The Micromarkets of the Internet

I suppose I should find the consequences of network effects so commonplace by now that they're hardly with a certain notice, much les the awe—a dawning realization of just how unlikely and improbably something would have been before Web 1.0 (or at least the democratization of Newsgroups). In this case, it's a sideline business so narrow and specialized that it's hard to see how it could ever be made to work if it required print ads (even in the most specialized of magazines) to find an audience.

First a little background. Northgate was one of what was probably the 2nd generation oro so of PC clones. Compatibility issues were in the past mostly. Northgates like Zeos (along with an equally fledging company called Dell and the still surviving though hardly as successful Gateway) sold most of their systems through ads in the like of PC Magazine and Computer Shopper and direct mail. However, Northgate also made a line of keyboards. Now, back then, a lot of people actually cared about keyboard feel. The IBM PC's with their "buckling spring" clicky feel were the gold standard. (Clones are still made.) But the IBM laylout and the sizes of some of their keys weren't to everyone's liking. Northgate responded to this rather finicky need by releasing several keyboards that, among other differences, had a huge ENTER key the way god intended. I have a couple of these, by now, elderly devices.

Unfortunately, the Northgate keyboard on my regular desktop computer started acting up. Besides being filthy, a couple of keys were sticking. Si I tried cleaning it up, and through the over-zealous application of isopropyl alohol (and then other things), I managed to make it go from bad to worse. There was always the ebay option—another form of network effect—but what should Google find for me but someone who actually repairs Northgate keyboards, Robert Tibbetts. In his words:

My second computer was a 1987 Northgate 386 with a Gold Label 102, in 1999 it started going south. Three switches decided not to work anymore. So my brainstorm was to get some Radio Shack contact cleaner, big mistake. Instead of 3 switches, I had about a dozen that did not work. I scrounged around and came up with a used Ultra T that worked OK. It was bought as in new condition, NOT, but it worked. The GL always bugged me, so I decided to take it apart and pull the switches apart. I repaired them and made it work again, now I just replace them. It is just to time consuming to take a switch apart and clean the contacts. So after that I bought some on e-Bay, fixed them and resold them and here we are 4 years later and over 600 refurbed and sold. I keep around 100 keyboards in stock, some are new.Fixing keyboards is just a sideline that I enjoy doing. My main work is wholesale lumber for the last 40 years.

So, if you've got an old Northgate keyboard sitting up in an attic gathering dust. Consider dusting it off and sending it in. The cost for cleaning is about $60 with shipping and all; key switches cost about $5 each to repair. Above all, keep that Radio Shack Contact Cleaner away from them!

Slashdot Going Out of Style?

Jeremy Zawodny asks ifSlashdot is going out of style? [via Eric Boutilier] Speaking for myself, that seems a reasonable projection. I still find it somewhat useful as a pointer to interesting stories, but the actual synopses are often (typically?) incomplete and horribly slanted in one way or another. Perhaps more seriously, the comments never represented well-reasoned commentary from a broad population, but I think they did once represent a somewhat useful window into some of the most fervent Open Source vanguard. Today, with Open Source mainstream and made up of so many communities and diverse interests, the Slashdot crowd is increasingly such a small slice that I'm not sure its zeitgeist tells us anything useful.

Blockbuster Sinking

I'm usually suspicious of claims that the "old" way of doing things is dead because the latest favorite trend or company or business model of the techno-elite is going to blow it away. And 95 percent of the people who don't live within the boundaries of the Valley or Boston or RTP have never heard of whatever the hot new thing is. And wouldn't want it if they did.

But sometimes things really do change.

John Paczkowski of Good Morning Silicon Valley notes:
Blockbuster's days are numbered. Hemorrhaging cash, the video rental chain that once claimed a market value of $8.4 billion is today worth just under $700 million... If only it had purchased Netflix when it had the chance. A paltry $50 million and it would have had owned the company that destroyed its single biggest profit-earner -- charging late fees to customers who kept videos past the due date -- and forced it to invest millions in an also-ran online rental business that is too little, too late.

A commenter to the post counters that "i disagree. i used to feel the same way but those of us in the valley are too close to our own technology. go to anywhere outside of a high tech center like our own and blockbuster still thrives." Perhaps. But there's little disputing that Blockbuster's financials are a grim thing indeed.

There may indeed still be a significant demand for bricks-and-mortar video rental—but likely not at the scale or at the price structure of current operations. And here's a very real question: If the corner video store's costs require a premium over online and video-on-demand rental charges, how much additional business will it lose?

Still, we'll have video rental stores for a long time—perhaps most of all in cities. But we may well not have Blockbuster.

RSS Data

via David Berlind

Related: 31 percent of US Internet users use RSS, and most of them do so through My Yahoo.

Also related but contradicting: Recent research conducted by Yahoo! and Ipsos reveals that while 12% of surveyed Yahoo users know what RSS is, only 4% of surveyed Internet users use it (PDF) (and know they use it).

Interesting data but I'm not sure that I see the contradiction. My Yahoo uses RSS as a foundation technology. A fair number of people use My Yahoo but don't know (and in many cases care) how Yahoo is grabbing the feeds they're reading. That aside, a few thoughts:
In a recent post of mine about "Web 2.0" hype and definition, James Governor commented:

i am afraid you're showing your age buddy. kids under the age of 20 have never heard of web 2.0

they use services, but aren't thinking in the abstract. its us old bores talking about it. i hate to say it, but nicholas carr probably called it right when he called us elitist-hippie-boy-scouts

I tend to agree, at least up to a point. This data is a perfect example. People use RSS but they don't think about it. To take another example, how many people using various rich web interfaces like Google Maps have any idea what AJAX is?

At the same time, though, if RSS is truly used that narrowly outside of My Yahoo, that surely argues that blogs and related forms of personal syndication are relevant to a mighty narrow sliver of the population. Without RSS readers (and presumably the knowledge that one is using RSS), anything more than the occasional look-see at a favorite blog or two, or catching the odd post from a Google search, is well nigh impossible.

Wikipedia Experiments

I've written about Wikipedia in the past, most recently here. But I've stayed away from the most recent flaps for a variety of reasons: they were getting so much ink and so many electrons elsewhere, general fatigue with the topic, and a general sense that (as in so many things related to Open Source) positions were often so calcified that scoring points trumped legitimate discussion. Perhaps, as Nicholas Carr suggested in a recent post:

Wikipedia ran into trouble because it assumed - or allowed itself, not unwillingly, to have thrust upon it - a mantle of "authority" that it neither needed nor deserved. It became a cause celebre of techno-romantics who saw it as a harbinger of an internet-enabled era of egalitarian media and universal creativity...

Wikipedia is not an authoritative encyclopedia, and it should stop trying to be one. It's a free-for-all, a rumble-tumble forum where interested people can get together in never-ending, circular conversations and debates about what things mean. Maybe those discussions will resolve themselves into something like the truth. Maybe they won't. Who cares? As soon as you strip away the need to be like an encyclopedia and to be judged like an encyclopedia - as soon as you stop posing as an encyclopedia - you get your freedom back. You lose the need for complicated rules and restrictions and all sorts of tortured hand-wringing and navel-gazing. You don't have to worry about critics because critics don't have anything to criticize. Some facts are wrong? Hey, we never claimed they wouldn't be. Someone created an entry about an imaginary being from Planet Xenat? So what were you expecting - an encyclopedia?

Be that as it may, I've been conducting an experiment of sorts over the past few months. I've been keeping watch on a few Wikipedia articles on which I've done some work and/or interest. Obviously this is anecdotal (and on a small scale at that), but it probably sheds some light all the same. The articles in question are AViiON, Data General, Cambridge (MA), and InfiniBand. Read through the recent histories to get the full thread, but I'm going to pick out a few things I've found indicative and give my overall take.

AViiON: As I detailed previously, there were considerable problems with this article. I did a fair bit of reworking as a result. The result is not as complete as it could be, but IMO it's accurate as far as it goes. Recent changes by others? Minor, but positive. Moving a secondary and unresolved topic (where the AViiON name came from) to the notes. Fixing some minor errors and inconsistencies. I had to make my own meal, but no ones ruined what I prepared and has even added a few spices here and there.

Data General: Not so good. Breathless campaigns against supposedly non-neutral points of view here and only slightly less so here over what seem to me pretty factual statements. But, hey, we have to observe Neutral Point of View (NPOV) even though the concept is pretty silly taken to the extreme. Any conclusion is a point of view. It's only a question of how strong the supporting evidence is and how many people agree with you. No fixup or improvement of most of the problems I pointed out in my earlier post.

Cambridge (MA): Twiddles. No real harm though mostly pretty irrelevant. Additions to people who had once set foot in Cambridge, PC-ish fiddling with how to describe recent immigrants, one quickly-reverted piece of minor vandalism. The article could certainly stand some more history, but it's not seriously deficient for what it is and the changes haven't really been negative.

InfiniBand: I'm uncertain about this sentence that's been added at the end. (As far as I know, Horus is a 32-way chipset for AMD Opteron and can find no online support for what's described here.) That, of course, is one issue with the Wikipedia editing process. Although there are potential forums for discussions (e.g. Talk pages), it's not like a normal collaborative editing process really. As a result, you may respect the original material more than I would in my normal role as editor.

Overall take (for these statistically insignificant examples)?: For the most part-—piddle, twiddle, and resolve. Not one damned thing do we solve. Copyediting's great, but systematic editing, researching, and good writing are better.The "info rot" in these articles was minor. But neither were there step function improvements. Not an F. But a gentleman's C on these topics.

More on Bestselling CD's

In this earlier post, I noted that Chris Anderson had some interesting data which seemed to indicate the decline in CD bestsellers was even more dire than the overall decline. Chris is back with some additional data to reinforce this observation. I won't rehash it here for now because it really needs some normalizing and such to be really definitive. However, because I raised the question in my post, I did want to note that RIAA data on bestselling albums does seem to put one of my questions to rest:

One of the benefits of this multiple award process [a Platinum album is also a Gold album and so forth] is that it allows us to look at a question that commentors raised in the last post: do older albums simply have more time to become a hit, something that would bias this kind of analysis against newer albums? The answer appears to be no. If you go through the database you'll see that albums that achieved Platinum or Multi-Platinum designation almost all did so within a year of going Gold. So most of the sales do seem to be in the first year or two after release.

(In other words, the fact that there have been fewer bestsellers in the past couple of years doesn't seem to be because they haven't had time to "peak" yet.

Massport Continues to Fight For a WiFi Monopoly

As reported in yesterday's Boston Gobe,

Logan International Airport officials' [Logan is operated by Massport] ongoing quest to ban airline lounges from offering passengers free WiFi Internet services is angering a growing array of powerful Capitol Hill lobbying groups, who say Logan could set a dangerous nationwide precedent for squelching wireless services.

At least they've dropped the argument that blocking competing free WiFi would help keep terrorists (presumably cheap or poor terrorists who couldn't afford Massport's $7.95 daily charge) from communicating. They now admit that, along with a bunch of signal interference concerns to which basically no one else attaches any validity, ''Massport would also lose revenue associated with the operation of the central WiFi antenna system at Logan."

The Dying CD Bestseller?

As Ashlee Vance of The Register notes here: "2005 marked a banner year for quashing CD piracy in the courts and on the internet. You'd think sales would gave gone up once again, if you buy into the RIAA's way of thinking." They didn't, of course.

US CD sales in 2005 fell 3.5 per cent year-over-year, according to Nielsen Soundscan. That's quite a blow given that CD sales actually rose by 2.3 per cent in 2004. A sane person might suggest that higher energy costs throughout 2005 ate up a few of those sales or that pricey iPods left less cash to spend on albums. This logic escapes the Recording Industry Association of American (RIAA), which again attributes the fall in sales to piracy and which last year attributed the rise in sales to better anti-piracy measures.

The situation is even more dire for blockbuster releases. Chris Anderson, of "Long Tail" fame, presents some data that shows how few all-time topsellers there have been since the millenium (and none at all in 2004 or 2005).

I wonder a bit about the conclusions, some of which are also raised in the comments to the post. For example, I wonder whether it's partly a case of newer albums not having had time to become bestsellers yet. Chris Anderson's take on this (from the comments) is: "Typically albums do most of their total volume of sales in their first year. Looking back at the past years' releases it's a safe bet that there are no slow-burners there who are going to creep in to the Top 100 over time." Which seems reasonable enough. Perhaps there are cyclical trends at work here too, as there have been historically.

However, it's hard to dispute that we live in a world in which the mainstream (often manufactured) pop star may not be dead but where he/she/it competes with accelerated Internet time, (especially) solitary hits being sold on the iTunes Music Store, and a generally much more fragmented and information-rich market. It's much analogous to TV. As the network news shows change their guard, they'll never be the focal point they once were and, as Derek comments, "there will probably never be another TV show with the kind of ratings that "All in the Family" had in the 70's, because there are more channels now. Instead of one show with a 34-rating, there are 50 shows with a 1-rating."

This Week's Memo to Product Designers

Just because you can do something doesn't make it a good idea.

Case study #1: I received one of these travel alarm clocks as a gift. It almost goes without saying that the user interface was atrocious. Anytime instructions for setting a clock begin with "Press Button 1 three times..." you know it's not going to be pretty. But the general badness of UIs is a diatribe for another day. On top of everything else, this clock's "designers" (although that word implies a certain thinking about design that was clearly absent) felt a need to pollute their already over-complicated and overly-modal device with a stopwatch function. I would have hoped that the iPod's success would have encouraged more "Simpler is Better" designs. Apparently not.

Case study #2: The superbright LEDs that have appeared over the past few years have been a real boon for many applications. I do a fair bit of hiking, climbing, camping and the like. The headlamps that use these new LEDs are fantastic. They're small, light, have great battery life, and there's no bulb to burn out at the most inopportune moment. It's even practical to just carry a spare in an emergency kit. (And the latest generation of 1 watt SuperBright LEDs is even better with just a single LED throwing off quite a bit of light (as in this headlamp). But the same LEDs that work so well in flashlights aren't so great used as indicator lights in every piece of electronic gear. I do not, for example, really need a bright blue light staring out dolefully from the otherwise darkened direction of my television and entertainment setup. I've seen cases so bad that a piece of electrical tape was needed to mute the spotlight. Just because LEDs can be so much brighter than in the past doesn't mean that you have to amp up the brightness for everything.