Zeigen

“April 1st, 2009. For immediate release. Rackable Systems Announces Agreement to Acquire Silicon Graphics Inc.”

I saw this news release cross the wires yesterday, but thought it was a prank. Rackable Systems? Who? And only $25 million? Hah hah, very funny. Try harder to make it believable next year, okay?

Except it wasn’t a joke. When I opened this morning’s Merc, the front page of the business section had a big headline: “Fallen star SGI to sell most assets for $25M.” Engadget also covered the story today. So it’s true. SGI will not emerge from bankruptcy and is finally no more.

Its products will continue in some form, of course, and I’ve heard from a friend of mine who’s still there that “the early word is they want to keep most of us on.” But SGI is no longer an independent company.

I worked at SGI from 1997 to 1999, my first Silicon Valley job. (I’d worked as a corporate trainer in Oakland and as a freelancer on web sites and computer books prior to that.) It was my first exposure to the culture and work style of corporate high tech: casual start and stop times combined with long hours and passionate, smart employees; vendor food, beer busts, high end cafeterias, over-the-top holiday parties, and free soda; casual attire; dogs at the workplace; back-to-back meetings filling the entire day; 500 to 1,000 e-mails daily; the uneasy alliances and divisions between engineering groups and marketing groups and operational groups and customer support groups and IT groups and HR groups.

I went in already familiar with Unix and the web, and with broad computer and database skills, but I didn’t know anything about how large companies work, or what was involved with program management and product release cycles. The politics of big companies were also new to me. I learned a huge amount, from BOMs and ECOs and the importance of quarter end dates, to the different ways you need to communicate to, say, a VP of manufacturing vs. a director of marketing vs. a principal engineer. If I had to list the top skill I emerged with, I’d have to say it was knowing when to communicate with e-mail, when to pick up the phone and talk it out, when to walk over in person (not always easy with 40 buildings in Mountain View plus employees around the world), or when you needed to leave a note on a chair, sometimes with a plate of cookies. Or when to call from your director’s cube.

I witnessed first-hand a beaten company falling apart. In 1997 SGI was facing its first set of money-losing quarters, and I barely made it in the door before a hiring freeze. This was right around the time when long-tme CEO Ed McCracken resigned. I managed to survive several rounds of layoffs even as the new guy.

SGI was reeling from culture clashes between the Minnesota employees who came on board when Cray was acquired in 1996. On the high end, Sun and IBM and HP were brutal competition for SGI’s server and supercomputer business. (Especially Sun.) On the low end, Mac and Windows workstations were starting to catch up with SGI’s high-powered Irix-based graphics workstations made famous in movies like Jurassic Park.

Things bounced back a little in 1998 after Rick Belluzzo came on board. Rocket Rick talked a good game, the stock rebounded a bit, and new products shipped. He had two major strategies. The first, a Windows NT workstation, was a miserable failure. It took forever to develop, and when it finally shipped, it was far more expensive than anything else on the market, and while it did perform better, it was riddled with compatibility problems. Point releases from Microsoft for NT weren’t able to be installed until SGI could come up with their own patches months later. Sales of the Visual Workstation were awful, and competing Dell workstations caught up in performance before too long. His other strategy, to shift focus to servers, seemed pretty smart to me, since most of SGI’s business came from the government. But most of SGI’s reputation and PR came from Hollywood, and distancing itself from its Maya line of software and the special effects business was probably the wrong move. Belluzzo left in August 1999, Microsoft-bound. The company contracted rapidly after that.

By 1999, the dot-com boom was in full explosion. Gravity didn’t exist, and money grew on trees. Insanity ruled all. At SGI, the stock price was going down, the products weren’t selling, and every single week came announcements of major departures at all levels from all around the company. I ended up with three different bosses in three months. When the opportunity came along to move to TiVo, I resisted at first, because I was moving up the ladder at SGI pretty rapidly. But the writing was on the wall and none of my SGI co-workers were enthused about SGI’s prospects. Despite all that, SGI had thousands of employees and billions in assets when I left in 1999. So it’s not a surprise to me that they hung on for 10 years, despite losing money year after year. Many of the governmental and server contracts were multi-year, multi-million dollar deals.

SGI, I salute you. So long. Thanks for the experience and friendships. You made some amazing products and hired some absolutely brilliant geniuses. If only there had been a way out.

Take a look at the Twitter account creation page:

https://twitter.com/signup

Two big problems:

1. The password field doesn’t show you the password you enter, and also doesn’t require you to confirm.
2. The e-mail address field doesn’t verify you actually own the e-mail address you enter.

This evening, I went to create an account for my son Sammy. He and I have been talking about computers, the web, social networking (and sure he’s only three, but some of his insights are just as worthy of sharing as the junk older people tweet). He agreed he wanted to share updates on Twitter, so we worked together to create an account.

Unfortunately, the password I thought I entered wasn’t what I actually entered. And even more unfortunately, the e-mail address I entered for him was not his actual e-mail address. (In my defense, he doesn’t use it yet, and I created it three years ago.)

Even though we were still logged in, I couldn’t change the password (since it required I enter the old one to change it), I couldn’t reset the password (since I didn’t have access to the e-mail address where the new password was set), and I couldn’t change the e-mail address (since that requires entering the password). That meant I was stuck, and the only recourse was to delete the account and start over.

The harm of starting over is:

1. We lost the updates he and I had typed.
2. We spammed all the people he had followed, wasting their time.
3. We spammed the actual owner of the e-mail address I entered. (I sent an apology.)
4. We lost the username since the account is retained for six months, so had to pick an inferior username.

I sent a help request to twitter to purge the old account so I can reclaim the username from his new account; their help pages say that’s an option. They also say requests take 5-7 business days. We’ll see.

We also saw a few dozen fail whale pages (especially when trying to set the profile picture to the image he picked), so trying this on a Friday night probably wasn’t a good idea. Sammy did love the whale picture at least.

It’s not uncommon for people to fat finger both an unconfirmed password and an unconfirmed e-mail address. Given the design flaws in account creation, I would estimate that tens of thousands of twitter accounts are created each month only to be abandoned later when the user realizes they don’t remember their password and don’t have access to the e-mail address. A responsible account creation design has you confirm an e-mail address.

I recently experienced the other side of this with eBay. Someone named “Edna Stephens” created an eBay account. The e-mail address Edna entered was not hers, but mine. I started getting eBay spam. (I don’t use eBay.) I couldn’t unsubscribe, because the e-mail they sent me had an unsubscribe link that led to a page that required me to log in to change e-mail preferences. I couldn’t send customer support an e-mail because they require you to log in to send a customer support request. I couldn’t log in as Edna and change her e-mail address to something other than mine (or delete her account) because I didn’t know her password. I couldn’t reset her password because the password reset process required knowing biographical stuff like her favorite musician. I was stuck getting spam. Finally I created a new eBay account for myself for the sole purpose of sending a complaint e-mail to eBay. It led to an absurd chat session with several different eBay customer support reps who required me to verify my street address and cell phone number and jump through several other hoops before saying “they’d investigate.” But I haven’t received any eBay spam so they apparently did reset her e-mail address.

The simple fix for both issues is that there needs to be a step where you verify your e-mail address really is yours. The typical process is that after you enter your e-mail address, the site sends you an e-mail with a unique code, and then you enter that code back on the account creation page to continue.

eBay and Twitter don’t do that. They should.

I know why companies don’t always follow that step: Prospects will often not bother waiting for an e-mail and jumping through that hoop, and it also creates problems since a surprising number of people do not know how to copy and paste your code. Those prospects balk at that step and simply don’t join the site — and the prospect is lost.

(From a web development and QA perspective, it’s also more work to code in that verification step.)

It seems most companies would rather get the prospect and deal with incorrect e-mail addresses later. But I think the customer support burden and spam behavior really requires e-mail address verification during account creation.

The company I work for doesn’t require e-mail verification either — but then again, we create a tiny fraction of the accounts each day that are created for twitter and eBay.

Moral: Twitter and eBay should both reconsider their policies of account creation.

By the way, if you are so inclined, you can follow Sammy on twitter as e_sammy. I’ll be changing that back to esammy if I can.

I believe Microsoft Outlook is the most incompetent software in common use in the business world. Here’s reason #4,334 why I hate it:

• Suppose you’re in offline mode because it randomly disconnected you and is stuck in a login-loop when you try to reconnect.
• You forget you’re offline and you go to cancel a meeting by deleting it. It now wants you to send the cancellation. You click Send.
• You get an error message that you’re not connected. It asks if you want to try to send the cancellation again.
• Now you’re trapped. If you click Yes, you get the same error message. If you click No, you’ve just canceled the meeting on your calendar, but not on anyone else’s. There’s no recovering from that mistake. If you click Cancel, you’re back to the message that you need to send, but can’t be sent.
• Even reconnecting (assuming you can reconnect) doesn’t solve the problem. Now pick up the phone and tell your  meeting attendees they have to manually delete their copy of the meeting from their calendars.

So much hate.

Up in Tahoe on Friday, Sophie needed to nap while Sammy and Kimi were ready to start sledding, so I read in the front seat of our Forester while she slept in her car seat. After her 45 minute nap, we joined in on sledding. But when we were all done and I went to start the car: Nothing. Dead battery.

Fortunately, a kind man parked right next to us gave us a boost. Instead of the traditional jumper cables, however, he had a self-contained unit. He just hooked it up to our car battery, pressed a button, and a few seconds later I could start the car. (I think it was a Duracell DJUMP.)

So why the dead battery? Maybe I left a light on while I was reading. But on the other hand, the Forester is six years old and still on the original alternator and battery, so I suspected one or the other might have gone bad. Better to be cautious in snow country, so I had them checked out; turned out everything was fine. We had no further trouble with the car up in Tahoe.

The kind gentlemen had mentioned that his self-contained starter was only $20 at Radio Shack. Figuring this was a handy gadget to own, I stopped by a Radio Shack to pick one up, but they’d never heard of it. No luck at an auto parts store or Rite-Aid up in Tahoe either.

This morning Kimi tried to start the car, and again: dead battery. But this time I had certainly left a light on — after driving home last night and unpacking, I had been careless.

None of our neighbors were home, and we don’t have a set of jumper cables (whoops), and we didn’t have time to wait for AAA, so we took my car. I dropped off Kimi at work and the kids at school, and stopped by Target intending to buy some jumper cables.

There I spied the Black & Decker Simple Start 12-volt Battery Booster. This is a car starter with a twist: It doesn’t use jumper cables connected to your car battery. Instead, it charges the car battery through the 12-volt DC adapter (you know, that thing that used to just be a cigarette lighter, but now seems to be the universal source of power for anything, including pizza ovens).

The Simple Start also doubles as a flashlight and a cell phone charger. Smaller in size and cheaper than the competing battery starters that connect to the car battery in the more usual way, this seemed to me like a product that would be useful for my wife. After all, if she were ever stranded with a dead battery in the rain, she wouldn’t even need to get out of the car. Its compact and light, and designed to be stored in the glove compartment.

Now even if this product worked (note foreshadowing), it has two downsides, as disclosed in the manual. (Neither of these limitations are disclosed in the text on the external packaging, naturally.)

1. It has to be charged for 15 hours before first use.
2. Once you plug it into your cigarette lighter and switch it on, it takes 15 minutes to charge the car battery to the point where you can switch on your engine.

I don’t know about you, but when I find out I have a dead battery, it’s because I just started the car, which means I need to be someplace, so I’m not really in the mood to wait 15 minutes (let alone 15 hours).

A bit deterred, I pressed on and plugged the thing in and started to charge it at work this morning. Strangely, it seemed to already be charged, since the green light was solid instead of flashing. But just in case, I let it charge for over 9 hours. The green charged light never changed state. The flashlight seemed bright. It looked to be ready to use.

So, this evening, following the instructions, I set it to charge the Forester’s battery. Everything seemed to be working fine, so I left it for the prescribed 15 minutes. I returned to find it had switched itself off. Full of hope, I then tried to start the car.

Nothing.

Hmm.

I took it inside, plugged it in, and made sure it was still fully charged. Reassured by the solid green light, I took it back outside, and tried again. Once again, it flashed in the way that indicated it was hard at work charging the car battery. I waited the 15 minutes once again.

The moment of truth? Nothing.

I tried using a different adapter in the Subaru, since the manual said that some adapters may not have the right connection.

Now the thing just beeped at me, indicating that it thought the Forester’s battery was charged. But it was wrong. Dead battery wrong.

I called the customer support number listed on the manual. They were only open until 5 pm Eastern (apparently not caring about West Coast customers’ needs at all), so that was no help.

I tried once more, using both different possible adapters, and again the thing just beeped crazily at me.

That was four attempts with no success. FAIL.

I knocked on my neighbor’s door and asked him if I could borrow his jumper cables. They worked just fine, and our car started instantly.

In researching the Simple Start on Amazon, I find now that it has gathered 87 one-star reviews out of 189 total reviews. So it’s not just me. For most customers, this thing simply does not work.

It seems to me that for a product such as this, reliability is the most important feature. If you need a jump start, you really need it. I wonder, then, how Messrs. Decker and Black could, in all conscience, release this product and keep selling it for more than a year, when some 46% of customers find that it does not actually do its job at all.

We will be returning this junk and acquiring a set of jumper cables instead. (I will probably give the Duracell product a shot as well, since I saw it work first-hand.)

And now there are 88 one-star reviews at Amazon.