top of page
Search

Stories from the Wall: the UFG Adventures

  • Writer: Paul Hogendoorn
    Paul Hogendoorn
  • 2 days ago
  • 5 min read

These two gems arrived recently, and I’m thrilled to have them on my trophy wall.


I first got the chance to work with UFG (United Flower Growers) in 1987. It was the 3rd “Dutch Clock” system I would do for the three co-operative flower auctions in Canada, but it was the biggest and most advanced auction of its kind outside of Europe.


They were always pushing the envelope.


They also preferred home-grown solutions in those days (still do) and home-grown means west of the Rockies. I felt truly flattered to be given the chance to work with them.


It started relatively simple back then; they wanted two electronic clocks for their newly planned auction gallery, similar to two we had recently sold to an Ontario flower auction. Everything else came from local suppliers and developers.


It was a bit strange at first, as we were clearly the out-of-towners, but in time, we melded into the fabric and earned our stripes by doing whatever was necessary to deliver the results that this industry and thought leading client wanted. They set the bar high, and the only way to achieve it was focusing on the end goal and working together.


A few years later, they bought a third Auction Clock, making it the only flower auction outside of the Netherlands to have more than two. And then they started planning even bigger.


Pedal to the metal”, some might say, but perhaps “petal to the medal” would be more appropriate. They were clearly the leader in the industry outside of the Netherlands, but with regards to systems and methods, they were thinking even further ahead and looking to take the clear technology lead.


They would build an entirely new auction gallery, only a decade after building the one our first 3 auction clocks were installed in. It would accommodate 640 buyers, run 4 auctions simultaneously, and facilitate remote buying possibilities. This was 1997 – well before technology would make things like this easy.


Having earned our stripes in the previous projects, we were given the chance to develop the new keypad bidding system and auction interface. It was the most technically demanding project our company had taken on, and it was the first one that I was involved with personally where I wasn’t the primary interface system architect.


Up until this point, I had often taken on projects that seemed over my head and beyond my company’s collective abilities, but we selectively took them on anyway to continue to stretch ourselves and grow our capabilities. This one, however, was bigger, more complex, and way over my head. It was also the largest single project sale my company had made. I had a young, talented and dedicated colleague that was eager for the opportunity to take a big challenge on. He had worked with me on several small projects, and I believed he was up for the challenge.


And so, we took it on.


We designed everything from scratch; the interface, the keypads, the communications protocols. The interface would use 48 state-of-the-art microprocessors to communicate with the 640 keypads, and a new, powerful (also newly released) microprocessor to run the four auction clocks and communicate with the auctioneer consoles and backend system.    


The hardware came together reasonably well, and on schedule. But the firmware was more challenging. Trips to site that were planned to be 4 days became 2 weeks, and the 2-week trips started to stack up end to end, with just a few days at home days between each trip to recharge the batteries, get some fresh ideas, and get some fresh laundry.


Some technical challenges were known at the outset. Every 25 milliseconds, the price of a product being sold on a Dutch Clock declines by a cent. The 640 keypads have to be scanned 10 to 20 times every 25 milliseconds in order to know who was the first to place a bid.


Since it is not uncommon for many buyers to be attempting to bid at the same price, an idea was introduced that could dramatically shorten the duration of the auction. It was decided to accept the first 4 bidding attempts (if there were multiple bids at that price), providing there was sufficient product available for the bids. The first one would be awarded their requested quantity and the second, third and fourth bidders would get their requested quantities if there was enough left. There were 4 auctions running at the same time, connected to the same 640 keypads, compounding the critical timing challenge.


Other technical challenges were completely unexpected and had to be figured out along the way. We designed the keypads around the newest single chip microprocessors that were chosen for their ability to be reprogrammed in-situ. This feature (as it was published) would have enabled us to change the keypad program without having to physically unplug and remove the keypads. The feature didn’t work as published, so whenever a change needed to be made, and in the early days of trials and evaluations, there were a lot of iterations. All 640 keypads needed to be removed and reprogrammed manually with every change.


Eventually, we figured out how to do it, but then we encountered another new technology problem; the microprocessors in the keypads used a new kind of flash memory and sometimes small parts of the program would evaporate. If just one or two bits of program changed, the keypad wouldn’t function properly. So, we created a way for the keypads to review every byte of the program and return a 4-digit code. If even 1 bit of the program memory was altered, the checksum code wouldn’t match, and that keypad would automatically be reprogrammed using the in-situ programming function that we were eventually able to figure out.


It was 1997. They were the leaders in their industry and pushing the envelope to stay comfortably out in front, and we had to push the envelope by using the newest technologies available (some unproven, as it turned out) and striving to achieve objectives that stretched us to our very limit.  


Roughly 25 years after the first 2 electronic Dutch Clocks were installed, UFG modernized their auction once again, retiring the system we developed and installed in 1997 and the electronic clocks installed in 1987. When we first joined them in their journey, they had a single, mechanical clock auction with buyers calling out their quantities to an auctioneer.


Over the next 25 years, over $1B of perishable products were sold on those clocks.


They grew significantly, and so did we, proving two of my oft repeated adages: "choose your customers well, and then settle for nothing short of their success", and "the company with the best customers, wins". UFG was one of the "best customers" we were able to build our company on.


Postscript: I don't often mention people's names in my "Stories from the Wall" series (because there's usually so many people to mention), but after writing this story, I felt I had to offer a "tip-of-the-hat" mention to a few folks: to Tom, for his vision and drive for his organization to always be the best, and for giving us the opportunity; to Fred for sticking with it and finding ways to get the near-impossible done; to James for being a great collaborator, and to Dennis and his great team for all the support over the years - and especially for the two new trophies that are now on my wall.


 
 
 

Comments


It couldn't hurt to reach out...

London, Ontario, CA

+1 (519) 281-6772

  • LinkedIn

+1 (519) 872-7202

  • LinkedIn

Received! We'll be in touch shortly...

bottom of page