January 24, 2015

My wife and I have recently begun the home-buying process. After looking at our finances last fall, I realized that if we were to liquidate some of our investments, we could have a downpayment for a house. With this in mind, I began lining up the assorted ducks” needed in buying a house, namely:

  • determining a price range
  • determining a down payment for that price range
  • looking at our mortgage options

Around the turn of the new year, we finally got a chance to go to meet with a representative from our credit union and talk over the basics. While I was familiar with some of the info, my wife had little knowledge of the process (and to be honest, I wanted confirmation of my own knowledge).

Prior to all of this, I had occasionally perused Redfin and Zillow. After a while, I determined I actually preferred Redfin and thus began keeping active searches just so I could get an idea of what I liked in a house and what I didn’t. After we had met with the credit union, as aforementioned, I started spending more time looking at details and scrutinizing photos provided on Redfin. It was this closer scrutiny that raised to my attention a detail I hadn’t previously noticed, Redfin realtor ratings mean nothing.

What do I mean by that? Why would I say that? People love reading feedback about individuals in service industries and how they treat their customers. Of course the ratings matter. No no, you misunderstand me. The reviews mean something, the ratings mean nothing.

The Redfin Realtor Search site is the way I looked up ratings for Redfin Realtors in our area. Go ahead, take a look at the ratings for realtors. At least in my area, Seattle, there’s not a single realtor they list that doesn’t have a 4.5 out of 5 stars or better rating. That leads me to the corollary to my previous statement: When everyone is a 4.5/5, no one is a 4.5/5.

This is the same concept as participation ribbons and honorable mention.” If your scale isn’t actually differentiated, then the scale doesn’t matter. It’s the reason your college courses were curved. It’s okay that the highest grade was a C+ and you got an F, the professor is going to curve the grade so that the person with the C+ gets an A+ and now you’ve got yourself a C. The professor wasn’t interested in the absolute value of grades people got, he knows that shit is hard. Maybe he feels he didn’t cover a topic well enough and now because everyone bombed the midterm, he’ll correct it by curving the course. Curves have an interesting side effect in which they encourage the class as a whole to gravitate toward a small range of grades. If some asshole gets an A before the curve, everyone else is hosed. That’s a story for a different time, and now I’ve gotten sidetracked.

So if every realtor is rated 4.5, how do you differentiate? Well you meet with realtors until you find one that clicks. The homebuying process is incredibly nerve-racking for your first time or at least it is for Caroline and I. You want a realtor that can keep you grounded (within your price range, how much maintenance to take on) and keep you calm. You can’t discern this from rating nor can you discern it from a review. That brings me to my next point: The written reviews matter as they’re a way to narrow down what you’re looking for.

If you’re a high energy person, you probably like being around high energy people. If you’re prone to being a little neurotic, you probably should have some people that are level-headed in your life. Same goes for a realtor. While the reviews written about a realtor help you filter, you won’t be able to really discern how good of a fit you are with them until you meet them and spend some time chatting.

A good example is our experience in interviewing realtors. We talked with two realtors, whom shall remain anonymous. Both were great and lovely individuals. They both had qualities we loved. I’m a detail oriented guy, security engineer and all. I like people that I feel are on top of their shit. They got it together, they know what’s happening. My wife likes to make an emotional connection with the people around her. She wants to feel involved and connected, and wants them to feel involved and connected. The first of the two realtors (let’s call her Realtor A) we met, is very detail oriented. She is a clear Type A personality, very results oriented but very friendly too. The second realtor we met was less of a Type A personality, but she clicked really well with my wife and I on a personal level. She seemed to have a calming effect that worked well with my wife’s anxiety and yet she didn’t seem inattentive to the details, which calmed my own anxiety.

So if the ratings don’t matter, why does Redfin post them? My presumption is that, similar to having a merit scholarship in college, Redfin-affiliated realtors must keep a certain quality of review feedback as well as throughput on home sales & purchases. It is likely that in order to meet the 4.5+ rating in the first place, you’ll likely already be a 4.5+ realtor straight out. That is to say, those realtors which can sustain such a rating and sale/purchase numbers is holistically a 4.5/5 realtor. They’re just that good.

So what do you propose Zac? Now, I cannot 100% say that Redfin only displays 4.5+ realtors, but if that’s really all they display and/or affiliate with, there’s no need to show the rating. Absolutely keep the reviews. By displaying the ratings, they create false market where consumers believe that these ratings matter except that they don’t. If a consumer trusts Redfin and even consider Redfin’s affiliated realtors, they’re already bought in. I remain unconvinced that the ratings actually influence user decisions in going with a Redfin realtor. That said, it’d be interesting to see Redfin use A/B testing for a period to see if there’s a meaningful drop in users requesting Redfin realtor services.

All in all, the point is simply this: When everyone’s a 4.5/5, no one’s a 4.5/5.

opinion
January 19, 2015

Introduction

In the beginning, there was Zac. He discovered Linux and ran fifty thousand different distributions, never being satisfied. When one distribution fixed an issue, thirty new issues were created. Subsystems bickered, sound servers wept, and bus systems rended their clothing.

This went on for years and finally Zac gave up. He threw in the towel and bought a Macbook. He basked in the mostly-Unix environment provided by OS X. He occasionally tried to take Linux back but it was always a disappointment. One of the more recent forays into nix, he ventured into the BSD realm. Using FreeBSD from versions 7 to 9, he found a mostly harmonious environment. There were forays into other BSDs but these were but fleeting dalliances. Then the FreeBSD gods thought fit to introduce hellacious regressions/complications around version 9. Thus ended Zac’s adventures in nix yet again for several years.

Enter OpenBSD in 2015, stage left.

Segue

It seems every year, we hear that 20XX is the year of the Linux desktop. Inevitably, every Linux pundit from Brasil to Mongolia will extol the virtues of switching to Linux. Now don’t get me wrong, Linux does great things but it’s never worked out of the box well enough as a desktop. There’s always something to tinker with, some driver to compile, some knob to fiddle with.

I installed OpenBSD 5.6 on my old Thinkpad x201 and much to my surprise, it just worked. Better than installing Windows out of the box on this particular machine in fact. WiFi required a firmware update, but that was as simple as running fw_update. I configured a few settings with the help of the very thorough OpenBSD documentation and it’s pretty much been cake.

Maybe 2015 is the year of the OpenBSD desktop rather than the Linux desktop.

USB Key Setup

I won’t cover how to burn an ISO to a CD. That territory has been tread since before I started using *nix. In fact, I’m not even going to cover USB key creation with OpenBSD install media. OpenBSD does a better job at that in section 4.3.4 of the FAQ.

Basic Tweaks

I’ve taken to storing the various tweaks to config files in OpenBSD in my https://github.com/zacbrown/configs/tree/master/openbsd. There’s three main groups in the openbsd’ folder of configs, spread below. Folders are in italics and bold.

  • fw_update - this is required to update the WiFi drivers. My chipset is the iwl-1000. YMMV.
  • etc - everything in this directory can be copied directly into /etc/.
  • xorg.conf - a little chunk of Xorg config to setup the TrackPoint
  • rc.conf.local - the basic settings I use in mine, including specificying that apmd should dynamically scale the CPU, which daemons to launch, what flags to pass to PF, etc.
    • BSDNow.tv has a good coverage of some of the things you’d put in rc.conf.local.
  • login.conf - Settings for how much heap processes can take before the OS forcibly kills them.

The aforementioned settings are pretty key to a good experience with OpenBSD as a desktop. The xorg.conf file is necessary for the TrackPoint middle-click button to work for scrolling. While rc.conf.local changes aren’t required, many of the options I’ve specified in there are good suggestions for laptop configuration. The login.conf file changes are necessary since web browsers are terrible hogs.

Elaborate Tweaks

I don’t actually have any elaborate tweaking” that had to be done. One open issue is getting the hardware volume buttons to control the hardware mixer rather than routing the commands through X to the application with focus.

  • etc - everything in this directory can be copied directly into /etc/.
  • apm - this folder contains scripts that will be run for various apmd events (suspend, standby, hibernate, shutdown)
    • In my case, the suspend script is used to cause the Xsession to lock. See .xinitrc below for what happens.
  • pf.conf - firewall settings, block all inbound. Might need to allow ssh at some point.
  • bin - some basic helper scripts
  • wireless & wireless.cfg - wireless is a perl script that nicely connects to preconfigured wireless networks specified in wireless.cfg. You can find the original author of the script here.
  • powersaver-mode - uses sysctl to change hw.cpuspeed to lowest setting (0) to reduce CPU usage
  • performance-mode - uses sysctl to change hw.cpuspeed to highest setting (100) to maximize CPU usage
  • .xinitrc - This file is read before X11 is started. Notable entries in it are the launching of xidle which is used by the aforementioned apm scripts to trigger the Xsession to lock on suspend.
  • .kshrc - This file is loaded for each new ksh instance. Just some basic defined variables used in the terminal.
  • .profile - This file is loaded for each new logon session. In order to get it to reload fully, you need to logout and back in.

The configurations above are the extra tweaks I’d have made in some form in Windows had I just installed that. Power management configurations, settings to lock the machine on suspend, firewall settings, and wireless connection settings. None of these are earth-shattering settings to get some fundamental piece working.

Conclusion/Observations

Now at the beginning, I made it sound like it was all completely working when I installed. That may appear misleading considering all the files I describe above but they’re largely customizations as opposed to required steps. The basic tweaks section are the only real requirements.

As far as observations go, the ThinkPad x201 gets great battery life. It’s about the same as running Windows 7 and definitely better than any Linux distribution I’ve tried to run on it. WiFi is also better in OpenBSD than it was on Linux. Signal reception is more consistent whereas it seemed to fluctuate a considerable amount on Linux.

All in all, this is a pleasant surprise and I’ll be running OpenBSD on this laptop for the foreseeable future.

osdev howto openbsd
July 24, 2013

Glorious Nippon (Japan)

I’ve begun the long and arduous (but enjoyable) task of planning the honeymoon that Caroline and I will be taking. I’ve vacillated between several locations for the honeymoon. Much of it was concern over cost, time, and comfort traveling. Japan was the first real contender and after much thought it remained our best option. Other frontrunners included Turkey which has recently lost its shit and Montreal which takes just as long to get to as Japan from Seattle because there are no direct flights.

We’ll be traveling during the latter half of October of this year so as to enjoy the fall chill rather than the humid summers that most of the Land of the Rising Sun suffers under. Most people consulted that are in the know” regarding Japan say to visit either in October (fall colors) or in April (cherry blossoms). Additionally, the fall means fewer tourists so we can be all the more obvious as the crude Gaijin that we are.

Dreamliner”, our tube of pain

Our flying tube of pain of choice shall be a Boeing 787 Dreamliner” now that the FAA and Japan’s transportation agency have cleared the 787 for flight again. Boeing has redesigned the batteries which so inconveniently caught fire in the original designs. I figure six or seven months of the plane being in flight will demonstrate whether or not they’ve fixed the problem sufficiently. The plane’s unique design is supposed to:

  • lower cabin pressure
  • better humidity so that passengers are not dried out”
  • other stuff I can’t remember because effort

The fine purveyor of flying tube of pain services is All Nippon Airways (ANA) which consistently ranks as one of the best airlines. In fact, if you look at the top rated airlines, all but one of the 5/5 star airlines are Asian. I’ve wanted to fly on ANA for a long time now and went through great pains to be sure that we did not get put on a code-share United flight to Tokyo.

Tokyo

Once we arrive in Tokyo, we’ll be staying on the edge of the lovely bustling insanity that is Shinjuku Ward (think burrough if it were NYC). I spent a bit of time looking at the possibility of staying in hostels but the price savings to be had there are better if you’re a guy traveling alone. Girls staying in hostels complicates things as most of the time bathrooms are shared. While I like to think ourselves adventurous and open-minded, I don’t think either of us want Caroline to share a co-ed bathroom with mostly male strangers. Since shared bathrooms are out, this means looking for hostels that have private suites/rooms which ultimately end up being $10 cheaper than an actual hotel with 75% more hassle (hassle comes from things like curfews and lack of quiet). Since we’re not necessarily looking for travel companions (the main benefit of staying in a hostel), we don’t gain much for the trouble of trying to stay at one comfortably.

What we’ll do in Tokyo, who knows? You can spend a month there and barely scratch the surface. We’ll do some touristy things, wander through the madness of the city, try to eat at Jiro’s restaurant, and perhaps hit the fish market. Fortunately, I’ve got a nice big guide book to point us in the right direction as well as the internet to suggest destinations.

Osaka

After four glorious days in Tokyo, we’ll pack up, hop on a Shinkansen (bullet train for you uncultured swine), and head to Osaka. Osaka is the other major business center in Japan with a couple of notable surrounding cities:

  • Kyoto - the former capital of Japan and home to roughly one brazillion temples
  • Kobe - a luxury town with nice restaurants and delicious cow meat

Similar to Tokyo, we’ve got no specific plans. We’re simply going to get a little lost and see what we find. The word on the street is that the food in Osaka is unique in several ways and we do love to eat so we’ll probably just stuff ourselves with delicious and roll around in a swollen rice coma state.

The aforementioned plans only cover roughly 7-8 days of the 15 we’ll be in Glorious Nippon. We’ll need to find some other things to do to flesh out the rest of the trip. I’m almost certain Onsen (Japanese hot springs) will play a part as well as a visit to Hakone, a resort town outside of Tokyo near Mt. Fuji. Perhaps we’ll also squeeze in a trip to Kanazawa where there’s a rather large garden to stare at.

Sayōnara.

travel