Eating your own dogfood

"Eating your own dogfood" is a phrase that has a long history in the software industry. It refers to the technique of using your own product internally, yourselves, before, during, and after you deliver that product to your customers.

It's not always possible to practice this; sometimes your company writes a product for which you have no use, internally. I've been at those companies. I've also been at company where we do use our own products internally, and the difference is stunning.

So I was particularly pleased, earlier this week, when, at my day job, we upgraded our master internal server to the latest nightly build of the code. Not only are we continuing to eat our own dogfood, but this latest release contains a number of enhancements that I contributed to, so I'm excited to have that software in the hands of real users.

It will be a few more months before we decide that this software is sufficiently well-constructed that we are ready to deliver it to our customers, but this week was a major step in that process.

All of a sudden it seems like nothing but lawsuits

• Oracle sues Google


• HP sues Oracle


• Crossroads sues 3PAR


• Interval Research sues almost everyone.


• Everyone (via a class-action) sues almost everyone

What is going on in this world?

Just looking at the HP/Oracle thing for a bit, how will this story play out? I've seen it suggested that this all goes back to the pretexting mess of 2006; the New York Times article goes on to say:

According to “The Big Lie: Spying, Scandal and Ethical Collapse at Hewlett-Packard,” an authoritative account by the former BusinessWeek writer Anthony Bianco, Mr. Hurd was very involved in H.P.’s efforts to hunt down the leakers. After the scandal broke, he hijacked H.P.’s internal investigation, hiring an outside law firm and ordering it to report directly to him, instead of the board, which is the normal practice.

I've also read that Oracle had been recruiting Hurd for some time, and that they simply made their move when he became available. I know one thing for sure: I'm not the only one confused by what's going on!

When I was young, and just starting out in the computer field, Hewlett Packard was always associated with Management by Walking Around and In Search of Excellence and so forth:

we dropped by a small calculator and electronics store the other day to buy a programmable calculator. The salesman's product knowledge, enthusiasm, and interest in us were striking and naturally we were inquisitive. As it happened, he was not a store employee at all, but a twenty-eight-year-old Hewlett-Packard (HP) development engineer getting some first-hand experience in the users' response to the HP product line. We had heard that a typical assignment for a new MBA or electrical engineer was to get involved in a job that included the practical aspects of product introduction. Damn! Here was an HP engineer behaving as enthusiastically as any salesman you'd ever want to see.

HP: Wha' happen?

Drivers versus passengers

Interesting essay/rant by Dave Kellogg on his blog about drivers versus passengers:

Strategically and operationally, I think there is a huge difference between drivers and passengers that comes out when they are placed in a new situation. When placed in their next company:

• Drivers assess the situation and develop strategies and tactics appropriate for the new reality.


• Passengers do what worked last time.

I think this is fair criticism:

I am repeatedly stunned by the number of otherwise very intelligent people who show up and do what worked last time. Often with the very same cohort / entourage with whom they did it.

It is, however, a very easy trap to fall into.

Meanwhile, completely unrelated except that I was reading his essay at about the same time, Paul Graham makes an interesting point about the future of Venture Capital funding:

Different terms for different investors is clearly the way of the future. Markets always evolve toward higher resolution.

Having had opportunities to work for companies held captive by their investors, as well as companies which had other ways to manage their finances, I have seen both the good and the bad of the VC process. Graham has proposed a variety of interesting ideas over the last decade to solve some of the problems, and his essays are always intriguing.

It's not magic, it's physics!

Step 1: Shave head

Step 2: Run really fast.

Step 3: Kick it really hard, in just the right way.

Result: Bend it like Roberto Carlos!

It's all a matter of the tradeoff between aerodynamics and gravity:

one can identify sports dominated by aerodynamics (table tennis, golf and tennis) and sports dominated by gravity (basketball and handball). In between, we find sports where both gravity and aerodynamics play a comparable role (soccer, volleyball and baseball). Indeed, in the first category of sports, the spin is systematically used, while it is not relevant in the second category, and it only appears occasionally in the third one, in order to produce surprising trajectories.

This is what happens when you play chess!

The video proves it!


(I love the way he prowls around furiously, scowls, and buries his head in his hands constantly at the board; classic Garry imitation!)

Anti Semi Joins

I happened across an unfamiliar term today, "anti semi join".

I encountered it in this nice little writeup from the MS SQL support team.

So, of course, I went hunting for more information about semi joins and anti semi joins. It turns out that I had a fairly good understanding of them already, but it was nice to find several well-written articles describing them, including:

• Craig Freedman's Introduction to Joins


• Roger Schrag's Speeding Up Queries with Semi-Joins and Anti-Joins


• Conor Cunningham's Conor vs. Anti-Semi-Join Reordering

Cunningham's point about not needing to do "TOP 1" in existence sub-queries is a very good one.

Query optimizers are incredibly sophisticated algorithms. I've only been studying them for 25 years. Eventually I'll understand them :)

It's crazy days in the software business again

Check out this wild story on Mike Arrington's TechCrunch today. Here's a snip:

Sources close to Google tell us that about 80% of people stay when they’re offered a counter to a Facebook offer. But some still leave.