David's Blague

n does not equal n plus one

Project managers are often asked whether they can add a feature without impacting the schedule. The answer to this question is: no. If a fixed list of tasks can be completed during some duration of time, completing all those tasks plus one more requires additional time.

I met Casey on October 21, 2001. She was at Eastern Market that day, walking with my future wife. I was out of sorts, like I always am on that date, and my recollection is that I failed to charm my future wife, but succeeded in charming her Sheltie. They both charmed me quite thoroughly

Casey died in my arms on Sunday; being old, and full of days.

It's a time of transitions for our family, and Casey knew it, like she always seemed to know these things. She came into the family as Aunt Kate's dog, when JT left for college. She went to live with my wife when Kate made that transition herself. When we brought home our daughter, then a puppy, then our son, Casey always seemed to know what was coming.

She loved puppies, both human and canine.

We are working on the house, in preparation for selling it, so when Casey died, we were at the Bell House with my in-laws, where Casey lived as a puppy. There is a certain symmetry to it- we buried her next to Rosie, whom she knew as a puppy. Kirby, our little Cairn Terrier, didn't eat for three days- moping around the house, probably wondering where Casey had gone.

Our youngest wanted to sing a song, so we sang Good Bye, Old Paint.

The kids will remember Casey as the first dog they knew. My wife will remember her as a friend during her coming of age. I will remember her as the first dog that trained me so thoroughly, I knew what she needed solely based upon the bark, whine, or clicking of toenails.

Good night, sweet Casey. Wait for us by the white picket fence.

A fellow DC Dork (e.g. attendee of Dorkbot DC) asked for a visual aid after I was unable to describe this soldering technique via email. Basically, you coil up a piece of solder wire and let it be a jig for you, but since I've already proven my inability to describe it verbally, here's my attempt at modeling it. This is a piece of Radio Shack perfboard, with my butane soldering iron, and a diode sticking out of the perfboard in the back. Note the technique of coiling the solder on the table and reaching one end upward:


I can't remember who taught me this technique (neither could my fellow dorks), but it has saved more soldering hours than every other jig, clamp, and hack that I've got.

As I have occasionally done, I am responding to a Paul Graham essay. It also happens that I have had two (count 'em, two) people ask my advice on this matter, neither of which was Paul Graham, but one of whom mentioned his essay. As is my custom, this is a point-by-point real life satire, so really you gotta go read Paul Graham first.

1. Changing cofounders is easy.

"Cofounder" is a title, like "president" or "administrative assistant." If you get into business with a douche bag, get out of business with that person, even if they are a "cofounder." If you find someone later in the process who would like to join you, but wants to have "cofounder" on her business cards, you're an idiot to let this stop you from hiring the person.

2. Launch at the right time.

You've probably got a lot of ideas. When it's time to jump (it's not always time to jump), pick one that that is right for the time of jumping. Timing is important. Sooner is not always better- especially if starting the new thing means stopping something else (it almost always does), and that especially-especially if stopping something else means losing an income stream.

3. Evolve your execution.

Execution is critical. With good timing and good execution, even a mediocre idea can work. In fact, good execution is optional, as long as it keeps getting better. There aren't too many mistakes you can make early on that you can't recover from later, as long as your execution improves.

4. Understand your competition.

If your new idea doesn't fit into some ecosystem of competitors, consider that a red flag. Whatever you do, do it faster, cheaper, better than they do. Think of your competitors as (among other things) the folks who do free market research for you.

5. Better to be forgotten about and paid than loved.

Don't go for users that love you. Go for users that don't think twice about giving you money. Do something that is so critical, or so easy, and do it well enough that nobody ever thinks about leaving. If they love you, they're probably enthusiasts, and will probably love someone else in your place before you can blink.

6. Offer totally unnecessary customer service.

If your customers need it, give them a way to get it for themselves. If you rely too much on customer service, you will eventually suck. Nobody is as motivated as me to help me, so let me help myself, darn it.

7. Be careful what you measure.

This was, admittedly, Paul Graham's corollary. If you measure your number of users, your number of users will go up and up, sure. But you still might suck. Corollary: you make what you measure, so why don't you measure your money?

8. Don't borrow.

Spend your own money, and go with your instincts about spending your own money. It's one way to keep yourself honest about spending habits. If you got that money by selling part of your company, remember that it's now yours, and spend it accordingly. The best reason to spend money, is that the money you're spending has already made you more than you're spending.

9. Don't live on Ramen.

The main reason not to live on Ramen is that you don't have to. This isn't to say that you shouldn't be careful with your money, because you should. But skipping paychecks is a very bad habit, even in your early, early days. It's a form of borrowing, and rule #8 still (mostly) applies. If you can't make ends meet, either find someone to help you out (e.g. parents, friends, etc.), or get a job, or get some gosh-darn customers. But don't live desperately.

10. Stay focused.

On money. Focus on money. Your startup will succeed or fail on money. If you can't focus on money, team up early with someone who can. Money is boring, so it's hard to stay focused on it. Ideas are fun, so they're not. But if you're starting a company (as opposed to, say, pursuing a hobby or completing a project), focus on money.

11. Get demoralized.

Get demoralized early on. If you need to, have someone punch you in the nose or insult your parentage, but whatever happens, take an early, pessimistic view of what you're doing, and write it down. Think of all the ways it can fail, and catalog them in order of likelihood. You are very likely going to fail, so try to imagine how it's going to happen. Then, don't do that. Easy enough.

12. Fail.

And when you've failed utterly, stop doing what you were doing. Do something else. You might have heard this called "giving up." Fine. But once you've failed, understand why you've failed. If you have co-founders, customers, investors, or friends, get their help understanding why you've failed. Next time, fail differently. Or succeed. You'll probably do one or the other.

13. Never rely on a deal.

Succeeding or failing on the strength of a single deal is boring. It also leads to bad deals, since, if you need it to succeed, you'll probably accept crappy terms. Succeed or fail on your idea, timing, and execution.


If I had to pick one of these points, I guess it would be my number ten, which best drives home the point that you should start a business for the purpose of making money doing whatever that business does. It goes without saying that you shouldn't start a business doing something you dislike, and if you dislike focusing on money, it might be business that you dislike. Consider starting an open source project instead.