Software that lasts.

Anyone can ship. The rarer discipline is to stay.

Since 2002 the work has been finding the actual problem underneath the stated one, building the actual solution, and adjusting the sails until it sails itself.

What to call that has never quite settled. Architect, some would say. Web designer, once. Coder, engineer, developer - each true for a while, and all different hats to wear. Meanwhile the water keeps moving as AI arrives, automation takes the repetitive hours, the stacks grow taller and easier to stand on. Everyone can make something that runs now. Far fewer can make something that keeps running. Code is generated faster than it is understood - vibe coding, for as long as the name lasts - many times thin on foundations, security and questions that should have come first. Here we are at the new generation of coding.

None of that changes the work. It only raises the price of doing it properly. The stripes here were earned the long way: twenty-four years of it - with dedication, with determination, with the discipline to stay when leaving was easier, and with a name attached to every line of code, so there was always someone to answer for it.

A solutions builder is not a job title. It is the habit of asking what is really wrong before deciding what to make. On that, the mind is settled. Welcome to a solutions builder.

Ship is a verb. Stay is a discipline.

Twenty-two degrees north-northeast.

A compass bearing.

It marks a direction that has grown into the soul a long time ago and held ever since - through the years that were easy and the years that were not. It points at what matters most, at the place where the heart is and the dream lives, and it is not a place that can be reached by road.

That is the point of it. A bearing is not a destination; it is what keeps a person moving when the destination is out of sight. It is where the mind goes to rest, to regroup, to remember why the hard work was worth doing - and to stand on solid ground again before the next climb.

Direction is a discipline. So is silence about where the direction leads.

What has been built.

Software that holds.

Most of it carries money and the people responsible for it. Quotation systems that turn a conversation into a priced document. Invoicing that goes out on time and reconciles itself when paid. Payroll and wages calculated by rule, with the deductions and statements produced by the run. Ledgers in more than one currency, double-entry throughout, every figure carrying a name and a time. Behind each of them a back office where rules replace memory.

A great deal of it moves people and days rather than money. Websites and booking systems for the travel trade - rates, availability, itineraries and confirmations that agree with one another, and with the invoice. Calendars and appointment systems for anyone whose week is sold in slots. Dispatch that behaves the way people expect it to. Integrations - a long list of them - joining these to banks, payment gateways, tax authorities, accounting packages, suppliers and channels: systems that were never meant to meet, made to agree and to keep agreeing when one of them changes.

The rest is the ground it all stands on, looked after by the same hands. Servers on Linux behind Nginx and Apache; hosting and the panels that run it, from cPanel and Plesk to DirectAdmin and the open-source alternatives; domains, DNS, mail, certificates and firewalls. Cloud when the job warrants it, a single well-kept server when it does not. Every system breaks eventually, so every system is built for the day it does: deployments that go out without downtime, backups that are tested rather than assumed, and a restore that has been rehearsed before it is ever needed.

Native apps for the phone that record where the work happens, keep working without a signal, and settle when one returns. And the newer discipline of reading paper: receipts, invoices and statements scanned, the figures and line items extracted into structured entries, checked against the books they belong to, and posted without a keyboard. The aim is not to take the bookkeeper out of the books. It is to give back the hours that went on retyping what was already written down, and to spend the models only where they earn their place.

Going live is not the end of the work. It is where the proof begins.

The Method.

Dedication and thoroughness, every time.

Start where the work actually happens - at the counter, in the yard, on the warehouse floor, at the kitchen table - and learn what people really do there, not what a form assumes they do. Most problems are misdescribed at first, usually by the people closest to them, and the method begins by finding the real one underneath the one that was reported.

Build the core first and finish it properly before extending it. Scope earns its place by being relied upon, not by being requested.

Dedication is the part that shows up. Thoroughness is the part that nobody sees: downtime planned for before it happens, failure assumed and prepared for, nothing trusted until it has been broken on purpose and brought back.

The languages, the frameworks and the machines have been replaced many times since the first line was written, and will be replaced many times again. The habits underneath them have not changed. They were never about the tools.

Care is a habit, not a mood.

Since 2002.

Twenty-four years on the books.

The first years taught the obvious: that code which works on the desk is not the same as code that works once it has left it. The next years taught the expensive lessons: that a figure which cannot be traced is a figure that cannot be trusted.

The middle years taught patience. Financial systems, booking engines, payroll runs and back offices are not built in a burst; they are built in the unglamorous middle, where most of the risk lives and none of the applause does. The habit of finishing properly was formed there, along with the habit of saying no to the wide build until the narrow one could be trusted.

The recent years have taught humility, in the good sense. The tools now do in weeks what once took a season, and the temptation is to believe the tools did the thinking. They did not. Twenty-four years is what it took to know the difference - and to keep asking the question that the tools cannot: what is actually wrong here?

Twenty-four years is long enough to have earned the right to say less. There are no case studies here. No logos. This is not a gallery. It is a habit, kept daily, of doing the work properly.

Time is the only test that cannot be faked.