This paper presents 2 draft patterns - Natural Selection and Community of Interest. These are intended to fit into the pattern languages which precede the ChiliPLoP 2000 Hot Topic on Introducing and Sustaining Patterns in the Organisation. Their emphasis is more on the 'organic' processes which occur during successful technology adoption rather than action which can be taken by the tehnology champion.
You are in an established organisation. You have come across a novel technique or technology (in this case, Patterns) which looks interesting enough to try out.
You dont know if its going to work out in your organisation. Will it be too hard to do? Will there be any benefit? Will people lose interest?
Allow people in the organisation to evaluate the new technology by trying it out. See if the technology has the desired effect. If not, see if the actual effects were beneficial anyway.
Do your best with the new technology (see Related Patterns) but dont assume that it will automatically succeed. Be prepared to accept the opinion of the people involved if things dont work out.
One of the hardest aspects of this process is knowing when you have reached an outcome. You may decide to apply Natural Selection in a formal evaluation exercise with a set time scale but in practice it will continue to operate for as long as you are using the technology. Some (all?) technologies reach the end of their useful lifespan within the organisation sooner or later and will be weeded out by the same process which drove their adoption.
This is probably an inevitable fact of life. In an organisation of more than one person, you cannot control whether your colleagues value the new concept or not. Trying to force people to use a technology which you think is great but they do not value is unlikely to succeed - even if you are right.
This depends on the outcome of the evaluation. In any case you will have a better idea of the effectiveness of the technology.
If people found it beneficial, they will spread the word. You will also have some concrete experience with it and others around the organisation who were not directly involved in the trial will have seen its value.
The result may not be a firm yes or no - maybe you collectively decide to 'cherry pick' the most appropriate aspects of the technology rather than adopting it wholesale.
If the technology is not beneficial, people in the organisation will lose interest and it will not thrive.
Most new technologies undergo this pattern when introduced to an organisation. Positive outcomes are easier to identify then negative ones...
Linda Rising describes the process by which Patterns were adopted in AGCS [RISING96]. The idea was introduced by a Champion but only gained a critical mass after being thoroughly 'test driven' by a much wider community of 16 people. The outcome of this process was by no means guaranteed at the beginning.
In 1989 when Tim Berners-Lee first proposed the technology which became the World Wide Web he opened his proposal [BL89] with the statement "This proposal concerns the management of general information about accelerators and experiments at CERN". Natural Selection proved it both simple to implement, massively scalable and user friendly and its use spread world-wide, sidelining its predecessors (like Gopher) and competitors (like Microsoft's Blackbird project)
The compound expansion steam engine revolutionised the maritime world by minimising fuel consumption and allowing long voyages. Seeing this exciting technology, many eminent railway engineers tried to adopt it to their own domain - surely lower fuel consumption would be a good thing. It didnt take over - why? - because it compromised some vital aspects of railway locomotive design - simplicity, maintainability and therefore reliability and availability. Non-compound locomotives remained the predominant species right up until the end of steam.
Using a Less Disruptive Introduction will greatly increase the chances of a positive outcome from Natural Selection.
Another very different approach to this problem is the Checker pattern. In practice the two approaches are not exclusive - Natural Selection will often follow an evaluation by a Checker (whether you like it or not).
The organisation is in the early stages of adopting patterns, possibly undergoing Natural Selection and/or Less Disruptive Introduction. Some people in the organisation are enthusiastic about the technology and others less so. As we are dealing with humans, the areas of enthusiasm probably dont align with work group boundaries.
How do people who are enthusiatic about patterns get to feel supported if they are in the minority or alone in their work group?
Encourage the formation of Communities of Interest (also known as Special Interest Groups) within the organisation. Look for suitable communities outside of the organisation as well.
Groups within the organisation can meet during working hours or at lunchtime to exchange ideas. Each meeting should have a theme rather than being allowed to proceed without an agenda. See Related Patterns for some ideas.
Communities of interest provide the support and encouragement which for which most of us have an inbuilt need. By allowing and encouraging the communities to be cross-organisational the community can draw on a bigger pool of recruits and is more likely to reach a critical mass.
People from all over the organisation can benefit from peer support and feel less alone in their adoption of patterns. A productive exchange of ideas from inside and outside the organisation is promoted.
This is a very widely used pattern, both in the patterns field, other technology areas and also non-technical areas.
Probably the best 'large scale' examples in the patterns field are the annual PLoP, ChiliPLoP and EuroPLoP conferences. Other more local groups exist such as the Silicon Valley Patterns Group.
A number of mailing lists exist for exchange of information in the patterns community. Here the emphasis is on year-round on-demand communication rather than regular gatherings of people.
AT&T in the UK used Community of Interest some years ago to encourage exchange of ideas about new technologies. Meetings were held during working hours and were well attended, particularly by developers who were 'in a rut' technologically speaking and were seeking a way forward.
Many of the other patterns in this topic can be used to provide subject matter for the community to work on when it meets:-
Just In Time Training for the community by the community. A place where the Dedicated Champion can Preach To The Choir and disseminate the information needed to Keep It Going. A forum for a writers workshop for A Pattern of their Own Exchanging and selecting experiences for an Internal Drip Feed to the wider organisation
You should also consider the Gone to Maui approach when deciding on a venue and time for your meetings.
You might also Do Food or let community members bring a Brown bag to meetings.
[BL89] Information Management: A Proposal. Tim Berners-Lee. March 1989 http://www.w3.org/History/1989/proposal.html
[RISING96] Patterns: Spreading the Word. Linda Rising. http://www.agcs.com/patterns/papers/object.htm
I was first introduced to patterns in 1994 by an article by Kent Beck in Dr Dobbs Journal. They struck a chord with me and I have been trying to improve my understanding of them and put them to good use ever since. My early efforts involved using the pattern form to document aspects of the design of the systems I was working on and also evangelising the pattern form within the organisation. There was little material to use for pump-priming at the time - this was before the GoF book and all that followed it.
In late 1995 I co-founded AT&T Labs UK, which at that time consisted of 3 like-minded individuals committed to being smarter at building systems by using object orientation and other good things. The need for evangelism went away and the main struggle was to keep up.
I have been actively involved in the wider patterns community (i.e. more than just listening in on the mailing lists) since early 1998 when I submitted my first externally published patterns to ChiliPLoP. These have now reached the PLOPD4 book, so I'd better stop changing them.
During the last 4 years AT&T Labs in Redditch, UK has grown to over 40 people and has been joined on our side of the pond by AT&T Labs in Cambridge. As we are now comprised of a very diverse group of individuals, the need for patterns evangelism is returning - the original bearers of the message have been spread very thinly by the growth or the organisation and many people's idea of patterns is probably restricted to the GoF book and/or the get/set pattern in JavaBeans.
I have also promoted or been connected with the introduction of several other technologies over the past few years. A quick count shows a roughly even distribution of successes, failures and decisions still in the balance. In analysing the reasons behind the successes and failures I have come up with several common themes but I doubt whether I will ever be able to predict the outcome of the process at the beginning.
I'd like to finish with some suggestions for disussion points...
Date: 12 January 2000
Author: Andy Carlson