Now What?
Netscape survives its own autopsy

April 2, 1999


See, it choked on its own hairball

Let us now speak freely about Netscape.

Jamie Zawinski has finally given us permission. He posted a long resignation and postmortem on his own site today, during the hangover hours after Mozilla.org's first birthday party. The timing was kind. In an email sent to friends, he wrote "people who work for companies that didn't just sack 400 employees might actually be looking this party as a birthday instead of a funeral, and I'd like to let them keep their illusions at least for one more night."

But now the illusions are gone and the funeral is just starting. Jamie has seen to that by pulling the cover off the corpse and leading us through the autopsy.

Here are the high points from his pathology report:

  1. Innovation stopped early — like, right after 1995.
  2. Microsoft won when they made their browser free and "destroyed the market."
  3. Layoffs were demoralizing. Most of the good people are long gone.
  4. Engineering went to hell. "Netscape was shipping garbage, and shipping it late."
  5. The open source effort was a train wreck from the start:

If nothing else (and there is plenty else), Netscape tested the good will of the marketplace. And what it measured is still so enormous it verges on the unbelievable.

Consider this from Eric Kidd: "They released an unusable product. The Netscape 5 code that was released to Mozilla.org didn't compile. Once people got it to compile, it didn't run. Once it ran, it died on any major website. Now, open source developers will forgive bugs. But they rarely contribute to products they can't use. So developers stayed away in droves for the first six months." And didn't talk about it.

The ubiquity of this forgiveness almost amounts to a conspiracy. Why else would such a huge downside get no publicity? Think about it: Netscape's open source story was a box of parts on a high horse. But the people in a position to call this emperor naked said nothing. (Or, in my own case, gave it the benefit of zero doubt.) It was a kindness the company hardly deserved.

Like Jamie, I held my tongue for a long time about Netscape. Yes, I had family working there (still do); but the main reason I kept silent was that, like millions of other fans, I wanted them to succeed. They were my team, and I was as loyal as a dog.

But as a customer, I wasn't treated much better. Unlike most users, I paid for that damn browser. I must have called the company about ten times for various kinds of tech support (emails were almost pointless). What "help" I got was mostly useless and entirely disinterested. One support person told me that the browser "just didn't matter" to the company. "We're really in the server business," she said. "So why are you charging for the browser?" I asked. "Because people are willing to pay for it."

So maybe Jim Barksdale wasn't kidding a year ago when he said, "We successfully drove our browser revenues down to zero." They finally lost the sucker market.

Who's sucking now?

So is AOL the biggest sucker of all?

Hardly. Because Netscape isn't dead. Sure, the browser business is toast, but that's hardly the end of the story. Its code is still open to the world, and folks are still hacking away on it. In fact, the same Eric Kidd I quoted above suggests that perhaps Jamie left too soon, and adds, "I'm willing to give Mozilla a bit longer — they've followed a long, winding path, but they're finally starting to have a product."

In business, companies are not dead until their assets get auctioned or they disappear into the belly of another beast. If Netscape is dead, why did it sell for $10.2 billion to a company that's not only the world's biggest ISP, but Microsoft's biggest competitor — and the only company Microsoft has failed to beat at one single thing? Did AOL just buy the Netscape "brand?" Was this all just for Netscape's "portal" (the derivative and annoying "Netcenter?)"

What AOL bought was Netscape's mission — and the man who started it all.

For better or worse (and it has been both), Netscape has always been Marc Andreessen's company. Without Marc, Netscape would never have happened in the first place, and it never would have sold to AOL. Marc was as valueable to Netscape as Steve Jobs was to Apple. Without those two founders, neither company would be worth much more than scrap.

When saw I Marc last month he was chatting with a table full of Microsoft folks, look more fit and relaxed than he has seemed in years. When I asked about his new job, he said "AOL is an amazing company. They are going to do some fantastic things. Just watch." (Making the moment even more surreal was my mission at that table, which was to give one of the Microsoft guys a lift to a local Linux Users Group meeting where Eric Raymond was lecturing on Open Source.)

Of course it's easy to dismiss Marc's optimism as a result of equity-coupled wealth induction. But he has three good reasons to be optimistic about his old company's opportunities:

  1. The most important action has always been on the server side of the Net, not the client side; because clients only derive benefits that services provide. Yet the net effect of both the "browser war" and the world's constant fixation on Microsoft's "domination" has been to blur and mask both the importance of services (not just servers) and Netscape's well-earned leadership position there.
  2. The Net is still young, and it won't begin to fulfill a fraction of its long-term promise until it acquires real infrastructure in the form of ubiquitous network services at least as sturdy and useful as the ones that were old hat with Novell's NetWare and Lotus' Notes ten years ago. Those services are messaging, directory, management, security, file, print and other mundane nouns rarely heard in conversations about the Net because it's hardly worth talking about them yet.
  3. Microsoft is vulnerable in this whole network services space, because it needs to drag along its aging Win32 development platform, which was designed originally for clients, not servers.

Netscape has long since taken the lead in providing all kinds of infrastructural services. It has stayed ahead of Microsoft by adopting open and extant industry protocols (such as LDAP for directory access), saving development time and opening markets for participation by everybody. (For just one example of this strategy at work, read "A Bulldozer Through the Intersection." at the Linux Journal site.)

Take a look at Netscape's Servers page. Whatever the virtues or failings of the products here, they are exactly the kind of stuff we need to build real infrastructure. If you want to see where Microsoft really lags, check here. Then check this article in the current BusinessWeek and this article in the current Sunworld. Both are about the new Netscape Enterprise Group, which consists of 1000 employees each from Netscape and Sun, gathered in a "virtual company" to build e-commerce infrastructure based largely on Netscape's server products.

This is serious stuff. There should be a Linux play in here as well. But are the opportunities seem to be getting little more attention from the Linux vendors than they do from the mainstream press. Both seem to see the world through browser-framed glasses.

Take Bob Young's chapter in the new book Open Sources. It's full of examples from the client world of PCs, consumer products and users, and precious little from the server world where his products have the most real importance. He writes, "The benefit an open-source OS offers over the proprietary binary-only OSes is the control users gain over the technology they are using."

Users? Control? I'm a user with Linux, Mac and PC clients on my desk . The one over which I have the least control is the Linux box. Linux clients are packed with technology that is abundantly manifest to the hacker and opaque to the ordinary user — and even to less-than-ordinary users such as me. The box that gives me the most control is my Mac, which is also the most proprietary and closed of my three OS choices. (In fact I'm writing this right now on the Mac, because Adobe's fabulous GoLive 4.0 so far works only on the Mac platform. Go figure.)

For everyday work, my Linux client is 99% promise and 1% delivery. I expect the opposite ratio from my first Linux server: a Cobalt Qube2 that's due any day. I don't want to pre-review something I don't even have yet; but I've played a bit with this thing and it appears to be exactly what Cobalt claims: a simple Web appliance. You want a Web server? Here it is: a blue box that's hardly bigger than a phone. Plug it in, hook it up and configure it with a browser on any client you want. Now get this: you have to look at the small print on the back side of the Qube2 datasheet before you see that the thing runs Apache on Linux. Double irony: a Cobalt guy told me the Qube uses Red Hat Linux. No wonder Bob Young doesn't feel like talking about server stuff. There's no brand leverage. If you want the press, work the client story. (For what it's worth, the Qube also has a 64 bit processor of some kind. Who knew? Who even needs to know?)

But the Qube isn't the only server Cobalt sells. ISPs (including my own), have racks full of Cobalt's RaQ and RaQ2 servers — all running Apache on Linux. While we're busy debating Gnome vs. KDE and StarOffice vs. Applix and Corel — all of which are far more promising than practical — Linux kicks the most butt where it gets the least attention: in the server space.

Make mine Manhattan

Dan Bricklin recently addressed the infrastructure issue when he said, "It's a lot easier to build Levittown than Manhattan." His point was that open source software is still unproven at the Big Stuff that makes civilization run. He's right.

And he also gives us a terrific metaphor. We build with code. The open source movement is an argument for better building materials, better tools, better buildings, better architechture — and better conversations about all of them.

We need to start framing those conversations in bigger terms. And maybe we can, now that Jamie has conducted his autopsy. Because what Jamie talks about is not just the death of Netscape's original browser dreams, but the losing side of the "browser war" that began diverting media attention when Bill Gates declared it on December 7, 1995 (see Microsoft + Netscape, which I wrote four days after that event).

Web service is the first important chapter in the Linux story. Many chapters remain, and most of them will be about services, not browsers or the desktops that showcase them.

Let's start hearing those stories.