It is scarcely an exaggeration to say that the problem of communication is the axial problem of accounting.

Then Haried goes on to say that one of the major contributing factors to the problems of external accounting communication is insufficient standardization of the terms used in financial reports. Fast forward to Houghton in 1998, the birth year of XBRL, and you have another perspective that supports the general need for and use of XBRL and the utility of an agreed taxonomy specifically:

…the study of shared meaning between users and producers of accounting information is central to the study of the effectiveness of accounting communication.

But as Paul Baier at GreenBiz points out (with a few spelling errors):

When reading traditional annual reports, financial analysis (sic) quickly skim pass (sic) the glossy pictures, platitudes, CEO letter and other marketing fluff and head straight to the financial numbers and footnotes.

Savvy CSR report readers do the same. The value is in the numbers.

So it’s also good to hear that the Global Reporting Initiative (GRI) and the Carbon Disclosure Project (CDP) are working even closer together to maintain an explanatory mapping between each other’s sustainability and emissions indicators. That way the text and the numbers in GRI Index reports are more easily compared with those in CDP emissions reports.

Of course what would be even better is if those numbers could be standardized by referencing a single XBRL sustainability taxonomy. Then ‘scope 1′ emissions in a GRI report would always mean the same as ‘scope 1′ emissions in a CDP report – no matter what actual stakeholder report the number ended up in.

• OpenlyLocal is a site that lists UK local government sites that are making their data openly available.
• Rewired State has worked on various data mashup projects for the UK’s National Hack the Government Day 2011
• Armchair Auditor details all spending over £500 by the UK Windsor and Maidenhead local council
• The Online Goverment (sic) Open Accountability Ledger (GOAL) project for truth in accounting in the USA
• Google’s transparency map detailing requests from government agencies to remove content from their servers
• Germany’s Offener Haushalt – opening up German government finance data
• The Open Budget Survey for transparency in national budgeting
• A NASCIO transparency PPT detailing various US State open data sharing initiatives e.g. Texas transparency, openAlabama.gov and Utah.gov
• Gov 2.0 post on open data initiatives including Sir Tim Berners-Lee’s 5 star rating system for open government data

My point? Wouldn’t it be great – for us at least – if all these initiatives were reporting using XBRL?

In the long history of humankind (and animal kind, too) those who learned to collaborate and improvise most effectively have prevailed.

But what’s interesting about this graphic’s perspective is that it focuses on internal daring to share, rather than external. Even the latest ‘collaboration’ step in the graphic is about enterprises replicating microblogging and activity streams internally. But our perspective on Dare to Share here is less informed by ‘intentional collaboration’ and more by ‘intentional transparency’ – that is daring to share externally.

Filing XBRL to the S.E.C. is an example of a relatively old kind of corporate event/activity stream – financial reporting – that is shared externally in a new way (albeit mandated and via a regulator). XBRL is the medium, the S.E.C. acts as the channel to share this data with consumers (e.g. through their RSS feed) and the message is a package of financial statements and disclosures. In other words, it’s a communication and because it is intended to be suitable for machine consumption, the potential scope of the sharing is effectively unlimited.

Instead of a document designed primarily for humans can read, that often requires manual ‘scraping’ to get the useful data back out of it, data shared in XBRL can be interrogated directly and mined by rule-driven engines. These information robots can then manufacture new kinds of output from the data by reaching into other datastores to combine and connect information to construct new information. If you dare to share externally in this way you are stepping into a new world that is less about people collaborating and more about programs collaborating.

Both ‘direct access’ and ‘unlimited scope’ are key to daring to share externally – and they are something that sharing in XBRL helps to facilitate. But you’ve got to admit it’s quite scary and makes the need for this kind of external sharing to be underpinned by formal taxonomies all the more understandable and necessary.

However,  just in case anyone is mystified… Triplification does not refer to triple filtered Irish whisky or triple filtered Stella Artois, the Belgian lager with the famously smooth outcome, but converting data into triples, one of the pillars of the semantic web. Personally I find a few lagers and whisky chasers quite useful when trying to get to the grips with anything to do with the semantic web, but then I’m no Roberto Garcia (didn’t someone make a movie called Bring me the head of Roberto Garcia?). Anyway I digress as I sit here among the remains of a Stella six pack and a half-empty bottle of Jameson.

Garcia’s XBRL company dataset included data from the filings of 543 companies. By identifying specific ‘connector’ data they managed to link their ‘triplicated’ XBRL dataset to other web-available datasets:

Using the company ticker they managed to link 64 companies in their XBRL dataset with company data on dbpedia.org.

Using the SEC CIK code they were able to link 398 companies in their XBRL dataset with data in Joshua Tauberer’s (now CTO of PopVox) US Securities and Exchange Commission Corporate Ownership RDF Data.

Garcia cites other potential ‘connectors’ like ISO 4217 currency codes for example, which could be used to link to exchange rate tables.

So what?

Company data on dbpedia.org and previous (non-XBRL) filing data on JT’s RDF data repository adds context to the XBRL data and starts to build a bigger picture around an XBRL filing. This in turn illustrates the promise of using ‘standardized’ data formats – in this case XBRL and RDF – as a basis for generating new or at least potentially interesting information from data that was previously largely ‘siloed’ in various places.

At the moment, very smart humans like Garcia are required to ‘join-up’ these datasets but you can see where this is going. Once more datasets are triplicated in ways like those proposed by Garcia, vast new dataspaces are established that analytical tools can roam at will. Automatically and without the help of lager.

Consequently, our opinion is that the best short term approach in order to get financial and business data to the Semantic Web is not to propose and alternative language based on Semantic Web technologies, but to apply methods to map existing XBRL to semantic metadata. This also seems the best option in the short and midterm to populate the Web of Data with business information.

While much of this article is about the technicalities of mapping XBRL concepts across to RDF/OWL (a ‘semantic web’ technology) it’s at section 3.1  where things start to get more interesting as the authors figure out what other data – preferably that has also been semantically processed – they can link to the XBRL data based on common ‘connectors’ like CIK numbers, ticket symbols, company or director names. They discuss not just new information that could be derived from an enriched ‘external’ content but also how semantic processing can help with exception reporting between ‘internal’ XBRL instances i.e. current vs. previous filings.

While semantic enrichment of XBRL data has many possibilities, the authors concede that certain aspects of XBRL, e.g. the calculation linkbases vital to processing and reporting financial data, mean that XBRL and semantic web tecnology (e.g. RDF/OWL) are in fact complementary:

XBRL can be used for business and financial data representation and validation, while its translation to Semantic Web technologies can be the way to make all this data publicly available enabling cross analysis of this data thanks to semantic integration and a graph base model.

The basic assumption of Linked Data is that the usefulness and value of data increases the more readily it can be accessed and recombined with other data

The central insight is that its not the documents that are important, but rather the things the documents are about that is important. …The context of documents, the topics, characteristics, sources and provenance, and especially the interlinking patterns between documents is the true foundation for discovery and assessment of useful resources on the Web…Linked Data is a way to make it just as easy for people to establish and share context on the Web as it was for them to originally share documents.

When it comes to transparency, for example, context is king. And data linking is a key way of delivering rich-context. By being based on data linking standards, documents can be liberated from their current role as content ‘silos’.

Here at Rivet, we agree with Hyland’s proposition that:

…there is a very unique opportunity at present for governments, non-profits and commercial organizations to spark a true revolution in how individuals discover and utilize valuable data.

And that as she says: The transition to being “in the Web” is occurring.

See my next Dare to Share post for how all this relates to XBRL.

House Oversight and Government Reform Committee Chairman Darrell Issa (R-Calif.) held a hearing in San Jose on Monday April 18  to get input from technology executives about running the government more efficiently. Rivet joined Google and Microsoft to share our perspective.

According to the Politico.com piece (see link above), the government’s flagship data sharing site (data.gov) came under attack as the data vs. information dilemma resurfaced:

“Data transparency initiatives give the impression that data is accessible,” said Patrick Quinlan, president of RivetSoftware, a Colorado company. What the public needs, he said, is access to the underlying datasets.

“More data is not the answer,” said Stuart McKee, Microsoft’s national technology officer for the public sector. “We’re producing more data and less information.”

I have to disagree with my (misspelt:) namesake. More data is good. That is how you make the connections that generate interesting and unexpected information. But what is not good is data that is provided in a mismash of non-standardized formats. To me the problem is not about more or less data or how to get quality information from that data – technology will find a way to store it efficiently and analyze it effectively.

The real problem is that Government data is not treated seriously enough as a vital communication asset by Governments themselves.

The purpose of every Government initiative is to execute policy. To do this they spend taxpayer money. Often, one of the activities that that taxpayer money pays for is the collection of data and one of the key deliverables of any initiative is the communication of that data. Because let’s be clear: The initiative is only the custodian of that data – it does not own it. Taxpayers own it.

In my view, the remit for every new Government initiative must include:

a) The recognition that the data collected must be communicated back to the owners.

b) That data should be communicated in a standardized way so that it is easy to connect it to other similarly collected and communicated data and help drive the creation of new information.

So every initiative should include, as part of its initiation ritual, the development of a ‘mini-taxonomy’ for the initiative so that its data can be shared and combined with and connected to data from other initiatives. Otherwise you end up with what many people, Rep. Jason Chaffetz (R-Utah) included, think of as a bunch of data ‘silos’, interesting to look at but ultimately not that useful (except to hide the smell of manure or store a missile that when released could destroy everything the initiative stands for).

Maybe Data.gov and similar initiatives in the UK and elsewhere are, as our President put it, at the ‘pony express stage’ of information communication. But providing an accessible, easy to use hub for one of the world’s largest data warehouses is always going to be a challenge and you have to start somewhere.

If you don’t dare to share, then you can’t expect to share in the information spoils. But you can expect to spend a great deal of your time and money dealing with things called ‘leaks’.

I applaud the people behind Data.gov and Recovery.org – they are daring to share. And yes, maybe today there’s more sharing than daring, more congratulating than calling to account, but that will come. They are trying to overcome one of the hidden foes of transparency – surface data availability as opposed to deep-dive data accessibility. But to make these sites more than window dressing, we need initiative-level XBRL taxonomies to kick-start the process of making all this data easier to consume, combine and connect. Not more ‘silo’ websites with their plethora of ‘silo’ charts and PDFs chewing up yet more taxpayer money.