erode. Nevertheless, I think that the average business in this space will achieve double digit growth for decades to come, which if it was a country people would be calling an economic miracle. It's possible that linked data can explode at any time, but that's a pretty difficult game to predict. I think the correct business strategy for Linked data, is one which the name implies, ie to target interop and the building of an unrivaled ecosystem, and using the killer advantage of the semantic web to allow applications to talk to each other with a new kind of intelligence. With egov coming on board linked data in a big way this year, I think the BOLD is hear to stay, indefinitely. As a technologist, I find this the most fascinating area on the web right now. However, with my limited appreciation of business, I find the economic case to be equally compelling!
Just my $0.02
Kingsley Idehen
Welcome!
As usual, very good understanding of the burgeoning opportunity :-)
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
Adrian Walker
Thanks for a very thoughtful and thought-provoking posting.
Here's something that potentially has most of the desirable moats you mention. -- Social Media support for Executable English Q/A
It starts with the observation that (linked) data by itself is necessary, but not enough, for practical applications.
What's also needed is knowledge about how to use the data to answer questions -- for example, "how much could the US save through energy independence?"
There's emerging technology, live online, that leverages social media for the significant task of acquiring and curating the necessary knowledge -- in the form of executable English.
One can Google "executable English" to find this.
Imagine government and other web sites being able to answer questions like the one above, and also explaining the answers in English. Imagine citizens socially networking to continually expand the range of questions that can be answered.
As mentioned, the technology exists, and it's live online with many examples that one can view, run and change, using a browser. (The above energy independence example is there already.) Shared use of the site is free. Just Google executable English and follow the links.
Just my additional 2 cents.
-- Adrian
Kingsley Idehen
> Hi Melvin,
>
> Thanks for a very thoughtful and thought-provoking posting.
>
> Here's something that potentially has most of the desirable moats you
> mention. -- Social Media support for Executable English Q/A
>
> It starts with the observation that (linked) data by itself is
> necessary, but not enough, for practical applications.
How would you apply Executable English to the task of Annotation? Think
del.ico.us agging pattern with benefits (i.e., ability to make
structured statements about anything).
>
> What's also needed is knowledge about how to use the data to answer
> questions -- for example, "how much could the US save through energy
> independence?"
This is intersection of Linked Data and Answer services like
WolframAlpha and TrueKnowledge. Both offer APIs already.
>
> There's emerging technology, live online, that leverages social media
> for the significant task of acquiring and curating the necessary
> knowledge -- in the form of executable English.
>
> One can Google "executable English" to find this.
>
> Imagine government and other web sites being able to answer questions
> like the one above, and also explaining the answers in English.
> Imagine citizens socially networking to continually expand the range
> of questions that can be answered.
>
> As mentioned, the technology exists, and it's live online with many
> examples that one can view, run and change, using a browser. (The
> above energy independence example is there already.) Shared use of the
> site is free. Just Google executable English and follow the links.
Caveat re. "Executable English", our world is multilingual; thus, we
can't make Language based Data Silos re. Business of Linked Data :-)
Kingsley
Adrian Walker
You ask some good questions, and I'll have to think a bit about them.
However, you wrote:
This [executable English] is intersection of Linked Data and Answer services like WolframAlpha and TrueKnowledge. Both offer APIs already.
Sorry, but that misses the point, which is for NON-programmers to be able to input and refine knowledge in executable English, and to see English explanations of the results of running the knowledge. This goes much further towards leveraging social networking than just providing an API for programmers.
You wrote also
Caveat re. "Executable English", our world is multilingual; thus, we can't make Language based Data Silos
Sticking with APIs really does would not solve the multilingual problem, since there will be subroutine names and documentation in English, French, German....
It's a tough problem, but one step towards solving it is to observe that the executable English technology online at the site below takes an unusual approach. It's not the usual "controlled English". The vocabulary is open, and so to a large extent is the syntax. This means that one can actually include phrases in French, German and so on. The phrases take their meaning from their relations to other phrases, rather from the usual dictionary approach. Strict semantics is achieved via a trade off [1].
However, that's only a first step. When the machine translation folks achieve 100% (:-), we will be in better shape.
Cheers, -- Adrian
[1] www.reengineeringllc.com/A_Wiki_for_Business_Rules_in_Open_Vocabulary_Executable_English.pdf
Internet Business Logic
A Wiki and SOA Endpoint for Executable Open Vocabulary English over SQL and RDF
Online at www.reengineeringllc.com Shared use is free
Adrian Walker
Reengineering
Kingsley Idehen
> Hi Kingsley --
>
> You ask some good questions, and I'll have to think a bit about them.
>
> However, you wrote:
>
> services like WolframAlpha and TrueKnowledge. Both offer APIs already./
>
> Sorry, but that misses the point, which is for NON-programmers to be
> able to input and refine knowledge in executable English, and to see
> English explanations of the results of running the knowledge. This
> goes much further towards leveraging social networking than just
> providing an API for programmers.
TrueKnowledge offer APIs into these data spaces which allow you to query
using their specific variant of Natural Language Query Constructs.
Now, if you make a Linked Data mesh from these services, and then
combine (via lookups) with other Linked Data Spaces you get something
quite interesting. Basically, this is something we do with our Meta
Cartridges which are orchestrated by our Sponger Middleware [1][2][3].
The social dimension (as per my puzzle answer) comes from inserting
granular profile data into this mix :-)
Your key to social data injection is a WebID served up by a FOAF+SSL [4]
based platform (ideally, if you want to stay on the good side of privacy
matters).
>
> You wrote also
>
> /Caveat re. "Executable English", our world is multilingual; thus, we
> can't make Language based Data Silos
> /
> Quite right. Though, for certain tasks, such as air traffic control,
> we do agree to be unilingual.
Well I posit that one of the strongest USPs of OpenCalais is the fact
that it is multilingual.
>
> Sticking with APIs really does would not solve the multilingual
> problem, since there will be subroutine names and documentation in
> English, French, German....
APIs are anti-patterns in my world view. I am an Open Data Access guy
who sees REST as the answer to life long frustrations I've had with Open
Data Access. Remember, to me: Code is Like FISH and Data is Like WINE. I
can only tolerate thin layers of smart orchestration oriented code.
Basically, the dominant spaghetti variant of code makes my hair stand
(literally).
I believe in using data orchestration to drive the delivery of
application/solution experience. REST delivers on this approach with aplomb.
Links:
1. http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtSponger
2. http://virtuoso.openlinksw.com/screencasts/virtuoso-rdf-middleware3.swf
3. http://uriburner.com -- service based on the Sponger and its Cartridges
4.
http://www.slideshare.net/bblfish/building-secure-open-distributed-social-networks-presentation?src=embed
Kingsley
>
> It's a tough problem, but one step towards solving it is to observe
> that the executable English technology online at the site below takes
> an unusual approach. It's not the usual "controlled English". The
> vocabulary is open, and so to a large extent is the syntax. This
> means that one can actually include phrases in French, German and so
> on. The phrases take their meaning from their relations to other
> phrases, rather from the usual dictionary approach. Strict semantics
> is achieved via a trade off [1].
>
> However, that's only a first step. When the machine translation folks
> achieve 100% (:-), we will be in better shape.
>
> Cheers, -- Adrian
>
> [1]
> www.reengineeringllc.com/A_Wiki_for_Business_Rules_in_Open_Vocabulary_Executable_English.pdf
> <http://www.reengineeringllc.com/A_Wiki_for_Business_Rules_in_Open_Vocabulary_Executable_English.pdf>
>
>
> Internet Business Logic
> A Wiki and SOA Endpoint for Executable Open Vocabulary English over
> SQL and RDF
> Online at www.reengineeringllc.com
> <http://www.reengineeringllc.com> Shared use is free
>
> Adrian Walker
> Reengineering
>
> On Sat, Dec 19, 2009 at 11:36 AM, Kingsley Idehen
> <kid...@openlinksw.com <mailto:kid...@openlinksw.com>> wrote:
>
> Adrian Walker wrote:
>
> Hi Melvin,
>
> Thanks for a very thoughtful and thought-provoking posting.
>
> Here's something that potentially has most of the desirable
> moats you mention. -- Social Media support for Executable
> English Q/A
>
> It starts with the observation that (linked) data by itself is
> necessary, but not enough, for practical applications.
>
> Adrian,
>
> How would you apply Executable English to the task of Annotation?
> Think del.ico.us <http://del.ico.us> agging pattern with benefits
> <mailto:melvinc...@gmail.com
> <http://www.openlinksw.com/blog/%7Ekidehen>
Kingsley Idehen
I put this in the: slowly but surely bucket :-)
Read:
http://networksingularity.com/2009/12/14/patternndashbased-strategy.aspx
They focus on the Social Networking aspect i.e. Linked Water Cooler Data
Space .
A long time ago, my basic enterprise example re Linked Data exploitation
used to go something like this:
1. Employee joins a new company
2. Email comes in
3. Employee has the ability to click on the mailto URI exposed by email
client en route to description of mail sender, which unveils information
such as: This contact is associated with an Organization that placed the
largest order this year, to date, and there are 3 support cases open at
this point in time etc..
Kingsley Idehen
Quick raw dump that might become a blog post down the line etc..
Here is a list of companies that came to dominate markets at different
times in our recent history based on their ability to identify and ride
the Network Effects vector of their particular time.
IBM -- Computer (many people connected by time-shared computer access
via terminals).
Microsoft -- Desktop (at expense of Apple via MS-DOS which ran on
multiple desktop computer hardware)
Web -- Various stage related plays
Bootstrap --- Netscape (show masses the value of a LINK)
Commerce --- Amazon (unveiled what you can purchase via a LINK)
--- eBay (unveiled haggling via LINKs)
Search
--- Yahoo! (unveiled directories behind LINKs)
--- Altavista (unveiled smart full text pattern search
heuristics behind LINKs)
--- Google (unveiled entity rank + full text pattern
search heuristics behing LINKs + Scalable AD Model)
Web 2.0 -- Various stage related plays, but the key driver and
bootstrapper was Dave Winer via Userland and Frontier.
Bootstrap -- RSS (data container format) as foundation for
Blogging and disruption of Journalism (which naturally extends to
politics as demonstrated by Obama) via Web Services based PubSub mechanism
Collective Intelligence -- Wikipedia (this particular seed has at
best just passed germination stage: DBpedia is but one example of the
effects of Wikipedia)
Web 2.5 -- Twitter ushers in micro-blogging (#hastags and @identifiers
speak volumes about everything the Semantic Web Project has
fundamentally been about, but without the vital structured data rigor
and platform independence).
Web 3.0 -- company ??, I do believe this is most likely going to be
bootstrapped by annotation style nano-blogging (with "open world"
blogic[1], structured profiles, and Linked Data as behind the scenes
technology differentiators).
Links:
1. http://www.slideshare.net/PatHayes/blogic-iswc-2009-invited-talk
2.
http://netmesh.info/jernst/big_picture/from-1-to-a-billion-in-5-years-what-a-little-url-can-do
-- imagine what happens when you tack on #<somthing> frag ids to those
URLs, you get HTTP URIs :-)
Melvin Carvalho
Hi Melvin,
Thanks for a very thoughtful and thought-provoking posting.
Here's something that potentially has most of the desirable moats you mention. -- Social Media support for Executable English Q/A
It starts with the observation that (linked) data by itself is necessary, but not enough, for practical applications.
What's also needed is knowledge about how to use the data to answer questions -- for example, "how much could the US save through energy independence?"
There's emerging technology, live online, that leverages social media for the significant task of acquiring and curating the necessary knowledge -- in the form of executable English.
One can Google "executable English" to find this.
A very good idea. I suspect semantic messaging will have several 'handlers' to process different types of request. HTTP GET, SPARQL, SPARUL being low level, but executable english is definitely something worth persuing. A basic attempt is here:
http://sparqlbot.semsol.org/manual
Competitive Analysis shows if someone was to specialize i this area, they would probably become the market leader quite quickly, as the space does not see to have any esablished incumbant. Seems like a very good business opp to me, for someone that has the bandwidth to invest some time in it.
Adrian Walker
Many thanks for your comments, and also for the link to the sparqlbot manual.
As you say, sparqlbot is quite basic from the natural language point of view.. In contrast, a key point about full executable English is that non-programmers can write it, using their own words and phrases. This means that other non-programmers can also read and understand the knowledge -- and most importantly, link it and improve it.
When needed, the system online at the site below uses information in executable English to automatically generate and run networked SQL queries, and it explains the results -- in English -- at the business level.
Cheers, -- Adrian
Internet Business Logic
A Wiki and SOA Endpoint for Executable Open Vocabulary English over SQL and RDF
Adrian Walker
Reengineering
Kingsley Idehen
Here is a nice read that sums up the mercurial issue of: beyond
advertising business models, that are all about exploiting LINK density
across Linked Data Spaces.
The case study here is Amazon's affiliates network.
This individual demonstrates that individual media mogul at work. It
also show cases effect of URIs re. identifying contributors to economic
value chains (which is why URI visibility is so vital to fundamental Web
resource publishing patterns).