Crawford's Corner

Citation

Crawford, W. (2000), "Crawford's Corner", Library Hi Tech News, Vol. 17 No. 9. https://doi.org/10.1108/lhtn.2000.23917iab.001

Crawford's Corner

Perspective

The Inevitability of Ebook Readers

Give Gemstar International credit: they have big ideas and make big claims. Best known for those numbers next to television listings in TV Guide and elsewhere (I'm not making this up), Gemstar purchased NuvoMedia and SoftBook Press. NuvoMedia and SoftBook produce the only dedicated ebook readers that have made it to the market. This means that the competition for ebook readers is now between Gemstar and Gemstar. Sales of both devices were so low that Gemstar didn't need to report sales as a material issue in closing the takeovers.

According to Information Today, Gemstar has a "long-term strategic agreement" with Thomson Multimedia in which Thomson commits to "a multiyear product shipment plan aimed at placing tens of millions of ebook devices into consumers' homes and establishing ebook readers as the preferred choice for reading novels and periodicals."(Emphasis added.)

Henry C. Yuen, chair and CEO of Gemstar, says, "Thomson's leadership position in the consumer electronics business and its commitment under this agreement will virtually ensure the establishment of ebook readers as a popular consumer electronics category worldwide." (Emphasis added.) In other words, it's inevitable: in a few years, ebook readers will be the preferred choice for novels and magazines.

How can it miss? Surely you're aware that Thomson is the leader in consumer electronics, known for its steady stream of innovative products that always succeed? You say you've never heard of Thomson? Think RCA, Thomson's primary consumer electronics brand name in the US. Thomson also sells consumer electronics under the GE, Proscan, and Telefunken brands-you thought General Electric produced GE radios and TVs? If you look at Thomson's corporate history on the Web, you see that it intertwines RCA's history, even though Thomson purchased the brands from GE in 1988.

Thomson's separate history doesn't show any consumer electronics leadership that I could see, so let's look at RCA's remarkable record.

At roughly the same time that Columbia introduced 12" 33 1/3RPM long-playing records, RCA pushed 7" 45RPM records as a preferable standard.

RCA introduced CD-4, by far the most problematic of the quadraphonic LP systems.

RCA introduced SelectaVision MagTape videorecorders in 1973. Later, RCA's marketing of Matsushita's VHS helped assure that Sony's higher-quality Beta would be swamped in the marketplace-but RCA did not develop VHS.

When it comes to videodiscs, RCA shows up twice. The firm announced Selectavision Holotape in 1969, abandoning it in 1972. RCA came back with CED in 1975, although it didn't come to market until 1981, three years after LaserVision (Philips/Pioneer). CED was a technological fiasco, the only marketed videodisc using a grooved disc and physical-contact stylus. It had much lower picture and sound quality than LaserVision-but it was cheaper. The net effect was to deny LaserVision any real chance at widespread success. After 38 months, RCA dropped CED. RCA assured the half-million unhappy owners of CED players that there would always be discs for them. "Always," in this case, meant 24 months or less.

RCA's role in developing audiocassettes, Compact Disc, and CD-ROM is the same as that of Thomson: none whatsoever, although RCA did eventually sell cassette and CD players.

In 1987, RCA demonstrated DVI: an hour of color video on a CD-sized disc. Unsurprisingly, the announcement came a year after Philips and Sony introduced compact disc-interactive (CDI), and had what I believe to be the desired effect: it stalled CDI, which never really made it in the marketplace.

Thomson introduced Network Computers for household use under the RCA brandname. They sold a few thousand. Then they bought them all back: the services provider shut down and the devices were useless. We all know how well NCs have done as consumer devices: RCA was a leader in demonstrating the worthlessness of the market.

Let's be fair: RCA played a major role (possibly the major role) in developing NTSC color television and was an important factor in early radio and sound recording.

Thomson is part of the consortium that controls DVD, but I've seen nothing to suggest that Thomson played a significant role in the technological development.

There's no question: the RCA brand appears on a lot of television sets, although one wonders how many US buyers realize they're buying from a French company. RCA used to be a marvel at marketing-but their track record isn't such as to "virtually ensure" the success of ebook readers or anything else. That's particularly true given Thomson's market dominance in personal computing-which is to say, none whatsoever.

I could be dead wrong here. But if the NC fiasco is indicative, I'll bet Thomson has escape clauses in that multiyear agreement. If Thomson and Gemstar succeed in making ebook readers "the preferred choice for reading novels and periodicals"-which I would interpret as meaning that more novels and periodicals are read on ebook devices than in print form-I will happily eat a copy of this column. But then, I could be long dead by that time. Here's a shorter-term offer: If Thomson and Gemstar sell millions of ebook readers into American homes in 2000, as indicated by verifiable sales figures (not shipment figures), I'll eat a copy of this column during ALA Annual 2001. And I will say, publicly and in print, that I goofed.

Press Watch I: Articles Worth Reading

Askt, D. (2000), "Imagining the digital divide," The Industry Standard, Vol. 3 No. 16, pp. 51-2.

Askt suggests that the "digital divide" is nonsense-that it's intended to make dot-commers feel guilty while letting the rest of us feel good without doing much of anything. "Unfortunately, what separates the rich from the poor in this country isn't a computer: it's an education." If all children know how to read, write, do arithmetic, think, and organize information, they can pick up computer skills without any special effort. Isn't that how you did it? Isn't that how we all did it?

Askt imagines an alternative Clinton-say the president as played by Martin Sheen in West Wing-visiting a low-income school and asserting that reading is key to the children's future. He enlists publishers to donate books. He funds tutors for students having trouble. He sends notes home to parents urging them to turn off the TV and read books with their children instead. And, by the way, since it's hard to focus on books when life is a constant struggle, maybe a system of national health insurance, better gun control, and better earned income tax credits would be more useful than a bunch of free computers.

Askt is on to something here. But books and tutors aren't as sexy as free computers-and we all know that only Communist countries (e.g. every industrialized nation in the world except the United States) have national health insurance programs. So pull those cables and donate those obsolete computers: we'll bridge that digital divide!

Fallows, J. (2000), "Feeling sociable," The Industry Standard, Vol. 3 No. 19, pp. 51-2.

Remember the Stanford study that concluded that heavy Internet users lose contact with humanity? At the extreme, I buy that: people who become obsessed tend to lose touch with the rest of us. In this column, Fallows points out some problems with the Stanford study and its broader findings-and offers quite different findings from a Pew study.

The Stanford study collected all its data online, even though it was comparing serious Net users with others. Thus, its population of nonusers consisted of people who would deal with a set-top box and answer questions over the Net as to why they don't use the Net. Additionally (according to Fallows), the study used coarse assumptions-for example, if people said they spent less time talking on the phone, this was considered a reduction in genuine social interaction, dismissing increased e-mail use.

Pew's study used telephone interviews and asked fairly specific questions. It found a substantial increase in Net use among women (a finding matched by other data), to the point that we've reached rough gender parity. More to the point, many people use e-mail as a form of social interaction-interaction that might not take place otherwise. That's certainly true for me and most people I know.

"Macworld's Internet privacy guide," (2000), Macworld, July, pp. 62-77.

If you're a Mac user, this section bears reading; it's well-written and offers useful advice.

Engst, A. (2000), "Your PC passport," Macworld, August, pp. 84-92.

If you deal with mixed PC-Mac environments, read this article. In several sections and with a handful of "tip" highlights, this article reviews ways to convert files between platforms, transfer files and handle mixed-platform networking, run Windows on your Mac, and share hardware.

Perspective

The Web and Its Makers

In early June, I served as the second keynoter for a one-day conference in Boston titled "Digital Reality II." My assigned title was "The Future Impact of the World Wide Web on Libraries and Archives," and I followed Tim Berners-Lee, inventor of the World Wide Web. In preparation for the talk, I read Weaving the Web, a book by Tim Berners-Lee (and Mark Fischetti: never forget the ghostwriter; Harper, SanFrancisco, 1999).

This is not a veiled "emperor watch" piece. While Marc Andreesen did not create the world's first Web browser (it wasn't the first and it wasn't a one-man effort), I have no problem calling Tim Berners-Lee the inventor of the World Wide Web. Not the inventor of hypertext, and he goes to great lengths in his book to make that clear; Berners-Lee just made it work in a straightforward, scaleable manner. He thought up URLs and HTTP, and made them work in the real world. He's currently head of W3C, the World Wide Web Consortium, working out of MIT. Berners-Lee is a charming man, an excellent speaker, and remarkable in that he has not exploited his fame and accomplishments for a big-bucks dotcom position. He still works to improve the Web rather than to make a quick buck.

Do I agree with everything he says and do I think the book is flawless? Of course not. I think the book overstates the extent to which the Web is, or should be, "part of the very fabric of the web of life we all help weave." He says that he expected the Web to "be able to provide basic hypertext navigation, menus, and simple documentation such as help files, the minutes of meetings, and mail messages-in short, 95 percent of daily life for most people." I shudder at the thought that meetings and mail messages represent "95 percent of daily life," even if you add in all possible computer uses.

Berners-Lee has a slight blind spot when it comes to the accomplishments of W3C. The book touts W3C's development of the PICS rating system, PNG graphics format, and SMIL-but none of those has had much real-world effect. Anyone care to guess the ratio of JPEG and GIF files to PNG files on the Web-or the number of people and sites who make any use of PICS? As for SMIL, if you don't know what that is, that says a lot about its impact.

The vision of Berners-Lee that most unsettles me is also, I believe, the most unrealistic. He envisions a future in which XML tags make all Web documents understandable at a machine level-and a "Semantic Web" in which powerful computers "make all the data in the world look like one huge database." I find that vision horrifying. I said so at the conference, perhaps not in those words-but, as I also said, I take comfort in the sheer improbability of the Semantic Web.

Weaving the Web is an easy read. Berners-Lee is a remarkable man with remarkable achievements. It was a pleasure to meet him and listen to him. Oddly enough, he was the only other speaker that day who did not use PowerPoint, overheads, or other audiovisual aids. Both of us just stood at the podium and talked to the 357 attendees-and Berners-Lee was candid and forthright in answering questions. The rest of the conference was also impressive, with three afternoon speakers on aspects of preserving digital information. When NELINET announces "Digital Reality III" take a look: it may be worth your time.

You can read the written version of my speech (similar to the spoken version but longer). Go to my Web site ( http://home.att.net/~walt.crawford), click on "Coming Events and Links...," drop to the bottom of that page, and click on the speech title. If you want to avoid my site, you can go directly to http://www.nelinet.net/conf/pres/pres00/crawford.htm.

Network Solutions

When I read Weaving the Web, I was struck by Berners-Lee's dislike for Network Solutions (NSI), the outfit that has handled most Internet domain registration for some years. A brief piece in the June 19, 2000 Industry Standard, while not directly related to Berners-Lee's problems with NSI, certainly gives one pause.

Transmedia Communications owns an adult Web site, Naked.com. Suddenly, someone else had that domain name. Web.net and Web.org saw the same transition. In each case, the hijacker adopted e-mail header information from the legitimate owner (not difficult to do) and sent NSI a message asking them to change the contact information. Then the hijackers asked another registrar to transfer the domains to them, and the other registrar contacted NSI. NSI, without verifying the contact, transferred the domains. According to the people involved, "NSI ignored repeated phone, fax, and e-mail pleas from the legitimate owners."

In another case, a domain was hijacked for ten days, accompanied by a demand for $10,000 from the hijacker. This domain had been generating $500,000 a year in revenues from 13,000 unique visitors per day; when the owner got the site back, there was no traffic at all. That owner is suing NSI.

What we have here is (unforgivably) clumsy customer service and a set of events that must give pause to anyone who owns valuable Internet domains. Does your state library association or regional services vendor have an ".org" domain? Would you be unhappy if that domain suddenly turned up in hands unknown? Now consider the response from an NSI spokesperson: "I don't want to seem callous to what these people are going through, but compared to the overall growth of the system, it's relatively minuscule." No apology, no offer of compensation: essentially, "Stuff happens."

Press Watch II: Commentary

Bennetts, L. (2000), "e-Stress," FamilyPC, Vol. 7 No. 6, pp. 93-8.

"E-mail overload, crashing computers, and ever-changing technology. No wonder we're more frazzled than ever." Thus begins an article that has its merits, as long as you read it thoughtfully and take it with a bushel or two of salt. The author's preteen daughter agitated for her own email address-then got frantic to the point of tears when she ignored it for a month. That's "e-stress."

Bennetts' second example is her husband. She gave him "the expensive Palm organizer he wanted for his birthday," and six months later he hasn't programmed it. "At this point he's so frustrated that the whole subject puts him in a bad mood. At home, he's bombarded with faxes, phone calls, and-yes-emails from his office; evenings have become just another extension of the work day." Meanwhile, Bennetts doesn't watch TV at all anymore because her husband's love of home theatre gadgets has made it too difficult to bother with.

She quotes Walter Mossberg of the Wall Street Journal, an "expert" who recorded 23 incidents over seven days in which his six Windows PCs "either crashed, froze, or exhibited other unexpected or puzzling behavior." What's wrong with this picture?

Taking the last first, I'm always puzzled by the consistent claims that expert Windows users get three or four crashes a day. What's wrong with me? At work (Windows NT), I turn the computer on every Monday morning and off every Friday afternoon; I've had to reboot it an average of maybe once a month in the last year. At home (Windows 98), I turn the PC on as I need it-and I may have one freeze or crash in a month. But then, I don't consider myself a Windows expert and I can't imagine using six computers at once.

Letting work take over your home life isn't e-stress; technology just makes it easier to have bad habits. Later in the article, an expert says, "We're never unplugged," to which I say, "Speak for yourself." Cell phones have off buttons. Very few businesses force their employees to have full-time Internet connections at home.

The article cites one working mother who's a law professor, a Latin America expert and business consultant "who spends much of her time in Cuba," a TV commentator and an author-as well as a soccer and Little League coach. She does her e-mailing from 2 a.m. to 6 a.m. and basically doesn't sleep; her lawyer husband works more than she does. Her family takes computers on their "vacations." She finds this all empowering. I find it sad (that's the polite word), but astonishing that a case this extreme would make it into the article.

We're told that we'll be faced with technological overload at an increasing pace, and the "only alternative" to facing up to them is to refuse to face them at all. A sidebar reveals the truth: some people need to learn how to turn off their machines and get a life.

The author blurb for this charming article tells us that Leslie Bennetts has "been using computers for 25 years and is still afraid of them." No comment.

Kosmer, J. (2000), "The e-bill is in the mail," Computer Shopper, Vol. 20 No. 7, pp. 174-7.

Whenever I read stories about electronic bill-paying systems, I read the same claims: they'll save you time and money. This particular story reviews three current services and includes pricing for two of them-and for the life of me I can't figure out the "money-saving" aspect. (As for time-saving, I have yet to see the case in which getting to a Web site, navigating, and clicking on the appropriate places was faster than writing and recording a check, but maybe some people write very slowly.)

I use the claims made within the article itself: "Consumers waste an average of $200 a year on paper-bill-paying costs-check writing, postage, and so on." "Waste" is a remarkably judgmental word ("spend" might be a little more neutral), but let's figure that out. Postage is a flat $0.33 per bill. Checks cost three or four cents each; let's call it a nickel to make this a fair comparison. There is no "and so on" that I can see: the return envelope always comes with the bill, return address stickers are so cheap as to be meaningless, and you need more electricity to use a computer than to light up your desk.

If it costs $0.38 cents in supplies to pay a bill, then the average household must be paying 526 bills a year. That strikes me as a little high, but then my household must be a lot below average. (Checking past registers, we appear to write 150 to 200 checks a year-with 200 being the year in which we bought and sold houses, with a resulting move.) If a household writes 180 checks a year, total expenses will be roughly $68. But let's look at the fully priced e-bill alternatives, using the supposed "average" household (526 bills a year comes out to roughly 44 a month) and one of us "below average" households that consolidates most items into two or three credit card payments, writing 15 checks a month.

PayMyBills costs $8.95 a month for up to 25 payments, plus $0.50 per payment beyond 25. For my household, then, we'd pay the minimum: $107.40 a year, a bit less than $40 more than we pay now. An average household would average 19 extra payments a month, bringing the cost to $18.45 a month or $221.40 a year-still $21.40 more than the $200 currently being "wasted." How does this save money?

PayTrust does no better: its rates are identical to PayMyBills. StatusFactory makes it more difficult: you sign up for clumps of payees. Ignoring the lowest level (five payees, $3.50-that's $0.70 per payment, twice what checks would cost), you can pay $8.95 for 30 payees or $29.95 for unlimited payees. If you currently write checks for exactly 30 bills, this is a better price than the competitors-but to cover 44 bills, you'd be paying $359.40 a month, $159.40 more than you're "wasting" now.

I'm painfully aware that the track record of e-payment services is imperfect: sometimes the bills don't get paid. Of course, my check can get lost in the mail as well: stuff happens. I can see why the payees might like payment services (electronic billing should be cheaper than paper billing, and they don't have to deposit all those checks)-but in that case, the payees should pay for the service.

Maybe these services are convenient; maybe the time saved makes it all worthwhile. But save money? What am I missing here?

Unix and IBM Mainframes?

It's always interesting to see limited historical awareness. A one-page item in the June 26, 2000 Industry Standard tries to add perspective to IBM's decision to offer ThinkPads with Linux preinstalled. It's a strange little story. First, the reporter claims that IBM support for Linux may be a matter of revenge: "it was just 15 years ago...that another upstart technology, Windows, sparked the PC revolution and humbled IBM's ponderous mainframes." How many errors can you count in that sentence? I count three: that Windows had any significance in 1985, that Windows sparked the PC revolution, and that PCs "humbled" mainframes. Windows 3, the first Windows of any significance in the marketplace, appeared around 1990. In terms of widespread use of serious personal computing, the PC revolution was sparked by-well-IBM itself. And anyone who thinks PCs have "humbled" mainframes doesn't understand minicomputers or mainframes.

Then there's a truthful text slightly undermined by the accompanying chart. Linus is "the third fastest-growing desktop OS, behind Windows 98 and NT Workstation-and ahead of Macintosh OS." That's true enough: comparing 1999 shipments to 1998, Linux increased by almost 80% while the Mac OS increased by 28%-while Windows NT increased by 116% and Windows 98 by 224%. For that matter, the share of new shipments occupied by Linux was the third fastest-growing: which is another way of saying that its share relative to Windows 98 or NT declined during the year.

How can that be? Consider the numbers. Mac OS went from 4% to 5%. Linux doubled: from 2% to 4%. Meanwhile, Windows NT went from 11% to 21% and Windows 98 went from 14% to 42%. This suggests that Windows 95 and Windows 3.x declined from around 66% of shipments in 1998 to around 25% in 1999, leaving perhaps 3% for other desktop operating systems.

Here's what really caught my eye. "Nearly all of IBM's software and hardware is being retooled for Linux. The investment is hardly a resource drain for IBM: Software written for Unix, the core of IBM mainframes, can be reworked to run on Linux."

I must admit, in some 30 years of programming for IBM and IBM-compatible mainframes, that's the first time I've heard that Unix was at their core. Whatever happened to OS/MVT?

Breitzer, F., and Gore, A. (2000), "USB 2.0 vs. FireWire: the debate heats up," Macworld, August, pp. 31-2.

This article could be a straightforward discussion of two different high-speed PC interfaces-but a combination of peculiar reporting and platform wars make it a bit more interesting. The facts are that IEEE 1394, the Apple-developed high-speed interface otherwise known as FireWire, hasn't done nearly as well as it should have-and that a new version of USB, the Universal Serial Bus, may offer even faster data transfer.

The first sentence in the article gives a strong (if self-referential) clue on its bent: "Politics and technology shouldn't mix-but in Silicon Valley, they often do." If Macworld (and Andrew Gore in particular) was not so deeply imbued in platform politics, this would be a different story. Consider this sentence: "Call it connector envy or a sincere effort to improve its existing standard, but Intel is now developing USB 2.0..." Can you say "Poisoning the well"?

It goes on. Gore & Breitzer say that FireWire is "also known as iLink in Sony products or IEEE 1394 on Windows-based computers," which is mostly nonsense. Perhaps the authors don't know what the IEEE is, or don't understand that IEEE 1394 is the proper designation for an adopted international standard; FireWire, like iLink, is a trade name for IEEE 1394.

The article reveals one major reason that IEEE 1394 hasn't been widely adopted: Apple wants a hefty licensing fee for it. That's not the way to get a standard adopted widely-but it's the Apple Way. But there's always spin, as in the following quote: "Intel doesn't want to do it the FireWire way because it didn't invent FireWire. What they're trying to do is control the technology. Intel can't do anything to alter the 1394 interface...[FireWire] was designed so that one day you could forget the computer altogether." In this case, you must consider the source: a VP at LaCie, a Mac-centric peripherals maker that has traditionally charged much higher prices for its Mac-only products than you'd pay for comparable Windows products. Another way to read that statement: Intel doesn't want to buy into an interface that's exclusively controlled by Apple or to pay big licensing fees to a much smaller company. Big surprise there!

What makes this truly peculiar is the reality of Apple's own computers. Every iMac and iBook includes USB ports; indeed, all new Macs (desktop and portable) rely on USB connections for most peripherals. But some iMacs don't have FireWire ports, and neither do iBooks. The question arises: if FireWire is the way to go, why isn't it a standard at Apple?

Product Watch

Buy More Stuff!

The Industry Standard for June 5, 2000 reports on print hyperlinks from Digimarc, DigitalConvergence, and GoCode. What are print hyperlinks? In a way, they're a throwback to 15-year-old technology like Cauzin Datastrips (remember those, once published in this very serial?): marks on printed pages that can be read as digital information. But unlike Cauzin's straightforward "scan in the data" failure, these are ways of connecting print ads to Web commerce. Digimarc prints a little "D" on the bottom of an ad; that lets the hip reader know to hold up the page to a Webcam equipped with the right software. The Webcam reads the hidden watermark and opens up a Web browser to the site, so that the reader can Buy Stuff. Digimarc sent 25,000 Webcams out to Wired subscribers to kick off the idea-after all, who better to understand that new technology is all about buying more stuff?

Digital Convergence prints a small bar code on the page, taking up a little more space than Digimarc but using simpler technology. They plan to send 1.2 million free scanners to readers of Forbes and Wired-scanners with no purpose other than to connect you to advertisers' Web sites. They're working with Cross to produce portable readers-and they're working on special broadcast tones so that, if you have your PC connected to your TV, a TV commercial will let your PC go directly to the appropriate site so you can Buy More Stuff. You don't have your PC and your TV connected? Now you have another reason to keep them apart.

Are you ready to add new hardware to your PC just so you can avoid typing in an advertiser's URL? Isn't this what convergence is all about?

$600 Picture Frames

Surely you must tire of that single picture occupying a frame in your office or at home. How can you resist Ceiva's $249 Picture Frame or Digi-Frame's $599 DF-560-which lets you use a single frame to show lots of pictures. Who wants a superb 8 x 10" print of a picture that really matters to you, when you can have the 5 x 7" or 3.5 x 4.5" picture-of-the-hour?

Oddly, it's the $249 model that has the large-snapshot size. Ceiva's unit needs a phone jack (What? You don't have phone jacks by all your photos?), through which it connects to Ceiva's Web site and downloads up to ten images-your own or ones you've borrowed from friends or relatives. Plan on $3 a month (if local numbers are available) for the service, but that means you get fresh pictures every day!

Digi-Frame can actually accept pictures directly from CompactFlash and SmartMemory cards or download JPEGs from a PC, a camera, or a Web site: there's no monthly charge-but you're paying $600 up front for an image smaller than many snapshots. But you can get fancy transitions between pictures or display them in black-and-white or sepia tone.

Or you could get a life, but that might involve more effort.

The sentence above ended this item when I first encountered digital picture frames in one of the PC magazines-but the August 2000 FamilyPC has a two-page spread including four desktop digital picture frames and a portable version. What changed since the earlier writeup? Ceiva's subscription is now stated as $35.99 to $94.99 a year and a new Digi-Frame DF-390 costs a mere $399. Sony's entered the fray with a "svelte" CyberFrame PHD-A55 for a mere $899, but it can even display short movies from its MemoryStick slot. Weave Innovations has a $299 Storybox Connected Frame; it connects directly to the Internet and displays your latest photos (as well as traffic reports and other information) for $3 to $10 a month. The $350 Wallet is a battery-operated device that takes a CompactFlash card and shows photos on its 2.5 x 3.75" screen. The article also notes software for displaying photos on Palm or Windows CE pocket PCs.

It's never safe to bet against Sony, but there must be better uses for $899 than showing a rotating set of snapshot-sized photos on your desktop.

Shots in the Browser Wars

Is there still a browser war? At this writing, the available version of Netscape 6 seems to be too buggy for serious use, but that could change. Meanwhile, the July 2000 Macworld includes a shocking review of Internet Explorer 5-shocking, that is, if you believe that the Mac and Microsoft don't mix.

The subtitle of the review and the last sentence may provide the gist: "Hands down, the best browser" and "This is a welcome case where the upgrade decision is a no-brainer: just get it." The rating is 4.5 mice. Why? It's speedy, it's the first browser on any platform that fully implements HTML 4 and CSS-1 (supposedly, IE5 for Windows doesn't), and it's appropriately Maclike. IE5 for Windows is much faster (and more stable) than Netscape 4.x or IE 4.5 as well. This does seem to be a case in which the Redmond Devils have improved the field.

Digital Whiteboards

Macworld for July 2000 has a photo and brief write-up of a wonderful new device from Brother International: the $699 CopyPoint 2000 electronic whiteboard. The brief review begins: "Copying notes from a whiteboard onto paper is tedious. But this thankless job-and the ensuing writer's cramp-could be a thing of the past." The CopyPoint weighs 32 pounds and includes its own printer.

It's a good idea-but it's not a new idea. We've had a printing whiteboard at RLG for at least five years, and I've seen a number of models since then. Unlike our old standby and like the Brother, some newer ones will download whiteboard contents to a PC (but not to a Mac, in the Brother's case), and the newer ones are doubtless cheaper and lighter.

I have to wonder why this device, which won't connect to a Mac, deserved half a page of Macworld's slender editorial space-unless it was to chastise Brother for not being Mac-compatible. Perhaps they've never seen a printing whiteboard in action?

Big Hard Disks: How Cheap Will They Be?

PC World for July 2000 reviews two huge hard disks: IBM's $625 Deskstar 75GXP (75GB) and Maxtor's $329 DiamondMax 60 (60GB). While the IBM offers higher performance, the Maxtor is cheaper for its size-less than 61 cents a megabyte!

How many people need hard disks this large? Quantum and Seagate doubt there's a huge market for such massive drives, but there are cases where such capacity would make sense.

Perhaps the most interesting part of this one-page review is the chart in the middle: price per gigabyte for PC (IDE) hard disks, including projections for the next three years. The chart begins at $261.84 per gigabyte in 1995-ignoring the much higher prices in the late 1980s and early 1990s-and proceeds down in a fairly regular manner, basically dropping by half each year. IDC believes the trend will continue for at least another three years, slowing slightly in 2002 and 2003-arriving at roughly $1.15 per gigabyte in 2003.

Is that plausible? To my mind, it depends on whether people want very large hard disks. You rarely see a current-production hard disk selling for less than $100 (maybe $95 at discount operations), which means that the smallest hard disk keeps getting bigger. Then again, the cheapest hard disks cost a lot more per gigabyte than the largest units-again, unsurprisingly. I believe you will see very large disks that cheap in 2003, but only if you want at least 100 (and more probably 150-200) gigabytes.

That may be a more plausible way to look at hard disk prices: for a given speed and manufacturer, you're paying a base price plus a reasonably constant price per gigabyte. I looked at the July 2000 ad for Dirt Cheap Drives, a long-time discounter of hard disks. For Western Digital IDE drives, you could reasonably say that the price was $62 plus $4 to $5 per gigabyte; for Quantum, $60 plus $4 to $5; and Seagate, $65 plus $3.50 to $4 per gigabyte. That's probably a more meaningful figure than the $6.50 per gigabyte that Western Digital's 30GB drive costs. I believe that the incremental cost per gigabyte for very large hard disks in 2003 is likely to be $1.15 or so-but you'll still pay $90-$100 for the smallest contemporary hard disks.

Review Watch

Desktop Computers

Broida, R. (2000), "Big PC on campus," Computer Shopper, Vol. 20 No. 8, pp. 154-62.

I'm not sure what makes this set of five PCs "back-to-school systems" unless $1,299 is a magic price point. That's all these systems have in common. Other than price, the configuration requirements included at least 64MB RAM, 10GB hard disk, 3D graphics with 16MB video RAM, an 8x DVD-ROM drive, 16"-viewable display, V.90 modem, three-part speaker system, and a productivity suite. It's an odd group, with one huge direct vendor (Gateway) and four small fry.

The small fry offer heftier configurations for the price: all but the Gateway include 128MB RAM and at least 15GB hard disk space. The magazine didn't enforce its minima: one system came with a two-piece speaker system. Computer Shopper consistently pushes price and configuration, with no weight for a company's reputation. The Best Buy of the lot is ABS' Performance 3: it has the best graphics card, although its non-graphics test results weren't tops, it arrived with defective equipment, and it comes with software that the magazine considers difficult.

Morris, J., and Venezia, C. (2000), "Pentium III hits 933," PC Magazine, Vol. 19 No. 13, pp. 32-40.

Why should 933 matter after gigahertz machines have already appeared? Because, unlike AMD's Athlon-1000, the Pentium III-1000 was a stunt: an anomaly required to compete with AMD, available in sample quantities of PCs from three manufacturers. The Pentium III-933 is Intel's fastest "real" CPU as of June 2000; this brief review covers systems from Compaq, Dell, Micron, and Quantex. The systems all include 128MB RAM (higher-speed RDRAM for the Dell and Quantex), 32MB double-data rate graphics RAM (with today's hottest graphics chip, nVidia's GeForce2 GTS, in all but the Compaq), 30GB to 40GB hard disks, DVD-ROM drives as well as CD-RW drives, 18"-viewable displays, Creative Labs' Sound Blaster Live! sound cards, and high-end speakers. Three of the four earn four dots out of five, although the Compaq lags behind the others in key respects. The sole five-dot system, and the Editors' Choice, is also the most obscure: Quantex' $2,899 SM933z. It had the top scores on most tests, a great configuration, and one of the lowest prices.

Digital Cameras

Grotta, D. (2000), "Picture these: 3-megapixel digital cameras," PC World, Vol. 18 No. 7, p. 84.

Although it's far too short for a reasonable group review and lacks any objective test information or sample photos, this quickie does offer another data point on the emergence of high-resolution digital cameras. Of the four cameras considered here, Olympus' $999 C-3030 Zoom offers the most flexibility and superb quality but with a significant learning curve. Epson's $999 PhotoPC 3000Z offers high quality with ease, although without the sophistication of the Olympus.

Grunin, L. (2000), "Picture this," Computer Shopper, Vol. 20 No. 7, pp. 168-73.

Set aside originality in article titles. This group review includes six two-megapixel cameras, not really approaching film quality but with a reasonable balance of price and performance. The cameras all capture at least 1600 × 1200 images (two cameras go a little higher) and cost between $501 and $734. There are photo samples, but they're too small to mean much-and there's no 35mm shot for comparison. The caption guides you to what the reviewer thought was most important in the image. Thus, while the Kodak image seems most lifelike, the Best Buy Nikon Coolpix 800 ($501) appears to have the best detail.

Johnson, D. (2000), "See what's developing," FamilyPC, Vol. 7 No. 8, pp. 116-20.

This mixed bag of eight digital cameras does include individual descriptions and feature summaries, but it lacks sample output and includes cameras ranging from 0.3 megapixels to 3.3 megapixels and $150 to $999.95. No surprise as to the highest rating: it's the $999.95 Nikon CoolPix 990, by far the most expensive and highest-resolution unit. More surprisingly, the second-highest rating goes to HP's $599 PhotoSmart C500 rather than Kodak's $799 DC290. No other cameras make the cutoff for the magazine's Recommended seal.

Displays and Graphics

Breitzer, F. (2000), "Flat panels: the next generation," Macworld, July, pp. 79-82.

If you're going to pay the premium for an LCD display, it makes sense to go digital (although you need an appropriate graphics card), and all six displays reviewed here take digital signals. That eliminates two conversion steps and should yield sharper images. This article offers good background. It also points out a problem with LCD displays that Windows-oriented magazines haven't stressed: their color accuracy and range aren't as good as CRTs, so they're not suitable for graphics professionals.

That caveat aside-and it's a big caveat for the Mac market-the editors find one 15" display and one larger display worthy of Editors' Choices. Among 15" units, Hewlett-Packard's $1,199 Pavilion FX70 offers generally fine quality, includes speakers, and can accept either analog or digital signals. If you need a bigger display, better have a big wallet: the choice here is Apple's 22" Cinema Display for $3,999-and you need a G4 Mac with an AGP slot to be able to use it. (If you can afford $3,999 for a display, you can certainly afford $1,600 or more for a G4 Mac.)

Pittelkau, J. (2000), "19-inch flat-screen displays," Macworld, August, pp. 38-9.

No, Macworld didn't review LCD panels in two successive issues (see the Breitzer review above). The ten displays reviewed in this too-brief roundup are traditional CRTs with 18" viewable "flat" screens. Unfortunately, the review not only lacks individual writeups, it also lacks any objective tests: all we get are a jury's subjective opinions.

The buying advice and ratings table also seem oddly mismatched. HP's $737 P910 (rated 3.5 mice) is lauded for image workers as well as for those using both images and text. Princeton's $539 AGF900 (four mice) comes in with strong color images but not particularly great sharpness. Two Mitsubishi models, the $659 Diamond Pro 900u (4.5 mice) and $479 Diamond Plus 91 (four mice), and the $625 Viewsonic PF795 (3.5 mice) also show excellent results across the board.

Here's the oddity. The last sentence of the review reads, "And although the Mitsubishi Diamond Pro 900u is expensive, you just can't beat its performance, features, and out-of-the-box Mac compatibility." In fact, that Mitsubishi gets the only 4.5-mice rating in the survey. But it's the only model named as "expensive" within the buying advice-even though the HP is more expensive by $78 and the Viewsonic is trivially less expensive.

Notebook Computers

Metz, C. (2000), "Notebooks," PC Magazine, Vol. 19 No. 14, pp. 165-92.

This roundup includes the latest entries from the dozen biggest names in notebook computers (as measured by market share), from IBM and Toshiba to Quantex and WinBook. New models make the lines between ultralight, road warrior, and desktop replacement notebooks fuzzy-but then, I've always felt that these lines mattered more to PC magazines than to consumers (particularly the line between road warrior and desktop replacement systems).

This is one of those detailed reviews where you really need to read individual writeups and draw your own conclusions. Editors' Choices go to three IBM ThinkPads (the $3,899 T20 84U, $3,799 A20P 6TU, and $2,000 i Series 1260) and Toshiba's $2,750 Portégé 3440CT. Quite a few notebooks earn four of five dots.

Somers, A. (2000), "Supercharged notebooks," Computer Shopper, Vol. 20 No. 7, pp. 154-61.

SpeedStep notebooks use Intel's newest Mobile Pentium III, which changes CPU voltage (and, as a result, speed) depending on the power supply-for this group of six computers, 650MHz with AC power (or if you force that speed), 500MHz with battery power. In past evaluations of SpeedStep technology, the real-world difference in battery life has been miniscule, typically one to three percent better at the lower speed. Some of this group do a little better, but with one exception the added battery life amounts to 19 minutes or less.

The criteria for this test included 128MB SDRAM, 15" display (or larger), 8MB graphics RAM, DVD-ROM drive, and at least 12GB hard disk space. The Best Buy goes to Quantex' $2,749 W-1511, very nearly identical to units from Dell and Windrover (all three systems are manufactured by Compal). The systems performed within a narrow range, but the Quantex offers better software and better DVD playback than the Windrover while costing considerably less than the Dell. Acer and Compaq units earn similar four-star ratings.

Optical Discs (DVD and CD)

Jacobi, J. (2000), "Spin city," PC World, Vol. 18 No. 7, pp. 143-52.

The lead may be a bit breathless, with its false analogy that "the old flexy floppy disappeared when faster, higher-capacity 3.5-inch disks caught on." It took about eight years for 5.25" diskettes to disappear after 3.5" drives became standard, and the difference in capacity was (I believe) never a factor: the smaller disks were sturdier, more compact, and a trifle faster. Set aside attempts to rewrite history and this is a worthwhile article, offering good background on DVD-ROM and DVD-RAM. As we all know, there aren't many DVD-ROM discs, and DVD-RAM is a specialized medium. But DVD-ROM still makes sense, if only to be prepared for the future and to watch movies.

The Best Buy DVD-ROM drive reviewed here is CenDyne's $147 CDI CD00042, which uses a Hitachi mechanism, mostly because it's reasonably well priced and offers useful features. If you're ready to write your own high-density discs, the Best Buy (of five DVD-RAM drives reviewed) is Creative Labs' $300 PC-DVD RAM 5.2GB, which costs a lot less than most DVD-RAM drives.

Printers

Read, D. (2000), "Workgroup printers," Macworld, July, pp. 34-5.

This brief roundup covers six monochrome laser printers selling for less than $2,000, with rated speeds higher than 15 pages per minute, networkable, and with paper trays holding at least a ream (500 sheets). The rule of thumb here is that you get what you pay for: the $1,749 Lexmark Optra T612 (four mice) is the fastest and offers the best print quality, while the NEC Superscript 1800N and Brother HL-2060 (each $999) offer the lowest quality and slowest printing. Unfortunately, the Optra is relatively expensive to run at 2.24 cents per page-almost a cent per page more than HP's LaserJet 4050N (which wasn't the fastest but did do the best job with graphics).

Service and Reliability

"Service & reliability," PC Magazine, Vol. 19 No. 13, pp. 192-204.

It's PC Magazine's annual service and reliability theme, but the magazine is so crowded with e-business advice that almost all of the narrative has disappeared. Instead, you get pages of tables covering desktop PCs, notebook PCs, Internet service providers, software, and printers. The historical columns don't appear, and once again PC Magazine ignores "buy again from this maker?" scores in assigning grades, even though Bill Howard elsewhere asserts (as I do) that buy-again rankings are "every bit as important as the grades."

I base my inclusion of "Top" vendors for PC values on this survey, and I think it makes sense to use the same criteria this year as last. To wit, I count a vendor as "top" if it meets the following criteria in the desktop PC survey:

At least 250 responses (out of the 8,500 responses to the survey)

A better-than-average buy-again score

At least three better-than-average scores out of the four other criteria

No worse-than-average scores (which I didn't explicitly mention last year).

Who makes the grade this year? Dell (the only "A" grade, with four better-than-average scores), Gateway, Hewlett-Packard, and Micron. That's almost the same list as last year-except that IBM has dropped off because they only had two better-than-average scores among the four other criteria. Then again, IBM has become a secondary player for desktop PCs; their primary interests in personal computing appear to be notebooks (where they tie with Dell as the only "B" grades; there are no "A" grades) and servers.

Who didn't make the grade? Apple doesn't have enough responses and has average scores across the board; Compaq has a solid D, with no good scores and two worse-than-average scores; Quantex would qualify based on scores but not on quantity (with 129 responses, roughly one-fifth of Micron, which is the smallest "top" vendor). If you care, Dell had the most responses (2,243), Gateway was second (1,769), and Compaq was the only other vendor with four-figure responses (1,400). The worst grade among desktop PCs was Toshiba, the only maker to have below-average scores on all four measures.

For ISPs, the table shows nine individual rankings and an overall ranking. Two ISPs stand out as having the most above-average rankings with no below-average scores: MindSpring (seven above average) and AT&T WorldNet (five above average); worst scores are shared by America Online and MSN Internet Access.

The only surprise among printers is that there are two A grades, with Panasonic getting the other one (HP has a lock on "A" grades among printers). I continue to be astonished that Hewlett-Packard's printers can get consistently above-average ratings on every single scoring category, even though HP printers represent 55 percent of all responses! On the downside, Epson drops to a "B" but is the only other brand getting a high buy-again rating. The worst scores go to Canon.

The Details

Crawford's Corner is written by Walt Crawford, an information architect at the Research Libraries Group, Inc. (RLG). Opinions herein do not reflect those of RLG or MCB University Press. Comments should be sent to wcc@notes.rlg.org. Visit my Web site: http://home.att.net/~walt.crawford.