Thursday, November 16, 2006

Kitchen Renovation Project and Systems Integration Project: compare and contrast

That long silence was caused by the Great Kitchen Renovation Project (GKRP), which started at my house on October 3rd 2006 and finished on the morning of November 16th. During the last six weeks I think I have seen almost every one of the things that go wrong on large integration projects go wrong on my kitchen refit.

I should probably explain that the old kitchen was a "temporary" facility, built as cheaply as I could manage it, and never intended for long term use. It featured 1.8 metres of benchtop with cupboards underneath, which I bought in the going-out-of-business sale of a kitchen supply company. I had a factory second sink installed in the benchtop, and added a variety of cheap, freestanding pine and melamine cupboards for more storage. A bench from Ikea, our existing microwave oven, the fridge that my husband bought second hand when he first moved to Sydney and 900mm wide Ilve cooker (gas top, electric oven, and the only thing I spent significant money on) completed the original fit out. The floor was concrete, sealed with Berger Jet Dry paving paint. We intended to upgrade the kitchen as soon as our finances recovered from building the house in the first place, keeping the expensive cooker and scrapping everything else.

Now how many times have you deployed a "temporary fix" that wound up staying in production for years? We used the temporary kitchen for 9 years, and the catalyst for the renovation was the expensive Ilve cooker breaking down for the umpteenth time, and me refusing the pay to have the lemon of a thing repaired yet again. Honestly, when you are on first name terms with the repair man, and he can comment knowledgeably about the progress of your landscaping, he's visiting too often. That Ilve cooker was nothing but trouble from the day it was installed, and I would never buy another.

When the cooker failed, I went out and bought a large Sharp Convection Microwave oven, with which I was extremely pleased. The gas cook top still worked, and with that and the microwave, we could feed ourselves decently while the new kitchen was planned and built. Little did I know that the new microwave over would be the factor that caused most of the problems with the new kitchen.

I went forth and shopped. I'm good at this: I shop thoroughly, with much research and attention to detail. I don't mind spending money, but I do expect to get quality, and exactly what I want. So I selected a fabulous new cooker from Falcon (I bought the cream coloured one); a wonderful fridge/freezer from Maytag, a Qasair rangehood (model CM ) and a Miele dishwasher. And I looked at a lot of kitchen showrooms, and settled on two potential suppliers, both of whom I asked to do a design and quote. I supplied all the details of my chosen appliances. I explained about the large convection microwave, which I wanted to keep. Each potential supplier sent a designer to measure the room, and talk to me about what I wanted. I showed each man the manual for the microwave, which stated that there was a special installation kit needed if you wanted to build it into a cabinet. And each man said words to the effect of "oh, we can build that, there's no need for a kit".

Both designs came back quite similar, and my husband and I chose the one we preferred (it cost a little more, but we felt the workmanship was better, and the cabinets would be made in Australia rather than in another country). We settled on finishes - cherry wood with stainless steel benchtops and splashbacks - and the supplier sent out their "check measurer", to ensure that the first set of measurements was accurate.

I insisted that the check measurer double check the fitting requirements for the microwave, and this time some one called the microwave supplier and actually got the real word. The answer was that the microwave vented heat at the back, and need a huge clearance (relative to the cabinet space available) to be installed safely. The kitchen person called me in distress to explain this, and it was a real problem: the microwave would not fit where it was supposed to go, unless I never used the convection feature.

Now I have seen similar things happen on integration projects: the presales team assumes that a piece of technology that they have never worked with before will work in a particular way, and when it doesn't, their plans and designs are invalidated. Or the customer tries to alert the sales folks to a problem, only to be assured that all will be well, and when it turns out that the customer was right, there's plenty of egg for all the faces concerned.

I was once involved in a project where the sales team assured the customer that all their shell scripts could be ported from DG/UX to Solaris. On the face of it this seems fairly reasonable, but the customer told sales that they had some scripts that involved print queues that were pretty special, and they thought there might be problems. They were more or less told not to worry about it, in a rather patronizing way. The contracts were signed, and I turned up on site to start the work. The customer showed me the suspect scripts, and I could see at once that there were problems. Their scripts relied on certain behaviours of the old DG/UX printing subsystem and the way DG/UX handled serial devices. I could have made the scripts work on Solaris 2.6, but large chunks of the Solaris printing and device subsystems had been rewritten to improve security, and what would work in 2.6 would not work in Solaris 9.

The customer knew perfectly well that this would be a problem, and I think they took some pleasure in informing me that, for them, this was a show stopper: either we provided equivalent functionality for the duration of the project (after which a new system would replace the old stuff and all its functionality), or the project got stopped. Fortunately I managed to concoct a solution which got around the problem, but it would have been better if the customer's legitimate concerns had been addressed at the beginning.

It was the same thing on GKRP: my legitimate concerns weren't addressed until the designs were completed. I wasn't prepared to sacrifice precious cupboard space to give the microwave room to vent, so I elected to replace the microwave with a non-convection unit. Since I was getting the Falcon cooker, that had two ovens, this seemed the best solution. The microwave cabinet could be made a bit smaller, the drawer unit next to it could be made a bit wider, and all would be well. So we agreed on a variation, and the diagrams were revised.

The day came when the cabinets were to be delivered to my house, and I got a phone call from the owner of the kitchen company. There had been a mistake. The news about the variation had some how not reached the cabinet makers. The microwave cabinet had been built too wide. Modifying the microwave cabinet was easy, it could be cut down and made narrower, but changing the drawer unit next to it would be a major exercise. He suggested that instead of changing the drawers, they would build a small wine rack to fill the space. This could be done quickly.

At this point I had the contents of the old kitchen stored in my dining and lounge rooms, and the kitchen was completely empty. I knew I was going to have to use the sink in the laundry for all tasks that would normally be performed in the kitchen sink for as long as the refit took. For cooking I had a single electric hot plate, the convection microwave (on a bench in the dining room), a sandwich toaster and a barbeque in the back yard. I felt like I was camping in my own house, and the prospect of significant delay was horrifying. I agreed to the wine rack solution.

Now how many times on a large integration project has there been a change in the design that has not been communicated to the people building the systems? And every time it happens there is a trickle down effect to other places in the project......

The cabinets were delivered, and all placed either in the kitchen or in my family room. Then came the appliances. The appliances are all large. The dishwasher and the microwave would fit in the family room to await final installation in the kitchen, but the cooker, the fridge and the range hood all had to be placed in the garage. The men who delivered the appliances looked at the access to the kitchen - my house is built on a narrow block of land, and in consequence some of the doors are narrower than standard - and expressed the opinion that getting the fridge into the kitchen at all was going to be a serious challenge.

Now how many times has a large computer/SAN/tape drive been delivered to a site, only to find that there's no way to get the thing into the building without removing doors or windows, or unpacking the kit in the loading dock and hauling it into the data centre piece meal?

Well, that's what happened with the fridge. Once the cupboards were all installed, a team of strong men arrived, removed the fridge doors, the hinge assemblies, and the freezer drawer, and carried the pieces into the kitchen. Then they put it back together. I am impressed beyond expression with the fact that they achieved this (including lifting the main body of the fridge and passing it into the kitchen through the counter-height servery hatch that connects the kitchen to the family room when there proved to be no way to get it through the door) without swearing once.

Then the installer tried to put the microwave oven into its cabinet. The cabinet had been made narrower, but it the cavity was still too tall: there was a big gap at the top. The whole cabinet had to be removed again, and taken away for adjustment. Things were now running quite late, so the project coordinator made the classic mistake that project managers make when a project is running late: he booked multiple trades to be on site at the same time. When the gas plumber arrived, there were two electricians and three rangehood installers already working in the kitchen. The plumber went away again, and came back later in the week. The rule here, and many project managers never learn this one, is that some tasks cannot be performed in parallel, and that throwing more engineers at a job may make it slower.

Another common problem on large projects is the part or service that is supposed to be provided by "someone else". The vendor assumes that the customer will have sufficient switch ports to plug in all the new equipment. The hardware people think that the software folks are organizing the software licenses. The software people assume that the system administrator will take care of their backups. Whatever, there are never enough switch ports, no one orders the software licenses and the system administrator doesn't have a big enough backup window to add the new systems. The thing you assume is "someone else's problem (SEP)" just gets missed. On GKRP, the kitchen company assumed that their subcontractor electrician would supply the chrome down lights, and the electrician assumed that the kitchen folks were supplying the parts. Consequence: no down lights, and a delay while they were ordered and supplied.

However, slowly it has all come together. The house is more or less back to normal - everything that should be in the kitchen is in the kitchen, not in some other room. I still need to get the walls painted and the flooring laid, but it looks great,

and works like a dream. My nerves will recover, eventually.


Bookmark and Share