John C. Dvorak comes full circle in his column about IBM possibly dumping its PC business. He starts off saying it makes little sense, but by the end of the editorial, he has his mind made up that IBM should have done it years ago.

Of course, this was probably written before word got out that its talks are more of a joint venture or spinoff than a complete sellout.

Here’s the logic: IBM makes about $100 million a year on its PC business. To IBM, that’s not a lot of money, especially since it sells about $10 billion worth of PCs every year.

IBM wants to be a consulting company more than anything. I heard it said when the PC division was losing $1 billion a year a few years back that the PCs helped them sell more than a billion in consulting, so it made sense to stay with it. Maybe IBM no longer feels that way.

IBM’s desktop machines are uninspiring, but business PCs are supposed to be uninspiring. The original IBM PC was designed to blend in with file cabinets and other office furniture. It was underpowered, but it had one thing going for it: the ability to take more memory than anything else on the market at the time, which made huge spreadsheets possible and made the machine a success.

IBM tried to make its PCs inspiring in 1987 by giving them 256 colors, but most of the M68K-based computers on the market at the time had better graphics capabilities and they all had much better sound. The IBM-compatible PC didn’t become inspiring until the early 1990s when SuperVGA graphics and the Ad Lib and Sound Blaster sound cards became available and software started making use of them. None of this had anything to do with IBM, but it didn’t have anything to do with Compaq or Microsoft or Intel either.

IBM’s PS/1 and Aptiva lines of home computers sold reasonably well, but when the dawn of the $999 PC came around, IBM got out. Compaq and Packard Bell were outselling it, and then Emachines upped the ante with a $399 PC. A company that sells 7-figure mainframes isn’t going to be very interested in trying to compete with a $399 desktop. Besides, IBM tried before, in 1984, to compete with $400 PCs made by Commodore by responding with its $699 IBM PCjr. Commodore sold about 30 million computers while the PCjr took its place in history next to the Ford Edsel, New Coke and clear Pepsi.

So what’s one to do?

IBM’s PC business is going to be like a business that sells filing cabinets. Every office needs at least one, but it doesn’t matter much who made it or whose name is on it. A certain brand name on your filing cabinets doesn’t make people think any more or any less of your business, and, with very few exceptions, the same is true of computers.

By the same token, IBM’s ThinkPad is the best laptop on the market. It’s been the best laptop on the market, almost without exception, since 1993. During some points in history it’s been not only the best, it’s been the only laptop worth buying. So why doesn’t everyone have one? I don’t know. You can get a ThinkPad for as little as $899. It’s hard to find a cheap and nasty laptop that’ll fail the day after its warranty expires for more than $200 less. So for 22 percent more, you can get something worth having.

I suspect that if people knew that, and if they could go into a store near them and see one, they’d sell a lot more of them.

I don’t blame IBM for not wanting to sell home desktop PCs. HP and Dell are able to sell a lot more of them than IBM ever did, and Gateway/Emachines isn’t far behind IBM, but if any of them are actually making money doing it, it’s Dell. Dell’s PCs are as uninspiring as they come, but somehow their corny marketing works.

But that brings up a question. Dell is able to sell buttloads of PCs and make a lot of money doing it, even on razor-thin margins, so why can’t IBM? Sure, Dell has its sweetheart deals with Intel, but why can’t IBM negotiate a sweetheart deal with AMD? IBM’s manufacturing facilities have been AMD’s Plan B or Plan C in case it needs more production capacity for years and years. Couldn’t IBM manufacture AMD chips under license for its own use and sell PCs containing them and get better margins than Dell? They’d need to go get a couple of ex-Calvin Klein models to advertise them and give them sex appeal, but they can do it. People still buy $10 pairs of Tommy Hilfinger socks even though they can get a huge bag of no-name socks for $5 at Kmart. If cotton socks can have sex appeal through marketing, why can’t a sturdy laptop, or, for that matter, a desktop PC?

Maybe this is what IBM is thinking. They know sturdy and attractive computers ought to sell but they just don’t know how to do it. So they seek out a partnership with a company that can make them cheap, give them economies of scale, make it someone else’s responsibility to sell it, and just lend their name to the effort. Maybe IBM figures they can cut $10 billion in expenses from their bottom line and reap more than $100 million in revenue from licensing their name.

They’ve tried goofier ideas.