The first elections to the National Assembly for Wales on 6 May will not represent some constitutional "big bang". Devolution is not a single defining event but a process.
No doubt the Assembly will have imperfections but it will change and adapt over the years; it will grow in confidence and stature; its power will increase and it will become the respected forum and authoritative voice of our country. The task of getting our devolution proposals through the Labour Party was difficult enough. The task of making our new democratic settlement work will, I believe, prove an even greater challenge. And we can learn lessons from Labour's 1992-97 journey that are relevant to choices we face now.
From 1994 a new vocabulary crept into Labour's lexicon. Party members were now supposed to be, at least in the media's eyes, either new or old Labour, although many of us were neither. Old Labour had little time for devolution. It wanted statist solutions; it thought devolution pandered to nationalism. New Labour had scarcely more enthusiasm - it certainly wasn't part of the original "project" but Tony Blair, to his credit, honoured the commitment inherited from John Smith.
There was, I always thought, a third strand. Not new Labour or old Labour but Welsh Labour. We were the enthusiasts for devolution who wanted a new pluralist approach to Welsh politics that was libertarian, decentralist and patriotic.
It was during these early debates that one of today's common buzzwords - "inclusiveness" - first emerged. It signalled a new style of politics which was less tribal and less dogmatic. It attracted most outright hostility from those members of Labour's Welsh establishment who believed that the party had a monopoly of wisdom on all issues. Inclusiveness implied that Labour would engage in a dialogue with others who shared a common agenda and be less assertive of its singular dominance of Welsh politics.
For devolution to be a success it was essential for the Labour Party to make clear that it had no intention of replicating the worst practices of some Labour-dominated local councils. Inclusiveness was not attractive in marketing terms but it was an essential foundation stone for the whole enterprise. It was my strong view that we would only be able to generate public support for a devolved assembly if we stressed the principle of inclusiveness. When the Prime Minister, at my urging, put the word "inclusive" in his speech to the Welsh party conference in May 1996 it paved the way for the Welsh party's acceptance of "an element of proportionality" in the electoral system for the new Assembly.
For devolution to succeed we must acknowledge the mistakes we have made so far in order to apply the lessons to the challenges ahead. First, under pressure from old Labour, we failed to broaden our support among other parties and across Welsh civil society. Only Labour had the strength and organisation to deliver change but that strength carried, and still does, a heavy responsibility to others. We should have had the confidence to work more closely with others to seek wider agreement and to identify common ground.
Second, this time under pressure from new Labour, we failed to deepen our support among the population at large. New Labour's advisers and apparatchiks were cool about devolution, seeing it as not relevant to "the project", and a potential vote-loser, in England at least. Neither before nor during the general election did we take our case to the people. In Wales we had one brief press conference at the start of the six-week campaign on our plans for the Assembly. Although it was officially one of the party's election pledges, as far as Labour's national campaign was concerned Welsh devolution was a non-issue.
Third, we failed to acknowledge the extent of the changes that must flow from devolution. We should have made it clear that devolution was an essential part of transforming the way we govern ourselves; it has far-reaching consequences for government, the economy, society at large and the political process itself.
Devolution is not an optional extra to the modernisation of the UK. Nor is it a panacea. The early years of our new democracy will present us with some difficult choices and painful realities. For example, the temptation will always exist, even after devolution, for money to follow UK government priorities. This year, additional resources were made available to cut hospital waiting lists and to reduce class sizes below 30 for the under-sevens. This was perfectly sensible: Labour was elected in 1997 on specific pledges to do these things. After devolution, however, when the co-operation of the devolved assemblies will be needed for such policies, there can be no guarantee that targets as arbitrary as waiting lists or class sizes could withstand rigorous scrutiny in a Welsh context. There will be clear tension between those wanting "UK solutions" and those wanting national "opt-outs" or other variations.
Again, there will be conflicts when each constituency in Wales has two directly elected representatives, even where they are from the same party. Both can claim to be the authentic voice of the people. How will a constituency party respond, for example, when its Assembly member supports a call for more influence over Westminster legislation while the Westminster MP strenuously resists it?
These matters can be resolved but the political parties will have to change and innovate. Everybody concerned will need a degree of tolerance and creative thinking.
The writer was secretary of state for Wales until November. This article is an edited extract from his pamphlet "Devolution: a process and not an event", published this week by the Institute of Welsh Affairs, 01222 575511