from Friday through the end of May, you can sign up for an 8GB iPhone 3G in Japan for a grand total of zero. ... so long as you sign up for a new two-year contract ....

...

SoftBank subscribers will still have to pay monthly voice and data charges, of course, although as AppleInsider notes, the carrier is chopping its standard data plan (about US$62 a month) by more than 25 percent.

The charitable interpretation is, I guess, that it's simply assumed that everybody pays for a monthly cell phone service contract, so that should reasonably be excluded. Since I represent a walking, breathing refutation of that assumption, I'm more prone to be irked by people sweeping a thousand dollars under the rug.

It doesn't take much further inspection to notice the prominent product placement for a certain retailer, which makes me wonder about the ratio of "journalism" to "advertising" in this "story". It's a reminder that print newspapers had some standards, however tattered and breached, that we're probably going to miss in the online world.

The speaker works in the field, managing installation of broadcast television studios. The following is my paraphrase of the speaker unless otherwise noted.

Why is this delayed? No project management. Fighting between agencies, no one in charge, etc.

Q: Whose project is this? A: Nobodies. Nobody is in charge of it. Stakeholders: NAB, the public, FCC, NTIA, consumer electronics companies, broadcast equipment companies, system integrators, cable companies, satellite providers.

Why was it delayed? The NAB went to the incoming Obama administration warning that some people weren't ready for the transition, in total about 1% of all Americans.

The history of analog TV. Some tech terms. Ampex developed the first commercial video tape recorder. ATSC is the name of the technical standard for digital TV in the US and the committee that came up with the standard. It used to be called the Grand Alliance, although that term is out of use. Europe has a different standard. DTV is not the same as high definition. ASTC is a set of different standards, some of which are hi-def and some of which are not. Up to five standard-def streams fit in one former analog channel. The major US networks have chosen different ways to divide up their channels; only two are actually using the full resolution, 1920 by 1080.

HD dates back to before 1980. Broadcasters wanted to switch to HD to gain a competitive advantage over VCRs and other new competition. "Now they're getting their wish but it's almost too late, right? ... Broadcasters are almost playing catchup at this point."

1600 local stations already broadcast in DTV in parallel with their analog. Total spending for equipment, government coupons, etc, is roughly five billion dollars. The government plans to auction freed up spectrum for 20 billion dollars (after apportioning some of it to fire departments and other emergency responders). The current transation plan, such as it is, dates back to 1996. There was a previous deadline in 2003 that passed without consequence.

The use of television as an emergency communication system. On 9/11, many internet sites were not updated in a timely fashion, and only TV had fresh data. That's changed somewhat since 2001.

HD is a wider aspect ratio than SD, and there was discussion of selling the bars on the side as advertising. When I was doing the KCAL transition, Neilsen didn't have high-definition boxes.

Q: What about the adoption of HD/digital in less developed countries? A: The prices for HD are dropping because of investment here. At the same time, all of the existing SD equipment in the US is obsolete and will be sold cheaply.

Q: What would you do as president in 2001 to make this run better? A: Appoint a project manager ... this was run by politicians and their interests ... do traditional project management, get the stakeholders, get a plan .... Q: So you are saying we can use this as an example of what happens without project management? A: yes, it's a trainwreck.

This presentation was a bit disappointing. I've been following the DTV switch for years, through my own reading and, of course, from Harry Shearer's ongoing "Digital Wonderland" updates on Le Show. I though this event might feature some insights, stories from the trenches, etc, but it was relatively shallow.

The first attendee introduced herself as requested, and her motivation for attending was that her company is starting a new Scrum project. The second attendee introduced himself as on a date with his wife, the first attendee. The room then erupted in laughter.

The rest is my paraphrase of highlights from the speaker, Sutanu Ghosh, unless otherwise noted.

The PMBOK (Project Management Body of Knowledge) does not conflict with Scrum. PMBOK provides a framework; Scrum is a specific practice that can be understood in that framework.

Toyota is big on agile manufacturing, which is very similar to agile programming.

Existing processes may be too heavy and offer too little reward. This is a premise of six sigma—pruning useless processes.

Information saturation: the more aware you are of your teammates' activities, the less documentation you need.

Traditional view: the trilemma of scope, time, and cost. In the agile view, think of time first, then resources (cost), then features (scope). Joel's note: I'm increasingly questioning the validity of this model, either flavor, in software. The overwhelming cost of software development is salaries, so aren't time and cost co-variant instead of in opposition?

Stopping the release, or halting the assembly line, is positive to the extent that it represents the voice of the customer. E.g., the customer would not want this car or this software with this defect. But it shouldn't become a gotcha war between test and engineering.

Under what conditions is an Agile method applicable?

Scrum is extremely applicable to adding incremental changes to an existing system. It is also applicable to trying a novel assembly of existing parts. It is less applicable to problems with uncertainty in both sides: building a big complex whole with undeveloped parts. Scrum with long-term commitments is not Scrum, it's rolling wave.

Scrum requires sequestering your resources; they should be wholly committed to only one project for the duration of the sprint.

Discussion: if there are external commitments going far into the future, how can you use Scrum to get internal commitments to match? This comes back to the basic paradox of development: sponsors understandably want predictability out of an inherently unpredictable process. One partial answer is that Scrum is not suitable for an environment that can't accept scope churn. Joel's note: to me, the real difference is simply style. Declaring that something must happen in two years, and even threatening contractual consequences, does not cause it to happen. A waterfall process will spend weeks or months understanding the problem and then return an estimate of whether or not it's possible and what the risks are. An agile process will get started on something that seems like the first step today, but refuse to answer whether or not it's possible until it's almost done, which may or may not be in two years.

As teams grow, communication effectiveness drops exponentially. A team of 7 +/- 2 has optimal communication.

The product owner represents the customer. In Six Sigma, this is the concept of customer voice as opposed to engineering voice. The job of the product owner is not to give the team activities to do. It's to give them stories to realize, stories which can be sold to customers.

One difference between PMBOK and Scrum (and probably between PMBOK and reality) is that, in PMBOK, the project manager implicitly has some authority. In Scrum (Joel's note: and in my experience), the project manager does not have authority, only influence.

Scrum principles can help keep the counter-productive dart-throwers out. Pigs vs chickens. The voice of the customer is very important. People with an interest but not stake should not have influence.

Stories should be: independent, negotiable, valuable, small, testable.

A traditional requirements document tends to mix up structure and behavior. Attendee: I was taught, years ago, that the functional specification is written in the language of the customer. The design spec is written in the language of the engineer.

The recommended length of time for a sprint is four to six weeks. Some say two weeks.

People who write user stories often forget things like regulatory compliance.

Technical debt is cost that will be incurred later in product lifecycle due to negligence now.

Scrum plans are based on bottom-up estimates of how to meet top-down targets. There is still uncertainty, just as with other engineering. Accuracy is much more important than precision.

My questions:

• I've been working in software and IT since 1995, which is long enough to see some ideas come back under new names. (For example, SaaS, which used to be ASP). But I haven't seen a direct match to Scrum. What was the closest previous trend to Scrum, and when did it fade, and why?
• "It's called a sprint because you are running hard. As the program manager you drive the team to work very hard." From my reading I though the essence of Scrum was management putting giving a team three things: isolation from distraction, a clear target (or a commitment to give clear "warmer/colder" responses to output), and freedom to choose their approach; and this the ultimate purpose of all this is simply to get the team self-motivating. So why do you say that the program manager drives the team? Isn't that antithetical to Scrum?

The Obama administration is expected to impose a cap of $500,000 for top executives at companies that receive large amounts of bailout money, according to people familiar with the plan.

[...]

"That is pretty draconian — $500,000 is not a lot of money, particularly if there is no bonus,” said James F. Reda [of] a compensation consulting firm.

(New York Times)

The hell with nationalization! If we hear any more backtalk from Wall-Streeters-with-their-hands-out on how half a million dollars is "not a lot of money" it will be high time to round 'em all up and sell 'em to the highest bidder.

How much would you pay for a Vikram Pandit, John Thain, or Ken Lewis?

A further comment: you don't have to have a Che Guevara poster on your wall to perceive that people with money and power use that money and power to get more money and power. The current crisis is increasing the awareness of this basic fact. Perhaps that will lead to some changed laws. But what would help move the balance a bit further for a bit longer might be changes in mores and norms.

After all, CEO pay didn't start skyrocketing because CEOs suddenly got greedy. They've always been greedy. Pay started skyrocketing because that class of people, the executives and the board members and all that, had the dawning realization that they could simply give themselves more and more money and nothing bad would happen to them. Laws and SEC regulations count as "something bad", but can be circumvented. A general resetting of expectations would, I suspect, last a bit longer. Although I admit that the only way I can think to accomplish that is to send a lot of these people to prison or a life of well-earned destitution.