I've been on a roll lately, writing columns that seem to resonate with a lot of readers. It's nice to have people agree with me. But it also happens that columns like this one develop a kind of sludge that needs from time to time to be cleaned out. Well it is sludge-mucking time, when we revisit a few ideas presented in earlier columns and see just how well they are holding up.

First there's the iMac, which I predicted would have both shortages and reliability problems, neither of which have become apparent to the public. Right now the iMac is the number one selling personal computer model in the U.S. and easily on-track to meet Apple's projection of 400,000 sales by the end of the calendar year. There seem to be no shortages and nobody is telling me that the machines are breaking in the field. On the face of it, then, I was wrong. The whole story is a little more complex, but even then I was still wrong. Apple has done an excellent job of bringing the iMac to volume production.

Part of this iMac story includes Apple's early decision to test every machine before it left the factory. Normally this isn't done, but the relatively high failure rate among pre-production units - the very numbers I cited in my earlier column - prompted Apple to make the testing decision. And it paid off. Since every machine was tested before shipment, the out-of-box failures happened before the machines ever saw a box. But individual testing is an expensive thing to do with a $1299 computer. I understand now the failure rates are within acceptable limits and the testing has ended.

So here's to the iMac, which isn't in itself going to return Apple to greatness, but is nevertheless a great success story. The question now is what will Apple do next?

Next on the mea culpa, mega culpa list is last week's column about Microsoft's dropping support for Samba. As several readers pointed out, the files about Samba I said had disappeared from the Microsoft Knowledge Base have suddenly reappeared. Those who are into conspiracy theories have suggested that this was in quick response to my column. I don't think so. Microsoft sees me as an annoying insect, and it takes a lot of buzzing on my part before they generally react.

It is much more likely that I couldn't find the files not because they weren't there but because the Knowledge Base wasn't working. Now isn't this interesting? Microsoft's repository for support information fails and the failure mode tells users there is no information available, not that the system is down. Normally this would be viewed in the industry as bad design, but the Knowledge Base is out of action so often that Microsoft has apparently decided it is less embarrassing to fool users in this way than to return with the news that the darned thing is broken.

This doesn't however, change the fact that Microsoft still has documents available from time to time that tell how to work with Samba. And Samba, too, has been at work on the problem of how to deal with Windows encrypted passwords. The latest versions of Samba have a workaround for this, though readers tell me it isn't particularly easy to use.

Not that I should talk. If you've tried using the Baloney moderated forum on this site you know something is amiss in the world of I, Cringely. Here is what's wrong, and what we are doing about it. There is a bug in the code that truncates posts when it comes across certain punctuation marks, so lots of very interesting thoughts have been stopped in mid-sentence. We are working on the problem and hope to have it resolved shortly. We've had, too, a number of problems with the administration server. I can assure you, though, that the Baloney section is about to get much faster and more useful. I promise.

In recent days a number of interesting things have happened to me, personally. I was called by a lawyer from Utah who was looking for an expert witness to testify in an antitrust case. This had to be Caldera versus Microsoft, and since the lawyer was calling from Salt Lake, I expected it was Caldera looking for assistance. Besides, I have interviewed those people and we got along quite well. Nope, it was Microsoft.

Understand that I don't spend a lot of my time being an expert witness. In fact, I have only been an expert witness once before, during the divorce trial of Broderbund founder Doug Carlston. As an expert in both the software industry and divorce, that one was a natural for me. Antitrust is not quite as good a fit.

What Microsoft's lawyers were looking for was a well-established computer journalist who could be claimed to have broad expertise in the workings of the software industry YET HAD NEVER WRITTEN ANYTHING NEGATIVE ABOUT MICROSOFT. Obviously, these lawyers hadn't read my work before putting my name on the list.

In my phone conversation with Microsoft's lawyer I copped to the fact that just maybe his client might see me as having been in the past just a bit critical of their products and business practices. This was too bad, he said with a sigh, because they were having a very hard time finding a reporter who both knew the industry well enough to be called an expert and who hadn't written a negative article about Microsoft.

I imagine that poor lawyer, soldiering on at $300 per hour in a vain attempt to find a shill for Redmond. While this is only a single data point concerning the state of Caldera v. Microsoft, I get the sense that Redmond is sweating. Remember that Caldera inherited the lawsuit from Novell along with DR-DOS. Remember, too, that nearly all the money behind Caldera comes from Ray Noorda, who used to run Novell. Noorda has the resources and the determination to take this case to a successful conclusion. Microsoft should be sweating.

And I was sweating this week under the TV lights on a sound stage in Baltimore, finishing-up for PBS a program about digital TV. Like all the TV networks, PBS is under the gun from the FCC to begin broadcasting a digital signal by next month. As the guy who has to explain to this new technology and why it justifies purchasing a new and, right now, pretty darned expensive television, I have been learning a lot.

The part of digital TV I find most interesting is what PBS is calling "Enhanced TV." This is a peculiar cross between teletext and the World Wide Web. Excess bandwidth is used to encode additional information about the program being broadcast so users can access that data if they want to. Icons appear onscreen to indicate when extra info is available. Like teletext, this is a one-way system. You don't ask the PBS station to send you a link. Rather, the station sends all the links and the data is held in your TV or computer just in case you want to have a look. This buffer is periodically flushed as new data appears.

What makes Enhanced TV exciting is the amount of bandwidth available. Remember how teletext — and even Wavephore's current Wavetop product for encoding Web pages on TV signals — use the vertical blanking interval in the analog TV signal to carry the data, resulting in performance no better than a regular modem. Well, Enhanced TV has a much bigger pipe. Depending on what else is being broadcast, whether it is a High Definition TV (HDTV) show or up to four standard definition digital channels, there is a MINIMUM of two megabits-per-second available for extra data.

While I work with PBS and write this column for PBS Online, I have no inside knowledge of what that network, or any of the commercial networks, plans to do with this available bandwidth. But it seems only logical to me that they will put it to good use. Not all shows, you see, will have additional data. This very big data pipe that will over the next year or two begin broadcasting past every TV and PC in America will have a great deal of excess capacity. During some times of the day Ethernet speeds will be possible and from multiple networks at that. Whether people go out and buy digital TVs or not, I am absolutely sure there will be a class of PC networking products to take advantage of these new data services.

If Kenneth Starr is still issuing reports into the next century, we'll be ready.