Guess I could do with some help here…
In case you hadn’t noticed, I’ve been kinda pouring out the posts on enterprise-architecture and the like, over the past few weeks or so… (A few people have complained about the overload, and probably with good reason, too! 🙁 🙂 Oh well. My apologies, anyway.)
What’s happening for me is that it seems all of the work I’ve done on enterprise-architecture theory and practice over the past several years is suddenly coalescing into something big. It feels like I have a handle, or hook, or something, onto a radically different approach to how we do architecture-work in an enterprise-context, and in particular how we document that work and use that documentation to drive and support more-viable enterprise change.
The practical problem is that it’s all crashing around in my head, coming at me from all sorts of different metaphorical directions – metamodels, toolsets, methodologies, metaphors, the lot – and I’m having real difficulty getting it all down into a usable form. I tend to work solo most of the time, but in this case, frankly, it’s become way too large to handle all on my own. I can’t hold all of it in my head at once: it’s too big, too complex, needs way too many different skillsets and experiences, some of what I’m trying to do is likely way too complicated at present, and it’s almost certain that some is just plain wrong.
Hence, to quote the famous phrase, “I guess I could do with some help here…”
Please?
To give an idea of the scale of what I’m trying to bring together at present, there are around 300 posts on enterprise-architecture here on this website (excluding the weekly ‘A week in Tweets’ posts), of which perhaps half – maybe more – describe some new idea or concrete item of research on EA and related themes. There are eight finished full-size books on enterprise-architecture so far in the set, and at least another two or three under development at present. There’s another website just on metamodels, and another weblog on ‘big-picture’ business-themes. There are fragments of notes and experiments scattered around on seven different computers here, not to mention ten years’ worth of paper-notebooks and sketch-pads, and discussion-notes and emails from colleagues and conferences and the like over most of that time-period, too. And all of it seems to be coming together all at once. Yikes…
The real focus, as has come up in the most recent posts, seems to be about techniques and toolsets, to take any starting-point – such as a business-model in Business Model Canvas – and expand outward to tackle all of the whole-enterprise issues, including themes such as quality, security, sustainability, continuity and so on. Here are links to some of those posts:
- From business-model to enterprise-architecture
- Upwards and sideways from business-model
- Why business-model to enterprise-architecture?
- Where marketing meets enterprise-architecture
- The Enterprise Canvas – a really simple summary
There’s stuff about metamodels to move beyond ‘classic’ IT-centrism:
There’s stuff about Agile in relation to enterprise-architecture – in particular a metaphor of ‘backbone’ that I still haven’t seen discussed elsewhere:
- Agility needs a backbone
- Architecting the enterprise backbone
- Tackling uniqueness in enterprise-architecture
There’s also a real need to address the human side of enterprise-architecture, including architecture-as-story and the almost-unacknowledged issue of narrative-knowledge (as opposed to IT-based information) in enterprise-architecture:
- Two points of view on (enterprise)-architecture
- Modelling people in enterprise-architecture
- People, power and enterprise-architecture
- The relationship is the asset
- Where have all the good skills gone?
- Who are your anti-clients?
- The enterprise is the story
And perhaps the real core of what’s happening for me now, around toolsets to cover that whole scope:
- The toolset-ecosystem
- Next-generation toolsets for enterprise-architecture?
- Enabling enterprise-architecture conversations
So: that’s the range of topics that I’m struggling with at the moment. Because it’s so huge, different parts of it keep drifting in and out of cognisance at any given time, so it’s difficult to maintain a sense of the whole. And the problem is that it does need to be tackled as a whole if we’re to avoid the same kind of trap that earlier forms of EA fell into, first with IT-centrism and process-centrism, and now all too often with business-centrism. The only thing that will work is to create something – a methodology, a metamodel, and a toolset-ecosystem, all linked together – that will fully support the awareness that in enterprise-architecture, everywhere and nowhere is ‘the centre’, all at the same time.
To be honest, I need help in all aspects of getting this down into usable form. But where I most need help is around the overall toolset(s). I think that the ideas are just about ready now for a proof-of-concept: but I doubt that I still have the technical skill to do much if any of it on my own. (The last significant software project I developed on my own was a wiki-based ‘engine’ that was used for a number of system-prototypes such as the SEMPER diagnostic: but it was back in the days of PHP4,and way too crude for todays’ standards.) To cover the whole toolset-ecosystem we’ll need whatever-it-is to compatible in some way or other with all of this scope:
- large cross-enterprise repository-based formal EA system (like Troux Metis)
- team-driven repository-based formal modelling system (like BizzDesign)
- single-user formal modelling system (like Sparx or ArchiTool)
- free-form desktop, notebook and/or web-based modelling system (like Visio) but can do round-trip to formal modelling
- tablet-style with gestural interface (like a cross between Prezi and BMTBox app on iPad)
- handheld, mostly browsing, but also able to feed back updates/suggestions
and
- paper-based, whiteboard, driven from book or ebook, etc
So: if you’re interested in ‘pushing the envelope’ for EA, you’re keen to work on some kind of collaborative project, and have some expertise in system-prototyping, user-experience design, workflow, graphics or anything of that kind, please get in touch with me? Because, yeah, I could do with some real help here… 🙂
Thanks in advance, anyway.
Tom, I’m going to try to be uncharacteristically brief (sighs of relief all round).
I want to help you but I don’t think I can help you with what you say you want most. That’s a) because my skills in these areas are limited or outdated and b) because it’s not what drives me right now.
In your various recent posts (to the extent that I’ve been able to keep up with them) you’ve produced a lot of really useful insights into a different way (or ways) of looking at the enterprise, at enterprise architecture and at what EA is useful for. And these fit very well with the changes that are going on out there in the real (but also virtual) world. I think the last thing the world needs right now is another methodology. It needs instruction on and illustration of how to do the right things with the methods that are available to us – because the problem is (often) not the method but the user instructions. Even the better user instructions seem to be firmly anchored in a view of the enterprise which is increasingly irrelevant.
So I envisage a sort of catalogue of ideas, views, patterns (and anti-patterns). That can be a big book or a set of comic books or an accompaniment to a toolkit. And whilst a toolkit as such is not top of my priority list right now, if you were able to produce a toolkit which was inclusive rather than exclusive, which was descriptive rather than presscriptive and which had the look and feel you describe, you’d definitely be doing the world a big favour.
So I guess what I can offer, if you find it interesting and to the extent that I have time, would be a specification of what the toolkit should deliver (and avoid) – the kind of things I’d like to be able to do with it.
Any use?
Yes, definitely of use – yes please! 🙂
I agree with Stuart *emphatically and entirely*. I am fairly new to the scene and been doing a lot of the learning… I am finding that I need to catalog everything I’ve been reading because there is just so much. At least start with this… (I’m afraid a toolkit would be years in the making and suffer from incredible scope expansion.)
Principles, idealogies, views, patterns, anti-patterns (to build off what Stuart listed).
Tom,
I would need a considerable amount of time to go through your current ideas, approaches … (volume and quality is impressive) to build a complete picture of what you are trying to do and see where I could assist. You may have done this already, but perhaps a little of the old “eating our own dog food” from a vision, goals, model (business model canvas), challenges/opportunities, strategy, initiatives, projects… and road-mapping perspective may be required first to move forward in an optimal manner (and to manage that “crashing around in your head!” – although some crashing is good). Perhaps I can start helping there and then move into more “rubber meets the road” areas after that.
Tom,
Maybe my tubemapping idea can be of some help to visualize your storylines?
Tubemapping is a this moment an idea for a paper-based “technique” to build a story by starting to connect the story beats* into connected storylines and visualize those storylines as a tube map (subway map) based on the principles Harry Beck used when he designed the London Underground Map.
As my sketch shows I see tubemapping as a technique that can be used to structure ideas and sketches and that the tubemap forms the foundation for the story you want to tell. The sketch visualizes the continuous loop of doodling – noodling – idea generation – sketching – TUBEMAPPING – storytelling – feedback – doodling etc.
This sketch is also the basic structure for my Tubemapping the Enterprise book 🙂
*Reference story beats: http://rayannecarr.wordpress.com/2008/06/26/what-is-a-story-beat/
Hi Tom,
I don’t think you should apologise for the proliferation of posts that you’ve been authoring lately. Yes, there is a large amount of material to digest, but it’s instantly recognisable as quality work and that makes the effort of reading and understanding it worthwhile. Your blog is fast becoming my most frequently checked website, both for your posts and the quality of the comments from other learned professionals.
I’m putting my hat in the ring to help as much as I’m able. While I am very much in the early stages of learning about enterprise architecture and practically applying it, I have some skills which will be useful to your effort. My EA skills will continue to grow too, as I absorb and apply new ideas. The immediate help I can offer is composing workflows and other diagrams, producing and critiqing metamodels and contributing to the business architecture side of things. I can also do copy writing and editing as necessary.
Some type of web-based collaborative software will be necessary, I think, such as Basecamp or MangoSpring.
I agree with Bill Craig that “eating our own dog food” will be one of the best ways to start.
Let me know what I can do.
Anthony
@Joe G – Thanks, Joe.
Agreed, scope-creep is an inherent problem, and one we have to face right from the start, because by definition this whatever-it-is needs to be able to address any issue. Hence why I agree strongly with Stuart that it needs to be “inclusive rather than exclusive, … descriptive rather than prescriptive” and so on.
Will have to do another post on where I think this is leading, but I know there are a lot of people just starting to build a lot of loose collaborations around what Stuart has described, and you too: patterns, anti-patterns, principles, ideologies, views and so on. The practical problem is that, by the nature of the work, most of us are loners, or comfortable only in small groups – much like scriptwriters, in a way, and for much the same reasons.
On ‘eating our dog-food’, there’s a lot we could learn right now from the Open Source movement, in software. We need our equivalent of SourceForge – that shared repository that enables consistent access to an emergent variety of ideas and implementations at any level of ‘bakedness’. We need probably need our equivalent of Linux – a common ‘action-frame’ that has a consistent core but can be customised in any way necessary. We could usefully do with an equivalent of the LAMP stack – Linux, Apache, MySQL, Python/PHP – as a basis for ‘cookbook’ development and delivery. (Yeah, I know that LAMP is somewhat dated now, but the principle still makes solid sense.) And we urgently need some means to share structural information about architectures – hence my obsessing(?) over Archimate and the like.
I know I’m not the right person to do this. I’m no Linus Torvalds or (heaven forbid?) Richard Stallman: I’m painfully aware that I’m not good at project-management, and I don’t have the people-skills that are needed for cat-herding on this enormous scale. My skills at coding and interface-design were never wonderful, but are now hopelessly rusty. And, to be blunt, I’m probably too old for that game: I turn sixty in a couple weeks’ time. (That fact kinda brings its own urgency of a rather different kind… 😐 ) What I’m hoping for here is to at least start the ball rolling – hence over to you-plural on this?
Again, though, thanks for taking me seriously on this: it does help.
@Bill Craig and @Anthony Draffin – Bill, Anthony – yes, strongly agree that ‘eating our own dog-food’ needs to be the way to go.
One of the traps I’m struggling with is simplicity: it’s really easy to let this become too complicated to be usable at all! To me, Stuart has the right approach with his comment about comic-books, and likewise Peter with his idea on tube-mapping.
Over the next few days I’ll do yet another post (sorry…) that attempts to start off some of that ‘dog-food’: it’ll no doubt miss the mark by many a mile, but at least it should be something that we can (dis)agree with…
Again, though, what I really need right now is help with proof-of-concept prototyping for a toolset and underlying metalanguage that will cover the whole scope that we need. As I see it, we already have the core for the methods and tactics we need, with techniques such as design-thinking, systems-thinking, Dave Gray’s ‘Gamestorming’, Dan Roam’s ‘Back of the Napkin’ and so on; but the lack of any toolset to connect across that space is the one huge gap that we most urgently need to resolve. It’s true I’ve done a lot of work on this so far, but still not enough as yet to get to that essential proof-of-concept stage: any suggestions, please?
@Peter Bakker – As in my other responses above, I strongly agree with you about tube-mapping and the like.
(I took the liberty of updating your link there, to resolve a WordPress auto-conversion problem: I believe it points to the right place now, but perhaps please check? – thanks.)
And yes, also strongly agree with your point about story-beats (again, a great link – many thanks!). I do believe that a narrative-based approach to architecture is likely to be one of our most valuable techniques – and hence one that any whole-scope toolset needs to be able to support. More on that in future posts, anyway.
Thanks again, of course – very much appreciated. Let’s get this ball rolling, yes? 🙂
Tom,
After reading your ’emergency’ post… one word emerged in my mind: backbone. You need a backbone something that integrates as well as differentiates (ties together as well as sets free) at the same time the posts (and more) you mention. What is the Invariant, the Infrastructural ‘ingredient’ in them all? The wave on which all of them coherently ride?
Putting it in your own terms: “The only thing that will work is to create something – […] that will fully support the awareness that in enterprise-architecture, everywhere and nowhere is ‘the centre’, all at the same time.”
Continually creating living wholenesses by living centres. Centres coming into being, being there, shining, fainting away and disappearing again. Wholes creating centres; centres creating wholes. Source: The Nature of Order by Christopher Alexander.
(Your) Thinking mind cannot grasp ‘this’. Yet, you Know that ‘this’ is true. Coming to terms with ‘this’ requires you to skip and stop your dualistic mind/thinking for a while and quietlyand patiently wait for the answer to present itself to you. You have to let it all go, Tom… And, when you do, it will come back to you in unprecedented wholeness. The answer is already there. Only your thinking mind prevents you to really get (to) it.
Einstein used to take Light Creating Non Thinking Naps using two pebbles loosely held between the fingers of his two hands. When one/both of the pebbles would drop onto the (stone) floor… it would wake him up. One has to stay awake, but ‘nappy’ enough to stop/skip ones dualistic thinking.
Success!
@Tom G
Thanks for correcting the link, it works fine now 🙂
Do you now @SFDreverman on Twitter? He know an awfully lot about effective information sharing and he is working on a novel information sharing platform at http://www.infonotion.nl/blog-en/?page_id=6
He is also following Thomas van der Wal/Dave Gray’s The Connected Company initiative.
I think he should be able to help with setting up some kind of Sourceforge equivalent.
I will continue my tubemapping quest and will try to make some examples which relate to your Mapping the Enterprise book and recent blog posts 🙂
@Tom G
When it comes to this type of work, you have two main choices. To create your core engine from scratch, or to leverage an existing tool that allows you to create and execute your own metamodel. Obviously there are trade offs between these two options. The first offers the most flexibility, but you have to do a lot of legwork before getting any payoff. My suggestion is to go with the second, at least till you find it too constraining. My experience is with Holocentric Modeler and with Sparx Enterprise Architect. Of the two, Sparx makes its meta tools open and the cost to licence a seat is cheap less thant AU$300. It could be a good place to start. Have a read of this and see if you think it will be useful. The caveat is that you might find it overly UML based for your purposes. I’m interested to know if you think this might be a viable place to start.
@Jan van Til – Hi Jan: good point about the backbone, seems to be a theme worth developing in a lot more depth.
I wouldn’t actually call it an ’emergency’ post, more like an acknowledgement that I really can’t do this all on my own (or, more to the point, it’ won’t be very efficient or effective if I try to do so… 🙂 )
On ‘Your Thinking Mind cannot grasp this’, well, yes, strongly agree, of course – I literally wrote the book on that more than 35 years ago now, and you’ll see other variants on that theme in others of my books, such as ‘Inventing Reality’, ‘Disciplines of Dowsing’ and, of course, ‘Everyday Enterprise Architecture’. I’ve had about half a century of practice at holding an entire system in my head at once – it really isn’t a problem for me.
Getting it down into usable form that others can use, now, yes, that is a problem, and that’s the part I’m struggling with. In particular, what I’m really struggling with is the whole toolset space – all the trade-offs between different approaches to implementation, and the user-experiences themes that go with each implementation-approach. As I’ve said several times above, that’s the area where I most need help. If you have experience with iPad-style development, or web-based graphic-modelling, or user-experience design on different user-interface paradigms, or infographic design, that’s what I need right now. The rest isn’t all that important, until we have at least some kind of proof-of-concept to play with.
Thanks again, anyway.
@Peter Bakker – Peter, thanks for the link to InfoNotion, will definitely follow that up.
In the meantime, shall I happily nag you to get to work on your Tubemapping book? 🙂 – because it looks like it’s going to be a very useful one.
@Anthony Draffin – Anthony, thanks for the reminders about Holographic and Sparx.
When we did an evaluation of Holocentric about four or five years ago, the summary that we came up with was that it had by far the best ideas of any EA-toolset on offer in Australia at the time – the only one that really tackled the human-side of EA, for example – but unfortunately it had what seemed to be the worst implementation of all. 🙁 I haven’t heard anything about them since, and I’d be very pleased to hear that it’s improved in usability etc, because it really was a great concept. Any pointers/suggestions on that?
I was actually looking at Sparx earlier this morning: I’ve had an evaluation copy
on one of my machines, but unfortunately the licence has expired. I do know that it ‘s supposed to provide very good access to the metamodel, via a UML-profile editor, but I haven’t actually used it: how viable would it be for defining a metamodel from scratch? And also is the metamodel-editor available on the Desktop Edition? – the ‘compare editions’ page isn’t clear on that, unless it’s the ‘Create and Use MDG’ feature, which is only available as part of the Professional Edition upwards.
Another option would be Essential, I suppose – after all, it’s built directly on top of the Protege ontology editor. But I must admit that my previous struggles with Protege do not exactly fill me with enthusiasm… 🙁 – and Essential itself follows the same-old-same-old TOGAF-style ‘four architectures’, which means we would definitely be forced to start again from scratch there, whether we want to or not.
Good points, anyway – and thanks, of course.
Belt and braces logic leads me to post part of a response to a separate blog on this channel.
I ied previously to express what I saw as a limitation in the Business Model Canvas, which you correctly refuted. Those of us able to (literally in this case) think outside of the box don’t have a problem using the Canvas in the context of Extended Enterprise. You’ve provided some good examples of this. But the representational form is very much oriented to the single enterprise. That makes it susceptible to bad teaching and bad practice. That’s not the fault of the Canvas but it’s still a problem. I have tried with little success to create an ecosystem representation based on single enterprise iconography. Imagine a picture with a whole bunch of Canvases representing different participants in the ecosystem. Doesn’t work. With slightly more success I have produced models using a different form of iconography. But my visual skills are limited (to put it mildly). Plus those models could only support very limited semantics. I believe we need a representational form, which allows us to create multiple views. An ecosystem wide view with limited semantics would be fine if it could be seamlessly (sorry about that word) linked to more specific views with richer semantics. So we can have the big picture to establish context and we can have various views of the ecosystem seen from the individual enterprise perspective and catering for the different roles of that enterprise (and its partners) in the ecosystem. No need to try to capture all the semantics in one picture. No point in trying. And if the representation form changes, as you have suggested, going downwards towards solutions (IT or otherwise), that’s just fine as long as the semantics survive the transition.
For more info on the contextual background, see my comment on the “upwards and sideways” channel. Cheers.
Hi Tom, always glad to help, although this time it’s a really big pie to eat:) I must humbly admit I do not feel that I could actually add that much value as I am merely an apprentice here;) And also coding is not my strength for sooo long already.
Regarding tooling, I am a big fan of Prezi as it offers infinite “paper” to write on and ability to zoom in to details and zoom out for bigger pictures. It’s tremendously flexible for brainstorming around structures as one can move things easily around. For structured information I don’t really have a candidate, the only experience I have is with EA from Sparx (might not be that bad idea as one can quite easily develop plugins).
Anyway, count me in to your movement:)
@Tom G
I just made a little start on my weblog: http://peterbakker.wordpress.com/tubemapping/
🙂
@Tom G
Holocentric has decided to differentiate itself within the tools market and steer clear of vanilla UML modelling. Instead it concentrates on its proprietary process modelling environment which is quite good, but doesn’t conform to any standards. They are also concentrating on a metamodel enforced Accelerator, SR&F (Software Requirements and Fulfillment) which have some business architecture-like features, but never quite get there. Unfortunately they are loathe to make available their metamodelling toolset, even though it is very good. It took me several years before they finally gave me a three month trial key (which has long ago expired). This combined with the hefty price of $5000 a seat, makes it a non viable option. Their Corporate Governance Accelerator developed by their ex-CEO and founder, John Forrest (now a silent owner) is excellent. It combines facets of BMM, Balanced Scorecard and MSP. John might be ideal to attempt to pursuade to join this effort.
In regards to Sparx Enterprise Architect, the feature you’re looking for is as you’ve guessed, the Create and Use MDG feature. As I understand it, it is quite viable for creating a metamodel from scratch. Even though the tool is UML profile based, the tool can still be used to create stuff outside of UML. I believe they created the Archimate add-in and the Zachman add-in using this same feature.
@Stuart Boardman – re comment #16, yes, I replied to this on the ‘Upwards and Sideways’ post, in the comments section here.
@Ondrej Galik – “I am merely an apprentice here”: is true of all of us, I think? (and I’d suggest that it’s the ones who think they aren’t “merely an apprentice” that we should be wary about? ::wry-grin:: )
Yeah, I too am a great fan of the Prezi ‘zoomable infinite canvas’ approach. I had a long play with it, though, and to me it still doesn’t quite work for this: I’m not sure we can conflate everything down to a single plane, no matter how ‘zoomable’ it might be. For a while I toyed with the idea of having a multi-plane Prezi-type structure, with separate planes for each of the extended-Zachman layers, because more information is added with each layer; but then I realised that that too was an artificial constraint that would likewise soon get in the way. Hmm…
I suspect what we really need is a kind of holographic version of the Prezi approach, where the ‘zoomable’ plane in use is an arbitrary selection from an infinite palette of views. The technology to do that is not hard – it’s just another form of tagging, really – but devising an easy and ‘intuitive’ way to navigate through the space is not going to be so easy at all… (Milan Guenther, where are you? 🙂 )
From what you and others have been saying, definitely looks like I need to have another look at Sparx – if only because it’s probably the only metamodel-oriented EA tool that I could actually afford! 😐
Thanks again, anyway, and yes, please do keep in touch?
@Peter Bakker – Peter – yes, looks good! Definitely keep going on that, yes?
@Anthony Draffin Anthony – Many thanks for the update on Holocentric, though unfortunately seems to suggest that they haven’t moved on much from where they were five years ago. Oh well.
Likewise, thanks for the advice on Sparx: as I’ve just said to Ondrej above, looks like I’ll have to bite the bullet and get a proper copy going on one of my machines here. Watch This Space? 🙂
@Tom G
Tom, another though about tooling went in the direction of MDA. Would it help if you could define a metamodel (or a set of interlinked metamodels) and then in more user-friendly way (via automatically generated GUI) fill in concrete models? I have one tool in my mind, but it’s not being sold as a tool, rather as a solution to a concrete problem with a concrete metamodel. One cannot however arbitrarily view those models. It’s more a graph like structure at the end. And it’s also not visually that strong, it was focused on efficiency. Should you be interested, I could check with the owner of the tool under which conditions we could use that. Perhaps he might be interested in developing even something concrete on top of that should there be an interesting business case in the future. Interested?
@Ondrej Galik Ondrej, thanks, of course.
Am obviously interested in the ideas behind it, but probably best if I don’t get too involved the tool itself at the moment – particularly because of “one cannot arbitrarily view these models”. As usual it’s a bit of a struggle to describe what I’m working towards, but it definitely would involved the ability to view anything from pretty much any direction or viewpoint or perspective or whatever we might call it. More on that in yet another post shortly. But in the meantime, thanks very much, because all of these ideas and side-views and the like really do help.
@Tom G
Just wrote down “The 7 Rules of Tubemapping” at http://peterbakker.wordpress.com/tubemapping/the-7-rules-of-tubemapping/
Happy tubemapping everybody 🙂