Complex or Complicated Solutions – a eGov issue

This started with my preparing a plenary keynote for the World Congress of IT, or WCIT, http://www.wcit2010.com/ due to be delivered on the 25th May. The theme calls for a new partnership between Government and Business in terms of the provisioning and use of technology in the creation of the emerging society we see today. Its long been a theme of mine at Government, or EU, events that what we need to focus on is not eGovernment, but eCitizens. Its not about delivering the IT processes of Government today over the Web with citizen access, but rethinking how a citizen will want to run their relationship with their elected Government. In short; ‘Government for the people’, though not as the saying continues, ‘by the people’, that seems a step too far currently!

Another way of putting this is ‘User driven services’ which makes it recognisable as the old alignment cry between business and IT. However the contrast between eCitizens and eGovernment is a particularly easy one to see as a use case for the complexity and scale of the change we are approaching. An eCitizen is going to expect a unique and individual outcome to their requirement that combines any number of government departments and services (meaning business capabilities) together as a delivery. I.e. a severe illness in the family requiring care cumulating in death with all the complex issues of registration, tax, inheritance, etc. By contrast a government will be driven by auditable processes each separate in administration and consequences, which, if mixed might produce wholly unforeseen tax or social payment outcomes to say nothing of muddled responsibilities.

Obvious society is complex and therefore will have complex systems, a topic that has, and continues to occupy a great deal of academic research time, http://css.csregistry.org/tiki-index.php. The most famous quote about Complex Systems comes from Aristole who said that “The whole is more than the sum of its parts”. Complex systems are systems where the collective behaviour of their parts entails emergence of properties that can hardly, if not at all, be inferred from properties of the parts. Well that’s probably what a Politian would like to think that they can create with eGovernment, or a CEO wants to get out of their business through leveraging core competencies. So in both cases the requirement top down is the same, and the expectation is that technology is going to provide the answer via the IT department.

Enterprise Architecture, and EAI middleware is not likely to be the answer to this, instead we should be looking at provisioning through using the granularity of ‘services’ as opposed to the monolithic approach of Applications. Actually its not one or the other, its both used together. The goal is introduce an abstraction layer between the core processes represented by applications connected together through closed coupled Middleware in defined relationship, and the loose coupled environment of services with the flexibility of orchestrations. At least in part the ability of drag and drop tools to produce orchestrations ‘on-demand’ starts the change towards User driven views on ‘outcomes’ that suit them rather than the computers database.

In reality we are talking about being driven by solutions and that in turn means solution architecture, but what does that mean in this new environment? For me this introduces complicated models, and principles, which is wholly different to complex models. The Cynefin Framework http://en.wikipedia.org/wiki/Cynefin offers a method of understanding and analysing complicated issues. Started in 1999 as a way of relating knowledge management to business use it has been extensively developed over the years by both business schools and leading technology providers including IBM, Microsoft, and Oracle. (I recommend Architects who want to start to think about this new area to read it), states; Complicated systems are defined as those in which the relationship between cause and effect requires analysis, and/or the application of expert knowledge with an approach of sense, analyse, and respond to create good practice. In contrast Complex systems definition says that the relationship between cause and effect can only be perceived in retrospect, and not in advance, requiring probe, sense, respond to recognise emergent practice. (Think of it as the approach to a public beta release of software).

To get a good picture on this whole topic then I recommend Harold Jarche’s blog on ‘the collapse of complicated business models’, http://www.jarche.com/2010/04/the-collapse-of-complicated-business-models which is its self a reply to a Clay Shirky blog entitled ‘the collapse of complex business models’ http://www.shirky.com/weblog/2010/04/the-collapse-of-complex-business-models If all of this seems too much to be bothered with then can I remind you of the similar situation in the mid nineties and the calls for approach what we now know of as Enterprise Architecture? At the time this was scoffed as unnecessary and the founding father of modern enterprise architectural principles was seen as making complex theories that weren’t needed!

Leave a Comment

33 Comments

Leave a Reply

Profile Photo Arvind Nigam

Awesome post @Andy. Infact cognitive science, anthropology and narrative patterns, as well as evolutionary psychology are very closely related to a branch of Computer Science which has seen deposition of several successful layers of advancements in the last fifty years!

I am aware of several rather obscure organizations out there that are close on edge of creating really impacting technology to handle complex situations i.e. where connection between context(cause) and effects are not easily understood (mapped). Intuitive technology is all about this. Fundamentally, there are many “learning algorithms” that dole out near-100%-accurate solutions in a probabilistic manner and help humans solve complex problems to a great extent.

Climate prediction for example.

I do not want to sound all jargon, but there is a silver lining of developments which fit your blog case immediately. It’s just about connecting the dots and making these things visible.

Cheers,
Arvind
http://bubbleideas.com

Profile Photo andy mulholland

as you say Arvind its been a ‘edge’ type subject but we seem to be getting closer to making these types of theorectical sciences practical to use and apply with new products/tools.

Profile Photo Srinidhi Boray

Enterprise Architecture is designed to be delivering along following business areas

Services for Citizens
Mode of Delivery
Support Delivery of Services
Management of Government Resources

With “services to the citizens”being mission critical.

I fail to grasp what you are attempting to discuss. Are you talking some solution architecture, tight coupled, loose coupled, fine grained, coarse grained and inaccurately alluding to Enterprise Architecture, that is more holistic.

Also in your statement

“In contrast Complex systems definition says that the relationship between cause and effect can only be perceived in retrospect, and not in advance, requiring probe, sense, respond to recognise emergent practice. (Think of it as the approach to a public beta release of software).”

Are you suggesting that all planning vehicles are working in hindsight so not necessary. I fear you are making statements calling for not creating emphasis on “line of sight” that IT spend have to demonstrate to ensure that correct agency motivations have been addressed, so citizens get the best benefits.

Most thing I hazard is “transparency” and “accountability” will get swept away allowing more room for obfuscation. The reasoning in your proposal rather simplistic in assumptions is quite ill founded.
It does not address any concerns of Clinger Cohen, OMB – Circular A-130. It does not consider the dynamics of capital planning and investment control, program management, issues in complex acquisition etc.

Very Sorry!!

Profile Photo Bill Brantley

@Andy – “Enterprise Architecture, and EAI middleware is not likely to be the answer to this, instead we should be looking at provisioning through using the granularity of ‘services’ as opposed to the monolithic approach of Applications. . . .”

Couldn’t agree more. EAI (as it usually described) seems like another iteration of Management Information Systems and Enterprise Resource Planning system which pushes a top-down view of the organization onto the organization’s processes. These views are abstractions and do not fully capture the dynamic complexity of the organization’s processes.

Using drag-and-drop services that are created by users better models the actual workings of the organization and how the organization interacts with the environment. Realizing that public agencies work more like the “garbage can model” calls for cloud computing, open data feeds, and apps that can be fitted together as the model for an agency’s IT infrastructure.

@Arvind – What organizations are you referring to? I ask because I have an interest in the fields you list.

Profile Photo Srinidhi Boray

Enterprise Architecture is a subject and a discipline and EAI is a middle ware solution. They are completely different things.

Guys!!! I am flummoxed. Probably author has confused what is EA and its application in the Fed OMB planning 🙂

Profile Photo andy mulholland

Hi guys – the debate looks to be hitting the point namely is what we do today with its current definitions correct for what we need to be able to deliver tomorrow? And in so far as the basic definition that most people would apply to EA and EAI would relate to state-full close coupled applications whereas the definition of EA in terms of defining business requirements can still be seen as relevant but we have to rethink the methods for stateless, loose coupled services.

Profile Photo Bill Brantley

@Andy – Exactly! IMHO, too often you have consultants who come up the latest three-letter acronym (TLA) with the associated propriety software solution. In order to sell their latest offering, they either redefine the organization’s issues as falling under the TLA or convince the decision makers that they have issues that needs the TLA and software solution. That’s why I don’t sweat definitions.

The better consultant is one who surveys the organization and crafts a homegrown solution from the assets of the organization.

Profile Photo Srinidhi Boray

@ Andy, Agreed these state-less services (SOA Portfolio) is a good thing to overcome redundancies. These things have been spoken and are being spoken, including much emphasis on Cloud and is a mandate these days. In the Federal sector EA has a specific role and they do allow for services to exist. But that is a different argument than EA being restrictive etc..

In Fed EA is a planning vehicle driven by budget and capital planning.

http://www.whitehouse.gov/omb/e-gov/fea/

Profile Photo Srinidhi Boray

@ Bill, you completely loose me 🙂 I must chuck my certification that the Feds created for realizing Clinger Cohen and for planning IT architectures. Or, I am out of my mind. Let me settle for the lesser reality, although both as hazardous.

Profile Photo andy mulholland

i dont think you are losing your mind ! as long as we stick to the current role of EA in connection with IT systems that we know today. The challenge is that Business Technology – and using the term deliberately as Forrester introduced it to indicate something different from IT – isnt the same thing. Its like applying Mini computer developement to PCs using Client-Server technology, a radical change of direction in what technology is being used for and in the technology being used. Bizzarely at the time the term IT was introduced to seperate it from Computing for the same reasons!

Profile Photo andy mulholland

Absolutely Srinidhi !!! thats the challenge – as long as we apply a client server state-full close coupled methodology and mentality to a web based stateless loose coupled world we dont get the point!! Is that easy to do? certainly not after years of training and developing expertise and methods but last time it happened it caused the death of the mini computer technology and the replacement with PCs using Client- Server, this time ….. well scale is endless, devices of all types, connections to anything and everything – and of course the cloud, all based on services and not applications. its pretty terrifying for many technologists and that includes myself in working out this new world!

Profile Photo Srinidhi Boray

@ Andy, even with stateless services and remote provisioning etc, one cannot keep the discussions such as “economy of scale” and related cost considerations away. Just mere stateless or state-full conversation does not complete the thinking around the “problem” and “solution” domain.

Stateless and state-full and such thoughtful discussion / implementation is an ongoing discussion to reduce redundancies and also avoidance of upfront capital cost and related IT operational headaches. The issue is how do these demonstrate that they fulfilled the strategic outcomes. Was IT spend thought-less or thought-full. Especially, I will be concerned when it is tax payers money.

I suggest you explore more about EA. At this time whatever you are trying to allude to EA does not make sense, and it seems you are confusing different subjects, terms etc.

I agree taking out application and bringing in virtualized services (SaaS, PaaS, IaaS) are all great thing. But show me the business argument. Are you developing these services on your own cost and then I am paying you monthly subscription and also need not own and maintain IT infrastructure.

Profile Photo andy mulholland

may be i am – or maybe the accepted definition of what EA means today might be the challenge? the European Architects Forum last may/june and the Open Group are both battling with this point. I really dont understand why you try to relate redundancies and capital to the issue? the whole point is that the shift is about addressing NEW areas with technology that current IT based technology cant address hence why the shift to calling it Business Technology instead – see forrester,gartner etc.

Profile Photo andy mulholland

oh and btw – on Tuesday 25th May the World Congress on IT addresses this point with Nellie Groes the head of the EU e Government making the opening keynote and you will see my reply on behalf of the industry as the followon plenary speaker. its all about changing from egovernment around pushing out from existing IT systems to eCitizen orchestrating their individual outcomes.

Profile Photo Srinidhi Boray

This idea is good, there is absolutely no doubt. eGov as remotely provisioned portfolio of services is a great upcoming thing. But one cannot discount the fiscal planning for IT modernization, this s the government’s main concern, how to save and how to pay.

All the best for your talk. Yes, they will revolutionize the idea of IT in future.

Profile Photo Bill Brantley

@Andy – Let me go back to what I think is your original question. Agencies have their services and data aligned around around their data model of the agency thus giving an agency-centric view of the agency’s services. What citizens need is the ability to pull those services from different agencies to craft their own consolidated and unique view.

I think one solution to this is to have agencies create their applications using a model-view-controller framework (http://en.wikipedia.org/wiki/Model_view_controller) with the goal of building in application programming interfaces (http://en.wikipedia.org/wiki/Api). A citizen can use a tool like Yahoo! Pipes (http://en.wikipedia.org/wiki/Yahoo_pipes) to graphically put the data feeds together into a personalized view. Data.gov is a good first step and what should be next is Pipes.gov which is the graphical drag and drop tool you talk about.

Profile Photo andy mulholland

okay Bill wins with the best summary !!! 🙂
my only build is the thought i have heard from other places to see egov as a Platform from which the ecitizen services as you describe can operate.

Profile Photo Srinidhi Boray

@Bill – “What citizens need is the ability to pull those services from different agencies to craft their own consolidated and unique view.”

And these magically will manifest. Someone is working from a magic hat 🙂

Profile Photo Bill Brantley

@Srinidhi – APIs are a well-established technology as the business success of 37Signals quite clearly proves.

And you might want to check out Yahoo! Pipes to see how users are combining views from different data feeds. Pipes has been around for five years and people have created some amazing applications.

You might want to read Hinchcliffe’s Web-Oriented Architecture – http://hinchcliffe.org/archive/2009/12/14/18179.aspx. IMHO, this is a better alternative to EA.

Profile Photo Srinidhi Boray

@Bill – My gosh, democratization of data and services are different things !!??

So what is in your definition EA. What is CCA, Circular – A130.

Its now time for mad tea party 🙂

Profile Photo andy mulholland

Just one final ppint from me – Srinidhi is right we cant discount the money side. one of the keys to the cloud revolution is that unlike classic IT which is an overhead funded by annual budgets as a cost of operation – ie centralised. Edge based services from the cloud are user pays based in a model that suits decentralised personalisation. that is a key change element!

Profile Photo Srinidhi Boray

@Andy, Yes that is the point. To me cloud iin theory s not revolutionary. Since early 2000 – ASP days, ideas like this has existed. Remember the famous Exodus. I have known many companies who used external hosting. Yes, remote provisioning for entire IT resources has undergone some very good technological advents. To turn the govt / fed invested existing asset suddenly into a remote services needs some fiscal thinking. Want to do with the existing IT investment. Many of which crunch realtime data and transactions oriented. Mainframes still exists. Billions are spent to manage them. What about them.

Also, if I don’t want to buy a PC instead use it as a service, no one is giving a good argument, why i7 PC should not be bought, should govt want to invest in technology refresh.

Sun was among the first to try the Network Computer needing zero administration – introduced in 1996, java changed the things the way software could be ported and invoked from a remote device on TCP/IP, etc. Lack of bandwidth and so many other things killed the product.

http://tldp.org/HOWTO/JavaStation-HOWTO/whatischapter.html

Profile Photo Bill Brantley

@Srinidhi: Quoting from Circular A-130 – “Agencies must implement the EA consistent with following principles:
(i) Develop information systems that facilitate interoperability, application portability, and scalability of electronic applications across networks of heterogeneous hardware, software, and telecommunications platforms;
(ii) Meet information technology needs through cost effective intra-agency and interagency sharing, before acquiring new information technology resources; and
(iii) Establish a level of security for all information systems that is commensurate to the risk and magnitude of the harm resulting from the loss, misuse, unauthorized access to, or modification of the information stored or flowing through these systems.”

To me, the ideas of a cloud and web-oriented architecture fit quite well in this definition. Nothing wrong with EA or Circular A-130 in theory. But, in practical experience, I have seen EA used as a vendor strategy to silo information resources, force agencies to buy bloatware, and build user-hostile systems. I’m not saying you are that kind of vendor but your comments are similar to other EA vendors that try to convince me that open-source, open data-feeds, and mash-ups will not work in public agencies. You make a good point about the existing IT infrastructure but being tied to a sunk cost is a logical fallacy. Government IT has to change and we have an administration that realizes that and is willing to make the transition.

You might want to read Chapter 1 – A Peace Corps for Programmers in O’Reilly’s Open Government book (http://cdn.oreilly.com/oreilly/booksamplers/9780596804350-sampler.pdf) which has a similar view to mine of traditional EA solutions.

Profile Photo Bill Brantley

@Andy – I apologize for continuing the rather unproductive tangent on EA. You made some great points (especially paragraph four and the final paragraph) and I think your focus on eCitizens instead of eGov is a great concept. Will you be posting your keynote speech at GovLoop?

@Srinidhi – I give up. Apparently in my 20+ years in IT with extensive experience in programming desktop and web applications plus a PhD in Public Policy and Management with a concentration in Knowledge Management along with an MBA in Project Management, I must have missed learning about this particular version of EA you speak of. I shall suffer in my ignorance.

Profile Photo Srinidhi Boray

@Bill – You got an awesome list of credentials, but yes you certainly missed this one.

Hope your are not blogging on office time 🙂

http://www.whitehouse.gov/omb/circulars_a130_a130trans4/

(2) The Enterprise Architecture

Agencies must document and submit their initial EA to OMB. Agencies must submit updates when significant changes to the Enterprise Architecture occur.

(a) What is the Enterprise Architecture?
An EA is the explicit description and documentation of the current and desired relationships among business and management processes and information technology. It describes the “current architecture” and “target architecture” to include the rules and standards and systems life cycle information to optimize and maintain the environment which the agency wishes to create and maintain by managing its IT portfolio. The EA must also provide a strategy that will enable the agency to support its current state and also act as the roadmap for transition to its target environment. These transition processes will include an agency’s capital planning and investment control processes, agency EA planning processes, and agency systems life cycle methodologies. The EA will define principles and goals and set direction on such issues as the promotion of interoperability, open systems, public access, compliance with GPEA, end user satisfaction, and IT security. The agency must support the EA with a complete inventory of agency information resources, including personnel, equipment, and funds devoted to information resources management and information technology, at an appropriate level of detail. Agencies must implement the EA consistent with following principles:

Profile Photo Gary Berg-Cross

The orignal post was indeed interesting and has generate several follow on discussions. I might like to explore the eCitizen idea more, but will make a point here about the main response thread on EA etc.. Complicated systems were contrasted with Complex systems definition and these were defined in the Cynefin Framework. You can see a discussion of all of this in the IBM article “The new dynamics of strategy: Sense-making in a complex and
complicated world” KURTZ AND SNOWDEN IBM SYSTEMS JOURNAL, VOL 42, NO 3, 2003

. I think that their take on EA might be as a modeling tool for part of the Ordered domain where their are “knowable” causes and effects. Unfortunately, in my opinion, not all parts of enterprises fit into this domain and our EA models are thus inadequte tools to guide us on the enterprises endeavor. If the EA modelers are wise (a minority status) they will focus on stable cause and effect relationships that exist in the enterprise domain.

But as the authors note such:

“stability may not be fully known, or they may be known only by a limited group of people. In general, relationships are separated over time and space in chains that are difficult to fully understand.
Everything in this domain is capable of movement to the known domain. The only issue is whether we can afford the time and resources to move from the knowable to the known; in general, we cannot and instead rely on expert opinion, which in turn creates a key dependency on trust between expert advisor and decision maker. ”

They go on to speak also about things that apply broadly, I believe, to the family of methods that EA and its methods fit into:

“This is the domain of systems thinking, the learning organization, and the adaptive
enterprise, all of which are too often confused with complexity theory… In the knowable domain,
experiment, expert opinion, fact-finding, and scenario- planning are appropriate. This is the domain of methodology, which seeks to identify cause-effect relationships through the study of properties which appear to be associated with qualities. For systems in which the patterns are relatively stable, this is both legitimate and desirable.
Our decision model here is to sense incoming data, analyze that data, and then respond in accordance with expert advice or interpretation of that analysis.
Structured techniques are desirable, but assumptions must be open to examination and challenge.

Profile Photo Srinidhi Boray

whether it is called eGov or eCitizen, the actor in context for who the systems are planned always has been the “citizen”, including BRM, the mission critical area has been “services to citizens”.

Cynnefin is Systems Thinking based, that few have begun incorporating into EA discipline.

Sense / Response making in the complex system might be something of KM interest, where correlation architectures work, and might be useful to data related services, while “Services to Citizens” are not mere data alone and belong to a different paradigm although might be triggered by the events designed within Event Driven Architectures to bring into effect the “sense/respond” mechanism.

Mitre work on ESE – Enterprise Systems Engineering Framework

http://www.mitre.org/work/sepo/pdfs/collaborations_vol5_num1.pdf