First Project Proposal – Wikitect

OK, so the primary purpose of my venture into the world of tech is to become a part of it. I am confident of the skills, creativity and work ethic I have to offer, but somewhat short on substantive CV matter. In the long term, this blog will hopefully serve as some part of a distributed resume – an aspect of a career bootstrap process.

One plank of this will be to publish project proposals – ideas for ventures of various kinds, more or less deeply considered – to give a flavour of the way I think, the areas I’m interested in, the thought processes I use.

Going public with ideas is an interesting undertaking – taking us straight to the heart of one of the startup dilemmas – are our ideas so precious (my precious…) that we need to keep them secret until we can unveil them in glory (or at least as an MVP), or is the need to find collaborators, to develop traction, so important that it requires us to come out in the open?

Do we need to fear that our ideas will be stolen/pre-empted/devalued by sharing them, or should we be confident that the process of sharing gives opportunities that will enhance the chances of success?

The answer, of course, is that this is a foolishly abstract question. The specifics are everything. Circumstances alter cases.

If, at one extreme, you have invented a method people can use for reliably assessing whether a new acquaintance is a viable life partner, and if furthermore, this method is relatively simple to implement after a brief explanation, then obviously you will keep it secret, working only with a few trusted individuals, then raise enough cash to launch big. On the other hand, if you are convinced that you know how to build a social media platform that will unseat Facebook, then it is a positive requirement to talk about it everywhere, in the hope of convincing someone, anyone, that it is worth investing five minutes in considering it.

Most ideas, of course, lie away from these extremes, somewhere on a continuum between them. To make an informed decision, you will need to ask yourself some questions, and chew over the answers. These are the sort of questions I mean;

  • Do I need ‘in the know’ help to get to a proof of concept or MVP stage?
  • What does the IP landscape around my idea look like (short-term / long-term)?
  • To whom I going to be talking about it? Under what conditions?
  • How difficult will it be to get the idea to market?
  • … add your own concerns…

I have a number of project proposals, of various kinds, in draft form – as no doubt do many of us. Some of them are easy to talk about, some less so, but overall, standing where I do, I am increasingly leaning toward increasing open-ness as the best position for me.

That said, I’m starting easy, in that my first proposal is not really my own, has no immediately obvious market, and is based on an open-source platform.

It’s pretty interesting, though, and connects to some strong themes in the world of tech, in my world of architecture, has startup relevance AND, imho, looks toward something that humanity desperately needs as it finds itself operating on a scale where we have powerful interactions with inherently complex systems.

Re-reading that last paragraph, it looks rather too much to attempt in a single post! I’m going to deal with it like this – I’ll introduce Wikitect as simply as I can, set out a few of the possibilities, and then flesh the whole thing out in a series of pages accessible from the PROJECT PROPOSALS menu item.


Wikitect exists as a collection of short scripts in PERL (not very far beyond proof of concept stage), written by Ward Cunningham. He describes it here as;

 ..a wiki inspired script for creating high-level architecture diagrams…

Ward had already written Wikitect when a proposal I had written made its way to him. We had some correspondence, and he released the code under the Eclipse EPL open-source license.

Put simply, wikitect allows you to structure a set of information in a diagram that is hierarchical, but not strictly so – mutliple linkages are allowed, and these can cross hierarchy levels, so allowing for rich models that map well on to real-world situations.

Rigid hierarchies can be captured in tree diagrams, where each node only connects with a single node higher up in the structure. Tree diagrams are simple to parse and eminently clear. However, in too many cases they map poorly onto reality – particularly  in situations with any appreciable degree of complexity. Tree diagrams cannot model feedback, or indeed any type of recursive connection.


Wikitect, (as you would expect from Ward, the father of the wiki), is lightweight and flexible. Crucially, it is able to export the network (example) in the open ‘.dot’ format, which can be used by a variety of applications to produce pretty, human-friendly output (example).

All this was in 2006. I was soon very busy with a number of architecture projects, and lacked the time and other resources to do what I had hoped, which was to get some coders involved with Wikitect as an open-source project. I had some fun finding out how to use git, built a network of my own in Wikitect (which I left open and quickly got spammed into meaninglessness). Very little else happened. So, you might ask, what is the point of reviving an idea that didn’t catch even when it had the good fortune to come from the mind of one of the giants of IT?

Perhaps, I might answer, the idea was ahead of its time…

Less vacuously, perhaps, what Wikitect can support is a way of structuring information that is infinitely richer than available models. Possible applications include: better brainstorm/mindmap software (all of which are stuck with tree-like structural paradigms), org-charts that map better onto the realities of organisations and projects, tools to allow people modelling processes to create legible diagrams that are based on ‘live’ (editable) data structures.

My strongest immediate proposal is for a tool that people can use to build ‘live’ business plans – ones that remain useful to the organisation beyond the date of the meeting they were last revised for. Ones that can do a much better job than unstructured wikis or social platforms in ensuring that everyone in an organisation understands why and how it works as it does, and how they can fit in. This sort of tool would be of great value to rapidly expanding startups, and also connects with the growing field around business rules approaches.

Finally, in the long term, I strongly believe that humans need better tools with which to cognise inherently complex systems. The reductionist approach has proved so useful for the last 400 years that we are in danger of assuming that it is the sole model with which to probe the mechanisms that underly the universe. As we need increasingly to make intelligent interventions into systems that are at the same time fundamental to existence and finally unpredictable (specifically, our ecosystem), we need tools that don’t rely on isolating one variable at a time, tools that can assist us in mapping and modelling complex networks.

Wikitect can be one such tool.

Resources: AboutUs page, with links – including git formation.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.