Starting a new Adventure

After 99.5 months (sounds more interesting than “8+ years”) at Microsoft, I’m heading off on a new adventure with Cloudera. Over the years with Microsoft I have worked with some amazing folks in support, consulting, and sales.

Thanks to all the folks from Microsoft who have mentored me, worked along side me, and who wished me well as I made the difficult decision to depart. I know that I started some lifelong friendships along the way.
Last day at Microsoft:

And thanks to all the Clouderans who have helped me find this new opportunity, believed that I was up to the challenge, and welcomed me to the team. I’m excited about focusing intently on the Hadoop ecosystem, and I feel like I’ve found a great fit with the energy and culture I’ve seen so far at Cloudera. It’s going to be a fun adventure!
First day at Cloudera:

Waking me up to complain about a low battery is terrible UX

If you design a product that might wake me up in the middle of the night, you should meet the following bar: I, or someone I love, must be in danger.  Not a hypothetical “If the power goes out and if something in the house is still managing to produce CO without power” type of danger, but something more actual and imminent.

Smoke and CO detectors seem to have been designed with a severe disregard for how the customer will react to the various interaction points other than a true emergency.  I can remember several times in my adult life when a chirping detector has induced me to get out a ladder in the middle of the night and climb up to check on various possible sources of the chirp, all in a semi-alert state.

Last night, the culprit was a CO detector that happened to be plugged in to the wall in our master bedroom, perhaps about 3 meters from where I sleep.  The acoustics of the alert sound on this device seem to have been designed to make it echo and reverberate around the house, which is great for an a alarm, but again is horrible for a low battery warning.  In my efforts to find the noise that only repeated every minute or two I got out two different ladders, stood on the ladder near the upstairs smoke alarm, then the downstairs one, then climbed into the attic to check that one, then stood in the room listening for the chirp again, then up to the attic to see if maybe there was another alarm up there that I didn’t know about.  Thankfully my wife happened to look down at the right moment and correlate a chirp with this CO detector plugged in to the wall.

I know that I should have probably replaced the battery at some point, but honestly I had totally forgotten about this particular detector.  It’s one of 3 CO detectors in our house, and I’m pretty sure that it’s been 3 or 4 years since the battery was changed because it runs off house power and the battery is just for power failures.  The end result of this particular warning was that the device was unplugged and batteries removed so that we could get back to sleep.

This morning as I tried to get back to sleep I was wondering how many middle-of-the-night ladder injuries are attributable to low-battery chirps.  I’m guessing that the low battery behavior is probably design-by-regulation and that manufacturers like First Alert don’t have a lot of latitude to make this better, but I see that Nest Protect is trying to solve the UX gap, although at quite a premium on price.

Fixing a Drippy Ice Maker

A few weekends back I spent the morning fixing a drippy ice maker in our Whirlpool Gold Side-by-Side refrigerator, so I figured I’d do another “fix-it” post. 

One morning when I opened the freezer there was a bit of a winter wonderland scene, with frost and icicles covering the upper-right corner of the freezer.  The in-door ice bucket was full of huge chunks of refrozen ice, and frozen up to the point where it couldn’t move the agitator to dispense ice.  Pretty much just an icy mess.

I originally thought that maybe someone had left the freezer a bit open, but there was still a bit of water dripping around and I know the door had been closed before I opened it.

Still not knowing what was wrong, I removed the ice bucket, emptied it out and melted all of the ice that was frozen around the agitator.  I put it back in the freezer and then went about my morning.  A bit later, I went to get some ice and remembered that I hadn’t turned the  ice maker back on, so I hit the switch.  I closed the door and waited for the ice to drop (It’s an in-door bucket so it doesn’t drop until the door closes).  Right when the ice dropped I opened the door and was then able to see first-hand the real problem with the dispenser.

Evidently in this model fridge, the water is dispensed into a small collector that then lets it flow into the ice tray.  The collector is circled in orange in the picture below.  Somehow this collector had filed with a solid chunk of ice, and was blocking the flow of water directly into the ice tray.  Since the float switch to stop the flow of water is in the tray, this means that the water stays on until enough water dribbles around the collector to fill up the ice tray (as well as coating a good portion of the freezer with water.

At first I thought I could attack this with a hair-dryer, but after removing all of the food from the top two shelves and then sitting with a hair dryer on the collector for about 5 minutes, I realized this was not getting me anywhere.  My next idea worked much better.  I used a turkey baster to squirt near-boiling water into the top of the collector (with towels positioned below to catch the overflow) and this unfroze the collector in about 2 minutes.


We haven’t had any further problems with the ice maker.  I’m not sure if this was just a fluke, or if there’s some sort of accumulation of ice in the collector that will come back again, but for now it’s working great and I know how to fix it quickly it it freezes up again.

If Your Dodge Caravan’s Power Sliding Doors Stop Working, Consider Replacing the Battery

I’ve now seen this on two different cars of the same model & year and I couldn’t find a reference to this specific resolution on the Internet so I figured it would be worth posting.

The symptom is simple.  The power sliding door on one side or the other will stop working.  I’m guessing that this problem is shared between the Dodge Caravan and the Chrysler Town & Country since they share the same parts for this mechanism.  Some people on the Internet say that the problem went away after they had the dealer flash the BCM (Body Control Module).  Other people had to get an entirely new BCM before the problem disappeared. We never went down either of those paths since I had a simple quick, although temporary, fix.  If you open up the fuse panel under the hood next to the battery:

Then remove, count to 5 and then replace the following fuse:

The doors should work again for a while.

With the first van that had this problem, we went for many months just pulling and replacing the fuse whenever a problem occurred, and it would usually only stop working every few weeks.  A while later we ended up replacing the battery after the van wouldn’t start one morning.  Since the door issue was only an occasional problem, I never really realized that it never happened after we got the new battery.

Fast forward a few years, and we ended up having to replace the van, and we opted for the same exact model & year.  When we had the “new” used van for a couple of weeks, the left power sliding door stopped working one day.  We were on a trip so I did the quick-fix fuse pull & replace, and the door started working again.  Within a week, the van wouldn’t start one day.  A jump start got us working again, but I took the van to Auto Zone to have the battery checked, and it failed the load test.  I replaced the battery on this van, and we haven’t seen the doors stop working again!

My guess is that the Body Control Module (BCM) that gets so much attention has a failure mode that is supposed to disable the sliding door motor if it detects a short or a stuck electric motor, but that when the van’s battery starts getting marginal, it trips into this mode unnecessarily.


Since this post still gets a bit of traffic, I wanted to add a bit of an update.  After many more years of wear, we have had a few more challenges with these sliding doors.  If you get to a point where most functions work, but one doesn’t then you’ve probably managed to actually break a wire in the harness.  This can manifest as something like locking or unlocking no longer working, or open/close not working, while other things still work.  If the break is still not quite complete, this can be an intermittent problem for some time.

I tried finding and splicing the broken wire one time, and it only lasted a couple of weeks since there is a fair amount of strain as the doors open and close.  I ended up just replacing the entire harness and I was pleasantly surprised how easy it was.  First time (passenger side, more use) took me a bit over an hour.  When the driver’s side failed I managed to replace it in about 45 minutes.  You can get the parts from Amazon (links for 2004-2007 model): Passenger Side [747-311], Driver’s Side [747-310].

Old New Traffic Bump

I usually don’t worry much about the traffic on my blog, but when I logged into the control panel today this caught my eye:

Page Views showing 800 views on the day Raymond Linked

 Now I already was expecting a bit of a bump since Raymond had given me a heads-up about the link, but I’m always curious exactly what traffic a link will bring.  Between December 31st and New Year’s Day about 900 extra views compared with my normal baseline is a pretty good reach for a pile-of-links post referral.  The results say a great deal more about Raymond’s reputation for putting up interesting links than anything else!