The Demise of TPI

In this interview, Jon Orwant, publisher of The Perl Journal (, discusses the fate of The Perl Institute. As one of the founders of TPI, Jon was there at the beginning and he was there at the end when the board voted unanimously on March 1, 1999 to dissolve TPI and transfer its assets to the Perl Mongers organization. The original idea behind TPI was that companies using Perl would support a non-profit to do a variety of good things for the Perl community. Even though it seemed like a good idea, TPI struggled. Two years ago, TPI was reorganized but apparently with no better results. Why was it so hard to make TPI work?

Jon Orwant: Ah, the killer question. Batten the hatches. I will be blunt.

When TPI started back in '96, we four (Larry Wall, Tom Christiansen, Randal Schwartz, and myself) wanted to create a non-profit organization that would help advance the interests of Perl. In Larry's words, "to help people help Perl help people." We each had slightly different ideas about what we wanted done, but we were all basically on the same page: to help legimitize Perl in the eyes of outsiders, to coordinate volunteer development, to provide a web site and mailing lists, and to give people who have benefitted from Perl a place to write a check if they wanted.

Typically, non-profits work like this: they have a board of directors, a secretary, and an executive director who actually gets stuff done, taking the board's suggestions and implementing them. The board was the four of us, the secretary was Sharon Hopkins (a longtime friend of Larry's), and the executive director was the friend of one of us.

That executive director was a complete nutjob. He knew nothing about Perl, or programming, or computers, or business, and spoke with bombast about how he'd build TPI into a million-dollar organization. Most of the board varied from skeptical to outraged. He came up with obviously unworkable get-rich-quick schemes couched in new age rhetoric.

From the beginning, Larry asked for "open accounting" -- a down-to-the-penny description of how we were spending members' money. We asked for that over and over again, and when we finally got a financial statement from the executive director, we learned why he was so reluctant to give us the numbers: he paid himself a lavish salary ($5500/month) out of our coffers. Maybe in his bizarro-world, that didn't seem like embezzling; after all, spending a hundred grand to make millions is a good deal. Your lucky Lotto numbers are 17, 28, 46, and 92.

When we found out about this, we clamped down on expenses. Or at least, we tried to. He held the pursestrings, so it was difficult. Eventually, we fired him.

For "TPI 2.0", we made two changes. Chip Salzenberg volunteered to be the new executive director, and we added Tim O'Reilly to the board. As any Perl guru knows, Chip is everything that our previous executive director wasn't: he's honest, smart, hard-working, realistic, and he knows Perl better than anyone.

Yet TPI 2.0 failed too.

To understand why, you need to put yourself in the place of a board member. You're busy. You do a lot of things apart from TPI. But every once in a while, you have a good idea about something TPI should be doing. So you send mail to the board.

No one answers, so you send mail again and yell, "Hey, is anyone listening?" One person replies, saying something like, "Well, that's a good idea, but X is already doing something similar, and we should coordinate." He doesn't say *who* should talk to X, because board members shouldn't be barking orders at one another.

Not all replies were like that. Some were more like:

"Hey, we're a corporate board. We can't make our decisions over email. According to the bylaws, we can only make decisions at an actual physical board meeting, where we use Robert's Rules of Order, put on powdered wigs, and refer to each other as the Right Honorable Gentleman From Mountain View." (Okay, I made up that last part.)

"But we're too busy to meet in person. Why can't we just try a bit harder and make decisions over email?"

"The corporation by-laws don't permit it."

"Then let's change them. I hereby propose to amend the by-laws to allow email votes without in-person board meetings. All in favor?"

"Sorry, the by-laws prohibit you from calling an email vote without an in-person board meeting."

Pathetic, no? We had no leadership. Eventually we did manage to have email votes, but even then we still did nothing. There was no one to crack a whip and tell people to do things. Every once in a while one of the board members would try cracking a whip, but it never worked, because other members wouldn't recognize his right to crack the whip. Meanwhile, the mounting embarrassment at TPI's impotence made each of us unwilling to take any sort of public action, for fear of being associated with a such moribund organization.

And so we each went back to projects where we could actually make the world a better place.

At the last Perl conference, Nathan Torkington took a few matters into his own hands, and single-handedly did more in the following month than TPI had done in the previous two years.

The lessons? Know when to insist on bureaucracy and when to chill. Choose leaders who will lead. And stay away from embezzling nutjobs. So, the assets of TPI are moving to the Perl Mongers. What does this mean for the Perl community?

Jon Orwant: "Assets" makes it sound like we're talking about a big portfolio. We're not: the assets are $3000, some computer equipment, and the domain. Maybe and a fax machine as well; details are hazy.

It's my hope and expectation that the Perl Mongers will be able to make better use of the TPI assets than TPI did. Ideally, they'll do some of the things that TPI never could. I hope so, but that's not a precondition of our gift.

In short, Perl needs an administrative center. For all you folks about to mail me The Cathedral And The Bazaar, please don't. I'm not talking about centralizing Perl technical development. I'm talking about a robust process for creating things like mailing lists, documentation support, and answering questions about Perl: not just how it works, but where it's going. When someone needs a statement on letterhead confirming that Perl is Y2K compliant, or demographic information on the Perl community, where can they go? We need one-stop shopping for this sort of boring administrivia.

Perl was built on the backs of volunteers, and we've been blessed with hundreds of people who have offered their time, money, expertise, and computational resources. But TPI was so feeble that it couldn't even say "Thanks for your offer, here's what to do" to the volunteers. An organization so lame needs to be put down, and that's exactly what we've done.

Visit the home of the Perl programming language:

Sponsored by

Monthly Archives

Powered by Movable Type 5.13-en