Why are we still talking about search?

Google is obviously pretty successful due in large part to its search service. Despite them not having the best social networking track record, it’s an activity that maps well to the social world. We search for things – our keys, a present, a holiday, our next partner, a restaurant to take a friend to - our next role - ourselves. Google doesn’t give us exactly what we want but that’s OK because we don’t know exactly what we want either. It’s useful to be given a few social curve-balls.

Despite us all mainly using Google or Bing (Yahoo, Ask etc. all gone now), new search engines e.g. Blekko still regularly surface with varying differentiations and significant funding.

What if, rather than vague searches (inevitably) generating prioritized yet potentially vague results (M:M) [in terms of question:answer]; we could ask a specific question that provided a generally accepted, precise result (or if that question had previously recently been addressed, a link to same) that we (and others) could refer to (to later justify our logic). What if we wanted - an actual answer (1:1)?

Consider - “Which is the largest State in the US?” The question is vague (by land or population?) also Google is mainly matching keywords rather than actually understanding the question (and of course we get many results so the full exchange is M:M). Actually, Google have been delivering factual answers to some queries for five years. It only works for some though (it doesn’t work in this case). Regardless, Google’s first result correctly tells us Alaska and additionally gives us comparison with other States and square mileage too. We aren’t guaranteed getting the same result when we query in future (so accountability goes out the window) but the answer is OK for most purposes. If we make the question “Which is the largest Republican State in the US?” we still get an answer but takes us a few minutes to manually cross-reference the links. If we ask “Which is the largest, happiest Republican State in the US?” we simply don’t get a usable answer.

Someone living in Alaska is worried about perceived rising Russian immigration and they want to know if it has risen in their adult lifetime. There are no obvious figures on the Web. They could tweet/email the Governor and (if enough others have asked broadly the same question) then he might post a response on his blog (if he has one) once his research team have arrived at some defensible statistics for him to refer to (assuming they can find some).

This simple/current scenario has challenges and inconsistencies on both question and answer sides. Here are a few:

1) People have to feel so strongly about the issue that they have to find time to interact – search for any similar questions/answers, write their email, monitor the Governor’s twitter-stream/blog for an answer. They feel as though they don’t want to put themselves forward unless others feel the same. Conversely, enough people have to use similar phrasing of the question in order to kick-start the answer process.
2) The answer (if it is given) is not explicitly tied to the question. It could be made on his blog, a successive TV appearance, in a press release or personal speech.
3) There is no disambiguation or prioritization of questions – meaning that there is just so much noise that the Governor or his team simply don’t have the time to answer.
4) The answer could be presented in different ways – perhaps to achieve political ends or simply because they are underused by Government e.g. taking Russian immigration over a twenty year time-frame may show no year-on-year increase but taken over a two-year time-frame it may show a very different story.
5) There is little opportunity to to-and-fro with the question and answer thread – blending anecdotal evidence, redefining terms.
6) It’s a naive voter with a question - asking a team of well researched people. Odds are there will be a soft soap answer as independent experts are not involved in the process to moderate.
7) Statistics could come from Federal or State government, commercial, charity or other bodies. Also there is unlikely to be any checks and balances e.g. does the US figure for people incoming match with the Russian figure for people outgoing?
8) Any presented figures will be difficult to understand (unless an intuitive and consistent graphical approach is used).
9) Lawyers may need to be consulted to define the term “immigrant”.
10) There is no accepted “end” to either the question or answer.

In business, we need - to commit resources to a business plan, to indemnify ourselves against litigious customers, to make our point in meetings and to provide defensible input to decision making. Consumers largely search for things (M:M) while companies have to query things (1:1), although as with social networking, there’s some overlap (M:1) in the middle. Companies lose out as they maintain a blinkered-view missing critical information not captured through their own processes. Consumers lose out since they spend so much time searching for things (and the Internet is so rich) that, like travelling, the journey becomes the destination.

Google Search Appliance has done very well in the enterprise. There is a need for a one-ring-to-rule-them-all "get things" function. Also, our work and home lives are merging and when the answer you present is (close enough to be) unanimously accepted and is able to be referenced in any future analysis of your decision - then it will always stand. Once you don’t have to think about the logistics of getting answers, you can concentrate on actually doing things with them (for both consumers and the enterprise). To do this, quite simply - we need a single answer, a 1:1 model that works for everyone - like search – a Q&A service.

We have “I’m Feeling Lucky” of course (M:1) but this is an indulgent lucky dip that actually loses Google revenue. The other way around (1:M) doesn’t work since it entails definitively stating the question e.g. using something like QBE or NLQ (both of which are hard to do right) but still getting a range of results. What other services are currently available? Facebook released Facebook Questions earlier this year. This is slated to be integrated into Community Pages (already includes Wikipedia content). It is developing but is still very social/free-text. Wikipedia is an encyclopaedia rather than a Q&A service. Yahoo Answers has more of a wiki/free-text approach. It probably is the widest used but is also known for being random and open to abuse. PeerPong is expert-focussed. Off-topic a little, for product reviews, there are Hunch, DooYoo, Epinions, Mahalo (their Answers service is a bit wider focussed than just products though but still very social). There are mobile Q&A services such as Mosio and ChaCha. Quora (founded by ex-Facebook CTO Adam D’Angelo) appears to stack up against Facebook Questions and again is very social. Social Q&A in general is big. StackExchange is an open-source platform for developing your own social Q&A service. In the enterprise, Opzi is attempting to be a corporate Quora. MSFT shut down their offering last year. Qhub runs a niche hosted Q&A service. Some social ones such as Hunch have repositioned themselves as recommendation services.  Many just cater to niches e.g. StackOverflow/development. OSQA is a free open-source one. Several dot-com bubble casualties have refocused on discussion based Q&A e.g. Ask has refocused on a mobile Q&A service similarly Startups.com.There are other start-ups.

All of these services are in some way niche, most require an account (which will put off many) and most provide multiple answers (1:M), although collectively they will probably end reliance on reference libraries (and on the product side – magazines like “Which?”). They helpfully hand you a piece of the jigsaw rather than solving the jigsaw for you. As has previously been said "None of these sites are Google-killers. In fact they make Google stronger because the questions and answers often will be indexed - extending Google's reach into the tail". Quora in particular has been recently heralded as a killer Q&A service but asking it "What are the most effective ways to engage news audiences?" receives a bunch of people's opinions; whereas we could have a single list in order of effectiveness that has been produced through actual operational data and curated by either a Journalist or Statistician. If there were any dissention around the term "effectiveness" then that could be hammered out through social interaction. There wouldn't be a need for a - discussion. As with all Quora questions, you need to read all responses to get a full answer but even then its just a subset of people that responded. Another example - RockmetIt better than Flock?

If the industry is framing Web 2.0 is the Social web and Web 3.0 as the Semantic web – it seems churlish not to leverage the powers of both in the Q&A service. The benefits of great data integration using common terms (Semantic) and crowd-sourcing (Social) in a Q&A service are obvious. By the same token, some answers will always require a (semi)/professional element to them (Expert) e.g. Someone asking a question about Alaska’s closeness to Russia (surprisingly - 2.5 miles) might get a quick answer but it will take an expert to plan how to get from one to the other through that particular route. Expert knowledge also comes into play around the presentation side – knowing which facts to present to most effectively answer questions e.g. the largest State question above – given to another search engine also gives Alaska but the square mileage is significantly less. A Geographer knows this is due to the second result not including water. There are other components but, in the main, both questions and answers have - social, semantic and expert elements.

Back to the Alaska question (“Which is the largest, happiest Republican State in the US?”). This requires both a semantic element (to determine the largest State) and a social element (who is happy?). If we make this “Which is the largest, happiest Republican State in the US liable to switch to Democratic?” then there is an expert element too.

Onto the proposed Q&A service - it needs to check a few boxes right off the bat:

1) Social. It would have to be based upon a popular, real-time social system. Despite perceived issues investing in their ecosystem, Twitter is arguably better than Facebook (or any other) for this purpose but it also needs to amplify and expand the reach of questions: present persistent questions that have undergone a process of disambiguation (both automated using semantic algorithms and manual through a public process of “backing” existing questions). Questions also need to be prioritized based on number of backers. Something like Kommons, Replyz  or maybe Formspring or Open Media could work here. Amplify is maybe too discussion focussed. The focus here is on directly (directly to the tweeter) quantifying the importance of a question and the validity of an answer. Social media monitoring like Crimson Hexagon could offer supporting services but they typically rely upon sentiment analysis whereas we really need direct engagement through voting or other mechanisms.
2) Semantic. We all need to be sure at the very least that we are talking about the same data. In short it will need an index. Google have Rich Snippet functionality but this is not a common approach adopted by most site owners. Google also acquired Metaweb earlier for this year for this purpose but something more open like Sindice would really work providing organisations have an incentive to cooperate (integration/ontologies/micro-formats etc.). Providing and articulating the incentive is the toughest piece of this by far. Then – we want an interface on the results (produced by the index) that allows us to de-select terms that we are not interested in so that the system knows for next time but also to publish the results of our tinkering and link it to the threads in the question or answer. Something like Sig.ma would do this.
3) Expert. We need data sets that are both selected and curated by subject matter experts. Curated answers can often be the most useful. Google acquired Aardvark earlier this year for this purpose but something like Wolfram Alpha or Qwiki would work. Involving Wolfram Alpha would add other benefits too – computational power (through Mathematica) to aggregate on-the-fly, leading NLQ support to translate questions, updatable widgets to support the presentation side and the semi-celebrity name of Stephen Wolfram attracting academic curators. All Wolfram Alpha content would need to be indexed and we would probably need to extend the publication side to be able to deselect returned facts as required (doesn’t appear to currently be available).

We are talking about potentially lots of social and semantic data needing to be agreed and presented. We’ll need simple data visualization with a basic level of interactivity so, with the example above; the user can slide the time scale to represent his adult lifetime. Something like Many Eyes, Hohli, StatPlanet or some improved Google Charts should do here. Also, we will want to store all our questions and answers forever so that we can reference them. Partnership with a very large EMC-like network storage partner will be required.

The resulting composite solution would be the most advanced Q&A service in the world. Q&A is so important and multi-faceted that it needs multiple solutions to work together. Its influence, if accepted, would be profound.

The service would drive out real answers (reference and topical) through a combination of social, expert and semantic approaches and actually referentially improve itself - as bringing social attention to the data would force organisations to improve data. Usability-wise - it wouldn’t hamper ad-hoc users by having a scary QBE type interface or by blinding users with spreadsheets of data. Conversely, it wouldn’t be so simple that it misses the point of what people want. It would be a natural extension to tweet activity (retweet, favourite, reply etc.). Opportunities for monetization would be rife: not least advertising with the solution’s wider collaborative user base, increased opportunities for interactivity and greater relevancy. New forms of credibility and relevancy would also become available. Journalism would shift to sourcing stories from datasets.

The traditional search space e.g. "Alaska Jobs" is already being eroded by location-based services. Google is certainly well placed to build a killer Q&A service as is Facebook with its obvious social advantages and semantic/graph plans. Google is confusing though - on the one hand, they have the right investments and infrastructure but on the other they still appear search/data/click driven - even imperiously so ("...designers...need to learn how to adapt their intuition"). They also have issues with innovation.

The big search players generally recognise current search solutions do not meet the needs of the user but they don't have obvious solutions. There is an opportunity for innovation and for weaning people of this style of information seeking. There are smaller players that, if they cooperated now, could produce a more potent, compelling, flexible and lasting solution than any. Their biggest problem would be being creative about incentives for organisational involvement on the semantic side. Let’s hurry though - its hard going searching for things. We've got work to do now. We have questions. 

Statistical thinking & the ability to read and write

Statistics are hugely underused. Even when they are used, they are improperly used. Much of the bolstering of weak arguments, miscommunication and ad-hoc ideology creation performed by individuals and organisations from the industrial revolution onwards are down to poor statistical use.

Entire industries are devoted to generating statistics; duplicating monstrous amounts of effort and when they are referred to e.g. in sales pitches, organisational reports, building/product tolerances, war crimes tribunals, political and environmental manifestos, news articles and business plans – they are so heavily caveated as to make them insensible. This is by contrast to the relatively deterministic, transparent and auditable approach organisations use to produce BI to support their own business decisions. Coupling this BI with governmental statistics is often necessary for the best decision making support, so commercial BI is itself hamstrung by statistics.

There are two key reasons for this:

1) Statistics are hard to find. If you are looking for, for example, the number of people that work in London currently (a simple enough request that many service organisations would need to be aware of), you will find this close to impossible. The UK Office for National Statistics (ONS) does not have this on their main site. Nor does the Greater London Authority or the newly released UK government linked data site. After you have wasted maybe twenty minutes of your time, you will be reduced to searching for “how many people work in London” then trawling through answers others have given when that same question has been asked. You will receive answers but many will be by small organisations or individuals that do not quote their sources. In the worst case, you may not even find these – instead using an unofficial figure for the whole of the UK which you have had to factor down to make sense just for London. If you search hard enough you will find what you are looking for at an ONS micro-site (completely different URL to ONS) but this data is over five years old.

2) Statistics have a poor image. The blame for this, in part, may be attributed to the famous Disraeli quote – “Lies, damn lies and statistics” in which he set generations of professionals into thinking they were akin to a practicable yet modish Victorian politician by disregarding statistics and cocking-a-snoop at the establishment in favour of their own experience. Showing you are practicable with maverick tendencies while (in overtly disregarding information that may cast doubt on your decision-making) shoring yourself up from failure - are powerful incentives. By contrast, other famous statistical quotes have been forgotten (“Statistical thinking will one day be as necessary for efficient citizenship as the ability to read and write” [H.G Wells]). Short-sighted governmental data integration, hugely delayed/over-budget government data-centric projects such as the UK National Health Service’s Records System, confusion over keys statistics e.g. number of asylum seekers and high-profile data losses haven’t helped matters since.

There is also a bit of a myth that statistical interpretation is an art and that the general public can be confused – even sent entirely the wrong message by engaging in statistical understanding. Only statisticians can work with this data. Certainly there is this side to statistical analysis (basically anything involving probability, subsets, distributions and meta-statistics) but for the most part, both the general public and organisations are crying out for basic (the answer to one question without qualifiers e.g. where/if etc.) statistical information that is quite simply - on a web-site (we can all just about manage now thanks), produced by or sponsored by the Government (we need to have a basic trust level) with a creation date (we need to know if its old). If statistics are estimates, we need to know that and any proportions need to indicate the sample size. We need this since we are now sophisticated enough to know “8/10 owners said their cats preferred it” has less impact than if we are talking about a dataset of ten cats rather than 10K cats (we don’t really need this though since we’re capable of working out proportions ourselves).

We don’t want graphs since the scale can be manipulated. We don’t want averages since it is similarly open to abuse (mean, median or mode?).  If we make a mistake and relate subsets incorrectly then the people that we are communicating to may identify this and that in itself becomes part of the informational mix (perhaps we were ill-prepared and they should treat everything else we say with care). We actually don’t need sophisticated Natural Language Processing (NLP), BI or Semantic Web techniques to do this. It would be nice if it were linked data but concentrate on sourcing it first. We really are not that bothered about accuracy either (since its unlikely we are budgeting or running up accounts on governmental statistics).

Mostly we are making decisions on this information and we are happy rounding to the nearest ten percent. Are we against further immigration? Is there enough footfall traffic to open a flower shop? Do renters prefer furnished or unfurnished properties in London? Which party has the record for the least taxation? What are the major industries for a given area? We just need all the governmental data to be gathered and kept current (on at least a yearly basis) on one site with a moderately well thought-out Query By Example (QBE)-based interface. That’s it.

Reading and writing have been fundamental human rights in developed countries for decades. Broadband Internet access is fast becoming one too. Surely we need to see access to consistent, underwritten government statistics in this vein too. Where other political parties dispute the figures, they should be able to launch an inquiry into them. Too many inquiries will themselves become a statistic – open to interpretation. It is absolutely in the interest of organisations and current affair-aware individuals.  

Semantic Web - Part 2 (Where is it?)

The enterprise in general has barely given these technologies a second thought to-date and the consumer has little idea about them (short of a vague idea that Web 3.0 will make the Web more intelligent). The concept of storing more data in a graph/RDF format remains disputed. MSFT, for example (as a populist bridge between the two), could be said to have a less-than-enthusiastic approach (all the main APIs are in Java with .NET versions managed only by enthusiasts). Few MSFT products use RDF internally (Media Management). None (including SQL) use OWL/SPARQL. Google have their recent Rich Snippets initiative but their Chart API currently only works with spreadsheets (rather than RDF). Facebook are actively pursuing the graph format as its on a secret logo inside Mark Zuckerberg's hoodie. Twitter have recently announced annotations - a way to add meta-data to tweets (which could be used semantically in future). Some breaking sites use RDF e.g. Glue, Drupal and Tripit but there are no killer apps yet.

The world awaits an application that inherently makes a tool of the Semantic Web. This will likely be focussed around disambiguation since wholesale data integration is a tougher nut to crack.

Reasons follow (descending order of importance to-date):

1) Openness. There are few clear reasons for the enterprise (the people who manage the vast majority of data) to be more open with data. Especially their raw data (as opposed to their massaged/reporting data). Government has a remit of transparency so they have more data in this format.
2) Federation. Of business processes. There are a host of facts (and rule-based logic linking them together) required to make the above scenario (and anything like it) function; all working over several different organisations with different remits. Each taking revenue share. Building federated applications; using other peoples data are also rife with issues of SLA and legalities.
3) Performance. Storing data as facts (triple or graph) results, in most cases, in a poor performing back-end. Relational databases are widely recognised as most efficient for most scenarios and therefore what most organisations use. Tests indicate triple queries are on average 2-20 times slower than an equivalent relational query. This alone instantly negates OWL/RDF and SPARQL for widespread use in the enterprise. There are also huge uncertainties over how distributed SPARQL queries and reasoners will work in practice - at Internet scale.
4) Ambiguity. Many believe that the real world is just too complex to be mapped:
a. It is just not able to be boxed and named in terms of categorisation and relationship (or ontology as it is called). This is essentially the same as the schema or Entity Relationship Diagram (ERD) for relational databases.
b. Linking ontologies together accurately (differing owners, world-views, drivers) is impracticable at Internet scale. Related to this is the almost philosophic issue around the use of the URI to identify a resource. It is hard to give a crisp definition of what 'representation' (of a resource) means in order to justify an assertion e.g. an image of Texas does not 'represent' Texas.
c. The recombination of facts necessary for inference are too simplistic e.g. the addition of - drunks (subject) frequent (predicate) bars (object) to our scenario might allow our agent to infer that the CEO is therefore a drunk. This may or may not be true but given known facts, it is a rash statement to make (especially if you are considering networking with him). You might not even know that the agent considers the CEO to be a drunk (it will just be one of the many factors that it uses to provide actions for you). This makes the situation much worse since bad decisions are then difficult to debug/improve.
5) Validation. Maintaining graph data is less structured than a relational format. It makes validating data integrity challenging. Data quality is a huge issue for the enterprise. Many CIOs will look to the ceiling when you ask them about their data quality and meta-data (needed by the Semantic Web to function) takes a back step in priority. Existing relational databases have much better tools and processes.
6) Semantics. The Semantic Web community have not helped themselves by casually using unclear terms, bolting on reasoning/inference to the Semantic Web definition and generally setting-up camps about well - semantics. Running parallel to Semantic Web development has been Natural Language Processing (NLP) development which, by contrast, has a clearer mission statement, can achieve some of the same goals and is actually more about human language semantics than the Semantic Web.

The first two above are related. There is simply an absence of reason for the enterprise to be more open with its data and link its transactions and business processes with other enterprises right now. However, it is fair to say there is a small revolution going on in respect of consumer personal openness or privacy. Consumers are seriously considering the benefits of publishing their purchases and sharing their screens for mass consumption. At the same time, Facebook is criticized for being too open and free-wheeling with personal data. What this tells us is that – the consumer wants to be in control of their data. This desire to control their data, all of it, all the time is essentially a form of self-actualisation; managing their on-line presence – facts, hopes, desires, history and image. Self-actualisation as popularised by Maslow is an established end-game for individuals and society in general. This will happen initially due to consumer demand. The enterprise will then be dragged into this process by more powerful forces of expectation and socialization (Web 2.0) – market forces. They will have little choice but to serve up their data and integrate with others. This could happen surprisingly quickly – within a year: Six-months of observed consumer demand, three-months to get an enterprise pilot going and another three for everyone to assess their new situation (slightly faster than Web 2.0 enterprise adoption) and bang – Web 3.0 is a competitive differentiator.

It is tempting to suggest that performance will become a non-issue due to better future technology (and network is more of a bottleneck than database for many applications) but the reality is the information explosion is so potent that it will broadly compete with Moore’s law to keep performance in play. Poor performance simply blocks enterprise adoption right now. There are creative solutions mooted e.g. swarm based reasoners. Similar high concept solutions are likely necessary since the performance gap is huge.

Ambiguity removal is essentially what the Semantic Web is all about. Objections around whether the world can be generally mapped or not are valid concerns; they can certainly be well illustrated by examples showing literally stupid inferences that can be made. Such examples are on miniscule datasets though. With the real live Internet (as with all complex systems) - outliers even-out at scale.

It is easy to imagine a massively connected system refining ontologies in real time based on user input (although certainly people would need some carrot to do this). It is less easy to imagine this happening for reasoning but a great algorithm could well emerge that will simply not be questioned (or even understood by the majority of the population) as it just works most of the time and anyway; when it doesn’t, you can’t really tell e.g. PageRank.

Tool interest will be stimulated once the inhibitors above start to be addressed. Tangential to this, a general consumer focus on their personal data will mean organisations are compelled to improve their customer data in order to meet new consumer driven activity (personal profile, reputation management).

The semantics point is ultimately an artefact of Semantic Web technologies not being commercialized. Once the enterprise gets involved, efficiency will drive out any hint of academic ownership. NLP is actually complementary to the Semantic Web since it is a front-end to allow access to the semantic data.

None of the inhibitors above are outright Semantic Web deal breakers. It is not conceptually flawed. If RDF/OWL or SPARQL have implementation requirements (Grouping, BI/aggregation, lineage etc.), they can change; they are still evolving. Collectively though, the inhibitors are assuredly deal breakers for widespread adoption in the five to ten year range. Before that time, as it seeps gradually into collective consciousness, performance and reasoning visibility will likely become the main inhibitors. It is not an all-or-nothing idea though. A little semantics goes a long way as they say. Next post will explore how elements of the Semantic Web can be utilised now.

Semantic Web - Part 1 (What is it?)

The science fiction future that futurologists love predicting is some way away because meaningful data is currently not well integrated. You cannot have anything like the rampant “Imagine...” scenarios (typically using “Agents” and RFIDs) that futurologists speak to on a wide scale until (in terms of data) a Single Version of Truth (SVOT) is agreed upon for the data we are looking at (or disambiguation as it is called) and it is integrated wholesale.

Many of these scenarios will happen since they portrait a more efficient, opportunity filled or simply fun lifestyle; one enabled by information. Someone in the future will find a way to monetize it (maybe in a Minority Report/advertising kind of way) because market forces always apply.

Let us try one out in the next paragraph to illustrate:

Imagine you are in a bar. A distributed agent *knows* your location through your mobile. It also *knows* that the CEO of a large company in your industry is in the same bar and that that company is hiring for a role one level above your current one. Of course, it *knows* that you have been in your current role for a while. It does not *know* whether the CEO has any people with him at the moment (or is waiting for them) as he has restricted this information by his personal privacy settings. The agent suggests though (by texting or calling you) that it is worth you going up to the CEO and introducing yourself but not before it has already emailed him your resume and informed you that his drink of choice is a dry Martini.

This scenario is by turns – fantastically cool, moderately disturbing, highly efficient, opportunistically enabling and culturally changing. Some version of it will likely happen. What it is not is - artificially intelligent. Aside from the technology existing to do it all right now (mine GPS location data in real-time to determine matches against pre-set scenarios e.g. connecting people for jobs then checking social encounter rules and privacy settings) and the rule-based logic involved is straightforward (if people in same location and same industry and networking opportunity exists then...) the fact that all data required to fulfil this scenario is in different formats and in any case – secured (since there is little reason for the owners to share) will prevent our vista from happening. A lesser inhibitor is the rule-based logic - straightforward certainly; but the types of scenarios we are talking about require a lot of rules and it is unclear who will maintain them.

Future agent does not *know* anything, it has simply traversed the Internet (using look-up tables or schemas) to find an SVOT (because data is well integrated your location is stored disambiguously) and acted upon them as directed by predefined rule-based logic. Basically it has acted like any program around today (but on better data).

To fully integrate data you need senders and receivers to agree on a standard for both storage and communication (storing data in one format and communicating it in another defeats the purpose of data integration). This standard needs to be simple (since we also want to exchange data with mobile and embedded devices and generally want rapid and broad diffusion of the format) and not restrict others from building niche and more complex standards on top. The simplest standard is - a fact (Sales, personal etc.). Facts – of course, are condensed down to, something that the fact is about (the subject), something about the subject (the predicate) and something that is related to the subject through the predicate (the object). Examples are:

You (subject) located in (object) bar (predicate)
bar (subject) place of (object) socializing (predicate)

You cannot decompose facts any more than this otherwise; they would not tell us anything. It is conceptually akin to storing data at a physical level as 0s and 1s. Any type of information in the world can ultimately be stored as a list of linked facts in this way.

I (subject) purchase (predicate) Jack Daniels (object)

What is missing here is (you might ask) the timestamp and location; don’t we have to add them as columns four and five? Surely our future scenario needs that information? No – the idea is that we stick with the simple triple representation and it becomes:

I (subject) purchase (predicate) Jack Daniels (object)
My purchase (subject) was timed at (predicate) 1430HRS (object)
My purchase (subject) was located at (predicate) O'Malley's bar

While it is certainly true that much rule-based logic will always be required to fulfil the type of scenarios above, the amount of it is significantly reduced by the ability to make inferences using facts. Consider our facts:

You (subject) located in (object) bar (predicate)

This fact is automatically generated by your phone broadcasting its GPS location and it being matched to its location as commercial premises and finally cross-referenced against its business type.

bar (subject) place of (object) socializing (predicate)

This is a core fact that never changes and was created by some internationally accepted community effort. Because we have a like-term (bar), we can now infer that:

You (subject) are currently (object) socializing (predicate).

You have not specifically told anyone that you are socializing. It has not been encoded. Indeed it may be the middle of the afternoon so, in lieu of further information anyone may have otherwise assumed you were at work. We could have built-in rule-based logic to achieve the same result (if you are in a bar then you are socialising) but we have been saved the trouble by inference. Performance has been maintained as inference was in memory. This type of logic – syllogism has been around since at least Ancient Greeks. The implicit knowledge that both you and the CEO are physically in the same informal situation at the same time allows an opportunistic suggestion to be made; opening and closing a loop in real-time without having written a rule for it.

If everyone used the fact format (and its inferencing - managed by reasoners) for data storage and communication then we should all be able to resign from our jobs and hang-out in bars; secure in the knowledge that we have reached a technological plateau and an agent will at some point fix us up with a new role. Imagine.

The existing Internet is still very page focussed. You go to a job search site and have to trawl through pages of job descriptions, applying your own experience to decide which ones are interesting e.g. is a “Sales Executive” the same as a “Business Development Executive”? Does that term differ by industry category? If so, should I follow these links instead? You have to do a lot of work to find the things you want. So much so that you either give-up or end up with things that you don’t want. Using the fact format at Internet scale with disambiguation removes the necessity for humans to contextualise the data and so enables machines to better process it; which in-turn leads to more pervasive automation and those Imagine scenarios. This is what is meant by the Semantic Web (Web 3.0/SemWeb).

Is the Semantic Web inevitable? Opinion is divided to say the least. The Imagine scenarios (or variations of them) are inevitable. They absolutely require disambiguation and wholesale data integration. This in-turn has to necessitate a standard for storage and communication of facts. The Semantic Web is an attempt (the only one in town) to deliver that fact standard. Inferencing must be considered an optional component of the Semantic Web. It may uncover previously unknown information or simply be required to make it work at scale. It may require too much effort to be practicable for many due to its current reliance on an Open World Assumption (OWA).

The core promise of the Semantic Web is - disambiguation and wholesale data integration (Linked Data). It is the primary enabler for data-mashups. There are certain parallels with the early days of the Object Orientated Programming (OOP) movement. The Semantic Web is inevitable but it won't be called such. It will still be the Web.

There is an established fact format right now – RDF (Resource Descriptor Framework). Much of the supporting systems are also in place e.g. query and descriptor languages (SPARQL and OWL respectively). They have been quietly developed over the last eight years or so and all focus around the core premise of the simple fact (or triple as it is known [subject/predicate/object]). Next post will explore why we have yet to see widespread adoption of these technologies.

UPDATE: Great source of further reading links here.

Where did those mash-up tools go?

Just eighteen months ago, mash-up tools were big. The promise of building applications quickly, with minimal development and with context directly reflected within the application (as they are made by SMEs rather than IT resources) remains appealing. They looked to be the perfect tool for civic activists and knowledge workers alike. They were high in Garner’s top ten technologies to watch. The enterprise was starting to take them seriously as a mechanism to reduce crippling data integration challenges and consumers - bored on a diet of pushing links - thought they would be fun and/or a showcase for themselves in much the same way as blogs have been.

Now, the wind has shifted and MSFT’s Popfly and Google’s Mash-up Editor are both gone. Other niche vendors e.g. Sprout Builder similarly disappeared. Of the big players, Intel Mash-up Maker and Yahoo Pipes continue. All of them attempt/have attempted to straddle the void between consumer and enterprise spaces. This is an important distinction since – mash-ups even within the enterprise rely upon an ad-hoc passionate approach rather than formal development. They are typically built at home by passionate non-programmers who want to invest time in a single non-niche tool so whatever they learn is portable (work and social/other organisations etc). Even more so than blogging (which also uses one tool across both domains) a single tool is required - as more learning investment is required. With the exception of SAP’s Visual Composer (if you are an SAP shop), none of the tools mentioned have been particularly successful in either space let alone both. Why is this?

1) No UX standards. For both enterprise and consumer, there are no standard for widgets (or gadgets or web-parts or whatever else you call discrete self-contained UX functions). There are standards for business cards (vCard) – why not widgets?
2) Slow linked open data adoption. More of a consumer inhibitor at the moment. Linked Data is a core component of the Semantic Web vision that uses a specific set of current technologies. It provides a way to readily mix-and-match data in a meaningful way and so is a key enabling technology for mash-ups. Sig.ma is a simple mash-up tool for RDF data. Unlike other data-based mashups which tend to be query-based, Sig.ma is search-based. You enter a search term, the search engine gets your data, you remove the bits that are not relevant and (if you like) re-publish the data again as RDF (or other formats). This is perhaps too simplistic for users right now but it is evolving and could become a potent research tool. Any mash-ups that rely upon open linked data (ideally the best data of all) suffer from a lack of it; although this is changing as Government initiatives in particular publish RDF data for transparency reasons.
3) Insular data integration. Although the various flavours (SOA/ETL/EII/EAI) have been core CIO agenda topics for over five years, they have been mainly confined to the particular enterprise itself; especially in the narrow form of web services and have been of limited success even there. Extranet take-up, where common data is shared between parties in the supply chain, has been leisurely and this is precisely where mash-ups are needed. Very few organisations treat meta-data with the same focus as data. This means, mash-ups have trouble vouching for data currency and lineage which detracts from user take-on. It is possible that the solution to data integration will be the Semantic Web and a greater openness of organisations to share data. If so, we will be waiting some years yet.
4) Industry standards. Have been slow to be adopted. A notable exception here is XBRL for common reporting.
5) SSO. This is not exactly an issue for consumer mash-ups; assuming you are using open linked data but it is still a huge inhibitor to data integration for many organisations.
6) Blogging comparison. Although superficially similar to blogging - mashing (if you are going to do it properly) requires a thorough data understanding and a lot more effort than committing stream-of-conscious thoughts before they float away into the ether (or linking to other peoples work). Blogging is simply an easier way to achieve microcelebrity and also - because the majority of posts are written in the first-person (I think...) they can be defended (if need be) by the simple statement - "These are my opinions". This is a segueway into a whole minefield of philosophy, politics and culture that is best left alone. People mainly do. Only a small proportion will directly challenge someone's written thoughts. Publishing a mashup however - where you are vouching for the legitimacy of the data opens you up to direct challenge (people may have provably better data) and so people resist it. Only when the number of single versions of truth in the world become smaller and more consolidated will this situation change.

Popfly showed early promise as a learning tool but never really got past being a Silverlight showcase. Its focus was on looking good (geo mash-ups and slick drag/drop) rather than data integration. It did not use RDF at all. MSFT have been slow to utilise semantic web approaches in general. Some of their media management technologies use RDF in the background but their main focus has been around semantic search through the Semantic Engine. This initiative uses recently acquired Powerset technologies and will be released through SQL Server. PowerPivot has been significantly downsized from its original Project Gemini remit which would have provided not just a potent reporting mash-up environment but the management and support processes and infrastructure to QA and promote the mash-ups throughout the enterprise. This latter point is a key inhibitor to mash-up growth in general.

There are still signs of life in the mash-up tool space. Dapper is advertising focused. NetVibes is portal focussed. Alchemy API takes a content management/annotation approach (similar to Intel Mash-up Maker). Birst takes an analytic portal approach. Jackbe looks interesting; it appears to take a sales analytics approach. Snaplogic is not exactly a mash-up tool but it certainly takes a non-technical approach to data integration. None of them really play in that sweet-spot between enterprise and consumer though.

The parallel economic downturn has influenced mash-up take-on. The enterprise essentially stopped unproven development and consumers have yet to be sold on the concept but let’s face it – the focus on one or two drop-downs for configuration and Google Maps didn’t help either. The future of mash-ups is secure because it is the future of building useful applications quickly by SMEs and that will always be desirable. Five back-end data sources (database, RSS etc.) linked to five middle-ware components (aggregation, integration etc.) and five front-end components (analytics, data entry etc.) generates 125 possible combinations of application straight off the bat. Adding tailoring through filtering, personalisation and general configuration takes in into the thousands. This simple logic guarantees a future at least in the enterprise.

Whether the name “mash-up” has been tainted by its recent hiatus and – like its raison d’etre will need to resurface as part of something new remains to be seen. We can be sure though that the next generation of mash-up activity needs to be three things in order to stick around; interactive, data focused and usable in both enterprise and consumer spaces.