The World As Best As I Remember It : General

Corporate Blogs -- Gradations of Value?

philipsu — Fri, 16 Jun 2006 03:43:00 GMT

First, let me say once again, there has been no corporate pressure whatsoever to cut short my previous posting (Broken Windows Theory). Nobody has said, or even implied, that I need to change anything about what I said. So conspiracy theorists, please rest assured that The Man is not out there monitoring and censoring the blog world. Seriously.

I pulled the content of the posting because productive discussion wasn’t happening. Of the 160+ comments, about five have had any real value from an “open minds, open discussion” point of view. I also pulled the content (once again, completely self-initiated with no pressure whatsoever from anybody) because there is enough internal debate within Microsoft about the value and ethics of blogging which I’d like resolved.

[Follow-up: I've restored the original post, after much internal discussion. Essentially, pulling the content was causing undue attention.]

Internal Debate

Many perspectives have been voiced to me, both publicly and privately, debating the value and ethics of employee blogging. Here, by “employee blogging,” I mean “blog entries that are openly identified as being written by Microsoft employees.” The rough gist of internal feedback from Microsoft employees falls in these categories:

Thank goodness someone is talking about this. “Kudos for having the courage to shed light on these critical issues.” “It’s great that we’re having open and insightful discussion about this.”
You need to put the entire post back up. Some folks are quite concerned that, with Scoble leaving this week and what not, there will be increased fervor behind conspiracy theories about how I’ve been silenced, shipped to Siberia, etc. This sort of feedback is much more concerned about posts staying up from a PR/media perspective, regardless of the content of the post. (Let me say here once again, for those who have deep-seated theories – my original post was shortened unilaterally by me. I was at no time pressured to remove any part of it.)
Employee blogs should be an extension of the company message. Folks in this category would say that the moment I identified myself as a Microsoft employee, my message should be on target with the corporation’s message, building a positive image, connecting positively with customers, etc.

Let’s Agree on Goals

From my perspective, it’s not a free speech issue. I’m employed by Microsoft, so there’s a valid discussion as to what sorts of posts are allowed for me to make as an employee of the company. Conditioned in my employment can indeed be restrictions on what I should and shouldn’t say – I buy off on that 100%. (For the last time, though, please remember: no one has pressured me to change anything.)

Second, the simple case that I think everyone agrees on is that nothing confidential should ever be divulged. This is where Mini-Microsoft, as entertaining of a read as it is, crosses a line. That’s also the reason it can only remain up as long as it’s anonymous.

The more interesting debate I’d like to have is not whether employees can or can’t post certain things, but should they. I have no interest whatsoever in the set of things that are clearly against company policy to post. I’m much more interested in the spectrum of things where people, even internally in Microsoft, disagree.

So How Does It Net Out?

The bulk of the internal feedback I have gotten falls on the side of encouraging posts like Broken Windows Theory. The vast majority of emails I’ve received have to do with how the article has opened up important issues for discussion. Folks in this camp say that Scoble has given a human face to Microsoft, has made Microsoft more accessible to the majority of customers. The openness, and in some sense, the vulnerability, of both addressing our strengths and discussing our weaknesses has been refreshing, these folks would say.

Another camp would say that blogs are a key part of how a company is perceived, and that they act as a megaphone of both positive and negative opinion. Part of an employee’s responsibility, then, is to at all times help build and reinforce that positive image.

What’s most interesting to me is that even within the company, we don’t quite agree on whether Broken Windows Theory is a net positive or net negative. If I take it purely based on numbers, the overwhelming majority of employees writing in say that it’s a positive thing. But I see merits to both sides of the discussion.

Thoughts?

Red Rain

philipsu — Sun, 09 Apr 2006 11:48:00 GMT

Tuesday morning, 9:00 AM. It’s sunny in the flat outskirts of San Antonio. Wisps of summer’s residual heat serve as reminders that it is, after all, only September. A child cries loudly on the tarmac. Strangers – stand, sit, stroll, play – blithely near the child. Not to worry: surely its mother is at hand.

Inside the hangar, thousands more. Standing in lines, sitting in groups, sleeping on cots. Everywhere the sights, the smells, the sounds, the touch – of people. Welcome to Kelly Air Force Base, Building 1536. Welcome to the largest makeshift Red Cross shelter in San Antonio.

Welcome to the aftermath of Katrina.

With the Trust of a Child

Late last summer, I seriously considered a radical career change. After several great years in the software industry, I began to feel like I needed to put my skills to something that more directly helps people in need. I was offered the opportunity to head the IT department of International Justice Mission, an ingenious non-profit with the noble vision of saving the oppressed and defenseless throughout the world. For weeks, my wife and I debated the prospect of quitting my cushy day job, moving across the country, and serving one of the world’s best non-profits. In the end, for a variety of complex reasons, we made the decision to stay at Microsoft. However, I set my mind to doing something noble for the world, even here in Seattle. Microsoft does, after all, have a long history and culture of philanthropy.

I talked with my manager about this decision and my subsequent resolve to support humanitarian efforts at Microsoft. Little did I know that I would only need to wait a few weeks before the opportunity presented itself.

Barely a week after Hurricane Katrina hit, my vice president and my manager offered me a chance to bring several developers and cases full of Tablet PCs down to affected areas. I sent out a few emails, placed a few calls, and found out that the San Antonio Microsoft office was helping local Red Cross shelters and needed developers. My wife and two teammates volunteered to go with me. Before you knew it, we were on a plane headed to San Antonio, carrying cases full of computer equipment prepared and donated by Microsoft’s Mobile Platforms Division.

Buried in a Sheltered Place in this Town

Suffice it to say that I had no idea what I was in for. I was told they needed manpower, computers, and perhaps some software automation. I assumed we would arrive, set up shop, and begin coding.

The Disaster Relief Fairyland in my mind goes something like this:

Several hours after the disaster, the military and the Red Cross have restored order. Shelters have been set up according to Red Cross’ decades of worldwide disaster response experience. The tired are resting, the hungry are fed, the lost are found, the grieving are comforted. FEMA, Red Cross officials, the military, and local politicians are doing a rousing rendition of USA for Africa’s We Are The World. A lone heartfelt tear rests on the cheek of a young audience member as the leading FEMA representative bursts into, “There’s a choice we’re making… we’re saving our own lives…”

The reality, which you by now have guessed or read about but could never truly believe unless you see, hear, and smell it yourself, was that we arrived a little over a week after the hurricane to the deafening roar of 17,000 fans ready to rock.

Hehe. Somewhere deep inside of me, I tend to indulge the Disaster Relief Fairyland more than I probably should.

We actually arrived at Kelly Air Force Base, where 17,000 displaced Katrina evacuees were housed in two mega-structures. The larger of the two buildings, Kelly 1536, is a hangar roughly a third of a mile long. It covers the area of fifteen football fields under one giant enclosure. You’d get the same effect if you joined four Costco warehouses together, removed all the shelving and merchandise, and invited all the shoppers’ extended families to live there. Cots were packed so tightly in this space that it seemed you could walk the building end to end without touching the ground.

I Can’t Watch Anymore… No More Denial…

But I’m not here to tell you about the overall experience helping out at the San Antonio Red Cross shelters. (If you’re really enamored, feel free to read the team’s blog on that subject). I’m here to tell you how software can radically revolutionize disaster response.

First, let’s talk about missing people. In ad-hoc mass evacuations, families get inadvertently separated. When we arrived in San Antonio, there were – this is not a typo – sixteen separate databases tracking missing people (well, fourteen, if you discount “paper” and “by memory” as non-ACID databases). This meant that separated family members were very difficult to find. Which database should volunteers search in? The International Red Cross’? Wait a minute, I hear the American Red Cross’ is actually the official one. Then there’s KatrinaSafe. But the City of San Antonio insists on their home-grown database…

None of the databases supported photographs of missing persons. A simple enough feature, one that wouldn’t be at all hard to implement. Why photographs? Well, some folks can’t identify themselves. In one of the shelters, a lost developmentally-challenged child was left unfed for a day because everyone assumed he was being taken care of. The parents, needless to say, had no way of searching for their child. A database that included photographs of people that can’t identify themselves would be a great way to solve this problem.

Take inventory tracking. Everyday, truckloads of donations would be unloaded into the hangar. Inevitably, the donations consisted mostly of toys. Especially crayons. I am not making this up. I saw tables stacked high with Crayola crayons. Fat ones. Thin ones. Mega packs. Scented ones. Pastel. The only thing they had in common was that every last box was unused.

Donated inventory was tracked on legal pads. Truckloads of goods counted on long yellow sheets of paper written in pencil. Do we have blankets? I don’t know – check the pad. Where’s the pad? I don’t know – under the crayons, perhaps. Which crayons?! There are like a gazillion tables piled full of crayons! It was ridiculous. It would have been Seinfeld-scale comedic if it wasn’t so tragic.

Imagine with me for a moment that Dell managed Red Cross donations. (Give this thought a moment to soak in… it’s a profound one… Ok, now you’re ready.) Donors would be able to check inventories in San Antonio and notice that diapers were in desperate need, instead of, say, vintage crayon collections. Shortages in one shelter could be offset by surpluses in another. Donors’ credit cards wouldn’t be charged until the very second that food enters an evacuee’s mouth. Every quarter, major donors would be called in and publicly spanked in front of their competitors for not having as timely or as high-quality donations. (Ok, that last bit is probably a dose of Dell-reality that we could do without in the world of humanitarian aid. But everything else stands.)

One last example of how software could greatly improve humanitarian response: let’s talk about volunteer tracking. Every day at Kelly Air Force Base, more than a thousand volunteers would come and go from the two gigantic buildings. At first, volunteers were tracked on whiteboards with Post-It notes (written in crayon, no doubt). A thousand volunteers a day! You can imagine the mess that this was. (“Uh, Bob, I think Joanne’s Post-It just fell off the board… where was she working again?”).

It’s mind-blowing how much the world of humanitarian aid needs software.

This Place Is So Quiet, Sensing That Storm

Why isn’t there Red Cross 9.0, a shiny little CD that every shelter manager gets on the first day of a disaster? It’d have all your favorite hits from the ‘80’s, as well as software that tracks missing persons, manages inventory, allocates volunteers, and keeps you connected to nearby shelter managers. It’d even search for extraterrestrials while your computer is idle.

While in San Antonio, our team wrote software that helped track volunteers at Kelly Air Force Base. By the time we were done, all volunteers could be registered into one common database from multiple terminals via a simple client-server design. Allocations of volunteers to different areas could be easily monitored and adjusted. Data re-entry was eliminated. No more whiteboards, no more Post-Its, no more blunted crayons. Similar systems were set up by the great folks from RackSpace who worked day and night to create a web interface based on Ruby on Rails. (I tried earnestly to embrace Linux, learn Ruby, and extend their system. But I kept getting odd failures in package installation. Seriously. I would frankly love to be able to say at dinner parties that I worked in a language that was “on rails,” but alas, it was not meant to be.)

The volunteers working at the shelters loved the software. It saved a ton of time and was a much-needed improvement over what was available. We deployed it at two other shelters, and FEMA also began to use it (hold the jokes, please). After its success at multiple locations, I was riding high on hopes that the Red Cross would spread the word and start including it as a basic part of running any shelter.

But that was not to be the case. As dust from the aftermath of Katrina settled, I got in touch with whoever I could at the Red Cross to hand over the volunteer tracking software. Free of charge! Ready to use! Clearly valuable! I was shamelessly begging the Red Cross to take it and use it. I contacted the head of IT at San Antonio’s Red Cross. A disaster-management consultant demoed it to the Red Cross and FEMA in DC. I tried to find anyone – anyone – that had an interest.

Dead silence. (Imagine tumbleweed blowing gently across the page)

I’m not sure what it is, but no one at the Red Cross seems to have an interest in the strategic long-term importance of software in improving Red Cross’ effectiveness. Whither Red Cross 9.0? How long will shelters be run on paper, and technology be cobbled together ad-hoc after every disaster, only to be discarded after each use? It’s a waste and a tragedy.

If you’re to remember only two things from this posting:

1) Software can make a huge difference to the effectiveness of humanitarian response. I’ve seen it with mine own eyes. The amount of waste and inefficiency in today’s system is tragic. We might as well put fistfuls of cash in a 50-gallon oil drum, toss in some crayons, and light the whole shebang on fire.

2) The Red Cross needs to embrace the strategic importance of software in their mission. Something is very wrong if developers from all sorts of companies are writing custom software for Red Cross shelters during each emergency. Something is very wrong if we can’t even give away free software to the Red Cross so that it can be used in the next disaster.

I belabor the point. Lest we end on a down note: the experience of volunteering in San Antonio was hugely rewarding and personally fulfilling. The dedication of the volunteers on the ground was awe-inspiring. I only wish there was a way to get the Red Cross to understand how truly critical software is to their mission. In the end, sheer human willpower was able to triumph over a tragedy of process and infrastructure.

Here’s to better use of software in our next disaster.

[From the Random Facts department: Peter Gabriel was apparently not making it up. There have indeed been incidents of red rain in parts of the world, the most famous of which was in Kerala, India during 2001. Read more at Wikipedia.]

Brin and Bear It

philipsu — Tue, 15 Feb 2005 09:21:00 GMT

I attended the same high school as Sergey Brin, co-founder of a verb called "Google." Sergey dropped out in 1990, the year I started at Eleanor Roosevelt High School. By all accounts, he was extremely bright. But from the little I've heard of him via direct anonymous sources (*cough* former-math-teacher *cough*), he was quite cocky about his intellect, often trying to prove teachers of various subjects (read: Math) just how wrong they were. But this is utterly forgivable, given that almost any successful self-respecting computer scientist went through such a phase during his teenage years. (If, however, you are remembered fondly by all your high school teachers and friends, chances are incredibly good that you're also a handsome business major now working at a consulting company who's wandered onto this blog by accident. But I digress.)

Who's Line Is It Anyway?

There's been quite a lot written about Sergey Brin, but hardly enough about a much lesser known Brin -- Dr. Michael Brin. In addition to being precocious Sergey's father, Dr. Brin was also my favorite professor at the University of Maryland. He taught what was easily the most enjoyable class in my college career. I would have taken more of Dr. Brin's classes if the university system allowed it, but alas a computer scientist of meager intellectual means couldn't afford the heavy IQ tolls imposed by his graduate-level math courses.

Dr. Brin is a first-rate orator who begins every class with a smoke. Well, usually he'd show up, pose a rather vexing problem, and then escape for a cigarette in the midst of the ensuing confusion. He expected an answer from the class by the time he returned. Invariably no answer was forthcoming, since that's the entire point of a vexing, smoke-break-enabling problem. (At this juncture, I know those of you in technical fields are just dying for an example of a vexing Dr. Brin problem, just so you can solve it faster than I could smoke a cigarette to prove me dumb. I'll bite and indulge. Answers at the end. And I don't smoke.)

Vexing Problem #1. You have an 8x8 grid of squares and 31 dominoes. Each square in the grid is exactly half a domino large, such that each domino placed on the grid covers two squares. How do you arrange the dominoes such that all the squares in the grid are covered except for the two squares in the outermost corners of the grid diagonal from each other? If this cannot be done, provide a proof of why not.

Vexing Problem #2. You have a solid bar of chocolate that's pre-scored by the helpful chocolate bunnies at the Hershey factory into four rows and eight columns (such that if you broke the chocolate along those lines, you'd end up with 32 pieces of chocolate). Prove the minimum number of breaks that are required to break the bar into its 32 pieces (you may not stack or otherwise cheat on how you count breaks -- a break is any one fracture along a pre-scored line on one solid piece of chocolate).

Vexing Problem #3. You are in a maze of twisty little passages, all alike. (Ho ho! You either laugh or scratch at this one, but in either case Dr. Brin would certainly not have posed this most vexing of problems.)

In addition to loving wacky problems, true enjoyment of a Dr. Brin class can only be had if you check your ego at the door. Almost half the statistics class he taught dropped out after the first session because they couldn't handle his assault and battery on their sense of self. Attending his classes was like experiencing the drill sergeant in Full Metal Jacket -- unless you braced for impact, your ego would likely leave wounded or even dead.

As for me, I was his Private Gomer Pyle, and I loved every minute of it. Dr. Brin would often hand back graded tests with a simple, "My sincere condolences." He'd openly make fun of wrong answers to his questions, which I believe he found amusing the way we find children amusing when they ask where babies come from. His spirited and equal-opportunity barrage on all students backfired only once in all my recollection:

Dr. Brin: It's time for an Odyssey of the Mind -- time for our statistics final.
[Most students understand the implications and put away their calculators. One student, let's call him "Bart," obviously didn't follow.]
Dr. Brin: Bart. You have your calculator on your desk. What -- do you ~love~ your calculator?
[A few students giggle nervously]
Bart (startled, confused, and a bit hurt): Every night.
[Dr. Brin is so caught off-guard by this genuinely clever answer, made in a spirit so congruent with how he ran his classes, that he allows Bart the unique privilege of using his calculator during the test.]

If crushed egos was the price for true learning, then I was ready for the abuse. Dr. Michael Brin's classes were always great and full of laughter (albeit at your own expense). I didn't mind being constantly reminded that my intellect was inferior. It was refreshing in an inexplicable, self-effacing way. And it was worth it to learn from one of the best.

How I (Almost) Made My First Million

One of the last conversations I remember having with Dr. Brin was one which I initiated. I was blissfully raving about this great new search site I had found, then in Alpha mode, called Google. It was done by some Stanford students! It didn't have ads! It was much better than then-market-leader AltaVista, and I wanted Dr. Brin to try it out.

He matter-of-factly replied that his son Sergey was a key member of the Google team. He also suggested that I write his son to see if there might be an employment opportunity for me.

Sergey never wrote back.

So here's to Dr. Michael Brin: May the laughter never end, may the smoking never stop. May only the humble learn at your fount of intellect, and may you always find interesting problems to puzzle and amuse.

Thiiiiiis Is Hoooow We Doooo It... (This is how we do it!)

You've stayed long enough. You've scrolled far enough. You want answers, and you want them now. And you're not going to settle for me promising them to you next week (We all know what happened last time when I suggested that. I hope you still have your day job as a programmer. If not, in the words of Dr. Brin, "My sincere condolences.").

Vexing Solution #1: You can't cover all squares but two diagonal outer corners. The proof is in the custard: Imagine the 8x8 grid was a chess board. Each domino would clearly have to cover exactly one white and one black square, no matter how you placed it on the board. Two diagonal outer corners are going to be the same color -- either both black, or both white. Given that you start with 32 uncovered black squares and 32 uncovered white squares, it's impossible to leave two squares of the same color uncovered if each domino covers one square of each color. This is called a parity problem, and is an excellent example of how many seemingly complicated problems are actually easily solved (and proven!) when you view them from the right perspective. (All three of the business majors that are reading this blog by accident should turn this principle into a trendy bestselling business book. I feel one coming. Pull my finger.)

Vexing Solution #2: I ask this one in interviews all the time. Inevitably, people start testing all sorts of conjectures. (You know the true computer scientists right away because they insist that breaking the chocolate into successive halves must be the best way. It just must!). Most interview candidates don't realize that to prove a minimum by such a method would require exhaustively testing all the possibilities, which is clearly too many (even for those of you who were just about to respond with some comment about symmetry [or how your job has gone to India and now you're browsing the web all day]). Far better to learn from Vexing Solution #1 and see this problem from a different angle. Here's the angle: no matter how you break a piece of chocolate along a pre-scored line, you always end up with the same thing -- two pieces of chocolate. Any break always adds one and only one to the total number of separate pieces of chocolate you have. So by induction, the number of breaks it takes to break one solid bar of chocolate into its 32 constituent pieces is always 31, no matter how you choose to break it. So be you Row-ist, Column-ist, Half-ist, or Whatever-Other-Compulsive-Chocolate-Breaking-Strateg-ist, know in your heart of hearts that it doesn't matter. 31. Always.

Prologue

philipsu — Thu, 01 Jul 2004 02:32:00 GMT

After all this time, finally got around to starting this blog. The World As Best As I Remember It will be centered around topics in software development management -- not technology itself, but the management of software development process and the industry in general. Rants and raves about the world sprinkled in between for crunchiness.

Quick professional bio: I'm a software development manager on the Tablet PC team at Microsoft. Previously a developer on Microsoft Money and Microsoft Word. Web developer at the University of Maryland. Taught C++ and XML at University of Washington. Co-authored Building Tablet PC Applications, the Microsoft Press book for Tablet developers.

Glad to sign on. Let's go!