# Crawford's Corner

ISSN: 0741-9058

Article publication date: 1 July 2000

## Citation

Crawford, W. (2000), "Crawford's Corner", Library Hi Tech News, Vol. 17 No. 7. https://doi.org/10.1108/lhtn.2000.23917gab.001

## Publisher

:

Emerald Group Publishing Limited

## Crawford's Corner

Perspective

The Post-PC Phenomenon

Personal computers are just too hard to use for the simple folk. And they're too expensive, too. Besides, they have all these functions nobody really wants or needsand they're ugly to boot. Add the known fact that Bill Gates is the devil incarnate, and it's obvious that we're on our way to the post-PC era. See, we all need to be connected to the Internet, all the time, all over the house (and in the car): when you're not on the Web (which is identical to the Internet, by the way), you're not functioning as part of the Digital Age. What if you're in the kitchen and e-mail arrives, or someone on your Buddy List wants to send you a message? What if you want to know somethingthe meaning of a word, for example, or maybe what's on TVand the source of all information (the Web) isn't immediately to hand? Whatever would you do?

The inevitable solutions come in many flavors. At work, you probably can't be trusted with the flexibility of a PC, so forward-thinking companies will install thin clients instead. They're a little like PCs, but with no removable storage devices so you can't get into trouble. The best thin clients don't even have hard disks: your files are stored on network servers, which also provide your software. Without a hard disk, you can't even use the Web to download applications not on the corporate approved list (of course, good corporations will prevent that with a combination firewall and filter in any case). What's the difference between a diskless thin client and a smart terminal? Two decades, for oneand, after all, putting a terminal on someone's desk would be denigrating, where a thin client is modern.

At home, you need Web appliances or Internet appliances: aren't the two interchangeable? IBM plans to make them; so does Intel; big outfits like Netpliance already do. Netpliance's $299 "i-opener" costs$299 (plus $22 monthly service fee) and gives you a big 10" screen and keyboard. Plug in the power, plug in the phone, and off you go. What could be better? Want to print off some mail? Why would you want to do that? Want to deal with attachments or download files? No, no, that's personal computing, and we're in the post-PC era. Some pundits believe Web appliances will replace PCs; others say that they're "supplements to PCs for Internet access." You know, for people who can run PCs but just can't figure out how to use them on the Internet. International Data Corp. projects that more than 55 million "non-PC devices" will be connected to the Internet by 2002, and that such devices will outnumber PC Net connections a year later. I suppose anything's possible. Is Post-PC Really Post-PC? Let's be fair here. When Scott McNealy says post-PC, he really wants to see PCs disappear. The same goes for Larry Ellison. But some journalists see it differently. A product review in the April 2000 Computer Shopper begins with these sentences: "When computer industry pundits speak of the 'post-PC era,' they're not referring to a time in which PCs are no longer in use. Rather, the term applies to the evolution of the PC into an integrated hub for all of the things we do." This precedes a review of "one of the possible forms this evolution can take," which is basically a new entry in the tired category of "home theater PCs," DVD-equipped PCs that use wireless keyboards and feed TV sets. That's not the way I've read it. Most often, "post-PC" pundits really do believe that PCs are just a passing fancy, that people would much rather have lots of separate single-function devices, even though they'd wind up spending more (particularly on a monthly basis) and probably getting inferior performance. Some see the return of the kitchen PC, so you can read your recipes right off the Internet. By now you've surely heard that your next refrigerator will order groceries automatically when it senses that you're out of something, and you must be wondering whether that Volvo driver in the next lane is reading e-mail at 80 mph... "Integrated hub for all of the things we do." What a charming projection, and so much in line with my opening paragraph. "All of the things we do" are digital, aren't they? Your house or apartment is really just a way to keep the rain from damaging your Internet devices. Of course my opening paragraph was intended as sarcasm. But if my job title was forecaster or market analyst and if I worked for some hotshot research company (and if I changed the wording of the third sentence), the paragraph wouldn't be much different than a lot of the stuff I've seen. More's the pity. NC, Nic, and Ellison's Obsession If this particular essay seems cobbled together, you're a perceptive reader. I began working on it in either February or March 2000, and bounced the (changing) essay from each "Crawford's Corner" in favor of more complete or timely comments. I'm glad that I did. The March 27, 2000 issue of The Industry Standard includes a long and fascinating article by Gary Rivlin, "The Network Computer Strikes Again!" If you have access to back runs of The Industry Standard, I recommend reading this article; you'll gain considerable insight into the "post-PC" phenomenon and probably get a few laughs in the process. The tease paragraph tells the key story: "Larry Ellison is pitching a new network computer, named Nic. The question is, why should it do any better than his old network computer?" It turns out that Ellison has his version of history about the NCand that version (naturally) involves evil manipulation by Bill Gates. Oddly enough, Larry Ellison doesn't even admit that the NC was a mistake. The company spun off from Oracle to handle NCs morphed into Liberate Technologies and changed its focus to interactive TV. Because Ellison owned a big chunk of Liberate and it went public while the Internet Bubble was still in play, he made billions (at least on paper). Ergo, the NC was a success: Ellison made money from it. On paper, that is. When this article was written, Liberate was "worth" around$9 billion, of which Ellison and Oracle own roughly half. When I checked it on April 21 (LBRT on NASDAQ), the capitalization was down to $3 billionstill a lot of money, to be sure, for a company with minor revenues and substantial losses. The article runs on the lower half of pages that have cartoon illustrations along the top. That's a reasonable move: this is the kind of story that belongs in High Tech Funnies as much as anywhere else. Apparently Larry Ellison was outraged because Microsoft's new-technology hotshot, Nathan Myhrvold, called him a "liar" in a Newsweek interviewbut, says this article, that's a misreading of the interview (it wasn't Myhrvold; it was a Compaq executive). In other words, if we're to believe Rivlin's article (and I've seen no letters from Oracle refuting it), a big part of the NC push stemmed from Ellison's inability to read articles carefully. Remember Ellison's claims: one million NCs sold by September 1997, 100 million by 2000. The truth is a little less glamorous: by the end of 1997, IBM and RCA (the top two NC manufacturers) had shipped some 10,000 units eachand RCA ended up recalling all its units. So are set-top boxes the death of the PC? Not for people who use their PCs for something other than low-quality Internet browsing and e-mail. So far, set-top boxes haven't set the world on fire, and neither has interactive TV. I know that, in my household, interactive TV is a losing propositionbut (unlike most pundits and futurists) I don't claim that my needs and tastes are universal or even typical. To quote what the Liberate people think this is all about. "During a Levi's commercial you'll be able to click on the girl with the pouty lips because you think you'd look good in the blouse she's wearing. Or punch the Domino's Pizza icon and charge a large pepperoni to your cable bill. ... 'We're talking about using the basic standards of the Internet to deliver a richer television experience.'" Easier impulse buying: that's the new mantra for interactive TV and the NC. The post-PC world is about easier ways to spend money. Maybe so, and maybe there really are tens of millions of people desperate to find new and faster ways to buy stuff. Are there enough to make a marketplace larger than personal computers (estimated at 120 million this year and growing more than 10 percent a year worldwide)? It seems unlikely to mebut then, so does Larry Ellison's apparent obsession with defeating Bill Gates. In 1996, Ellison said that the NC will "change our economy. It will change our culture. It will change everything." Want extra cheese on that pizza? (One of the "network appliances" mentioned earlier actually does have a pizza key: one press and it opens a connection to the nearest Web-enabled pizzeria. I couldn't make this stuff up if I tried!) When you read the article (if you read the article), pay attention to all the material about Liberate Technologies' "financial success." That success doesn't come from selling huge amounts of software or hardware at a profit; so far, Liberate is a tiny player in a crowded set-top market. The success is entirely on the IPO and (for a while) stock side: lots of people betting that Liberate would eventually make money. The bets don't seem to be as strong now (April 22, as I write this portion of this essay) as when the article was written. Way back then, at the end of February 2000, LBRT (the NASDAQ symbol) showed a market capitalization of$9 billion, making the Ellison/Oracle portion worth roughly $4.5 billion. As of market close on April 20, LBRT's market capitalization was$3 billion. That's still a lot of money for a company with trivial sales and huge expenses, to be sureand that's more stock market talk than belongs in a library publication.

What's the projection for Ellison's new Nic toy? "Millions" of machines will be shipping by June 2000. Running Linux, which we all know is the ultimate consumer operating system for people who can't be bothered with the complexity of Windows. (You'll be reading this later than June, and can check on just how many Nics have actually shipped or sold. This time, the company is New Internet Computing Company or NICC; the CEO is Gina Smith, a former journalist with no prior business experience.)

Interestingly, some 700,000 "network computers" or thin clients were sold in 1999. That's less than 1 percent of the PC market, but projections are that it could rise to be 2 or 3 percent in a few years. Does this make Ellison happy? Perhaps not: more than half of those network computers are Windows-based devices.

PC Values: May 2000

The standard configuration for May includes 64MB SDRAM, 24x or faster CD-ROM, AGP graphics accelerator with 8MB display RAM, V.90 modem, a 15.7-16.1" (viewable) display (called 17" by some makers), and wavetable sound with stereo speakers. "Pluses" and "Minuses" are shown where applicable, along with hard disk size, software, extras, and brand-name speakers. It may be noteworthy that there were no "minuses" and that each system includes 128MB SDRAM (typically special 100MHz RAM) and at least 16MB display RAM.

Top system prices are taken from corporate Web sites for Dell, Gateway, and Micronpc as of May 14, 2000. Other prices represent the best values encountered in the May 2000 PC World.

• Top, Budget: Dell Dimension XPS B600r: Pentium III-600, 20GB HD. Pluses: 128MB SDRAM, 32MB display RAM. Extras: MS Works Suite 2000, Altec Lansing speakers with subwoofer. $1,499, VR 11.49 (+19% since 2/2000, +22% since 11/99). • Top, Midrange: Micronpc Millennia Max 733: Pentium III-733, 30GB HD. Pluses: 128MB SDRAM, 32MB display RAM, DVD-ROM. Extras: MS Office 2000 SBE, Altec Lansing speakers with subwoofer, CD-RW.$2,125, VR 10.28 (+11% since 2/2000, unchanged since 11/99).

• Top, Power: Gateway Select 950: AMD Athlon-950, 30GB HD. Pluses: 128MB SDRAM, 18"display with 32MB display RAM, DVD-ROM. Extras: MS Works Suite 2000, Boston Acoustics speakers with subwoofer. $2,499, VR 9.17 (+8% since 2/2000, +31% since 11/99). • Other, Budget: CyberMax ValueMax 4: AMD Athlon-650, 20GB HD. Pluses: 128MB SDRAM, 16MB display RAM. Extras: WordPerfect Office 2000.$1,199, VR 14.00 (+27% since 2/2000, +20% since 11/99).

• Other, Midrange: Quantex SM800s: Pentium III-800, 30GB HD. Pluses: 128MB SDRAM, 32MB display RAM, DVD-ROM. Extras: WordPerfect Office 2000, Altec Lansing speakers with subwoofer, Zip drive. $1,899, VR 11.54 (+18% since 2/2000, +6% since 11/99). • Other, Power: CyberMax Enthusiast 4: Athlon-1000, 40GB HD. Pluses: 128MB SDRAM, 32MB display RAM, DVD-ROM. Extras: WordPerfect Office 2000, Altec Lansing speakers with subwoofer, CD-RW.$2,499, VR 10.21 (+18% since 2/2000, +21% since 11/99).

Press Watch

Manafy, M. (2000), "Circuit rewires the music magazine," EMedia, Vol. 13 No. 2, pp. 24-6.

Remember CD-ROM magazines? They were one of the next big trends a few years back, but they never really worked. Dozens were attempted, and one or two might still be around, but the form just didn't please people. Adding a music CD-ROM to a music magazine or a CD-ROM full of demos to a gaming magazine makes excellent sense and quite a few magazines do this now. But print magazines have special qualities which neither CD-ROM nor (I believe) the Web can replace.

So now we have a DVD-ROM magazine, Circuit. This article offers a clear look at how it works and why it could make sense. It's really a DVD magazine with some DVD-ROM featuresand the publishers have already realized that the magazine model isn't working. "Circuit will honor any existing subscriptions, but has put future subscription orders on hold for now." Will some modified version work? Check back in a year or soor check your usual magazine outlet to see whether Circuit is available.

Seymour, J. (2000), "CD-Rs and ­RWs: into the jungle," PC Magazine, Vol. 19 No. 7, p. 103.

This is the second of two Jim Seymour columns about CD-R and CD-RW, and I include it only to add some possibly useful comments. First of all, Seymour gets confused on compatibility issues. On the basics, he's right: virtually all CD and CD-ROM drives can play CD-Rs, but many (perhaps most) can't read CD-RWs, while almost all DVD drives can read CD-RWs but many can't read CD-Rs. His answer seems to be to ignore CD-RW; a better answer may be to check on the drives you're using. At this writing, I haven't yet added a CD-RW drive to my PC, but I suspect thatlike SeymourI'll use a lot more CD-Rs than CD-RWs. But that's not the thrust of his article or my comment.

Jim Seymour is all business: he's having none of this "make your own custom audio CDs." He sees the obvious application of writable CD as being "quick, safe, ad hoc backups." So far, so good. He goes on to note that most data files on a PC would fit nicely on a single CD-R, if you could find all the data files in amongst the applications. (You don't need to back up all your applications every week or month: you can restore them from the original CD-ROMs, at worst losing your preferences.)

While his underlying solution is the same one that I would recommend, he goes about that solution in a manner I find distinctly old-fashioned and limiting. The underlying solution is to put all your data files into a known directory structure and just back up that structure every week. Every MS Office application can be modified to save to a different default directory with a minor tweak, and I suspect that's true of almost every other application.

I disagree with Seymour on the methodology. He thinks you should subdivide your massive hard disk into multiple partitions (which become logical drives), putting all your data on one of the partitionssay "E:"while applications go on the C: drive. You can even name the E: partition "DATA." Then you just back up E: to your CD-R or CD-RW.

Finebut that means you must decide in advance how much of your hard disk you should allocate to program files, temporary files, and data files. Changing partition sizes is a nuisance even with PartitionMagicand there's no reason to do so if you're using Windows 98 or any other operating system that can handle large single partitions.

On my previous systemwith a 2GB hard disk, which turned out to be just a little too small for comfortI did the partition trick: 980MB for apps (to get the smallest possible clusters in Windows 95), 980 for data, and 40 for extras (a good place to put temporary files). Fine, until applications took up 900 of the 980MB on C:. I wound up loading CD-ROM titles onto D: (the data drive) and separating out actual data files by standard subdirectories within the drive.

So when I got my new PC with a 10GB drive, I probably could have allocated two or three logical drives. After all, 5GB (splitting the drive equally) is a lot of space. But why bother?

What I didat home and at workwas simpler and far more flexible. I simply created a "D" directory at the root level, and treated that directory as my data "drive." (I could call it "DATA\" rather than "D\" but who needs the extra typing?) All my data directories are organized under D\ (technically, C:\D\), and I backup everything in D\ once a weekcurrently to a Zip drive, until I move to CD-RW.

There's absolutely no difference in performance or organization capabilities. The only difference is that I haven't fixed the amount of spaceand I didn't need to buy a copy of PartitionMagic.

My big tip for the month, albeit one I've probably given before:

• Do create a structure for all of your data, and backup that structure on a regular basis.

• Don't bother with separate partitions for your data and program; a base-level directory will do just fine and not constrict you later on.

• Don't simply accept the default directories offered by applications.

Mowrey, M. (2000), "Thank you, please come again," The Industry Standard, Vol. 3 No. 11, pp. 196-7.

It's hard to discuss ongoing change in technology, libraries, and media without talking about money and the marketplace. Some pieces are particularly revealing, when you're considering claims of the e-business revolution and wondering just why some e-stocks can go down even faster than they went up.

This article reviews a McKinsey & Co. study of the books of some e-commerce companies. Some of these numbers are the kind that give averages a bad name: for example, "repeat customer maintenance costs" for the e-commerce sites studied may have averaged $1,931 (an astonishing number), but the range was from$0.27 to $16,000. The base numbers seem pretty horrifying, given my caveat about averages. On average, sites spent$250 on marketing and advertising to attract each new customer (partly because only five percent of those visiting an e-commerce site actually buy anything). But two-thirds of the customers never buy anything else. Given that the average purchase from a new customer is $24.50, this is a losing proposition. But then, the average repeat customeraccording to this studyspent$52.50 per quarter or $110 per year. Thus, after 2.5 years, the cost of acquiring that customer has been covered (but not the cost of the ones who never returned). Despite my partial entry into capitalism, I can't understand the overall numbers. If it's costing$1,931 to maintain a site for those $110 in saleswell, I'm obviously missing something. The good news seems to be in one chart that relates costs to the age of an e-commerce company. New companies average four times revenues for marketing and another 2.7 times for operating costsbut companies more than two years old do much better. Marketing costs appear to be a mere 172% of revenues and operating costs a trivial 108% of revenues. In other words, once you get past the first couple of years, you can coast on the marvel of spending only$280 for each $100 you earn. Meanwhile, to be sure, physical businesses (stores, restaurants) mostly cease to be businesses if, after two years, they're still spending three times as much as they take in. Willmott, D. (2000), "Surfing on the boss's time," PC Magazine, Vol. 19 No. 9, p. 97. This quarter-page thriller begins "Be honest. When you're at work, do you use the Web for personal purposes when you're supposed to be doing your job? Of course you do....but have you considered the economic impact of your actions?" A research report claims that 35 million U.S. workers and "all of them do at least a little recreational surfing." Then it comes up with an astonishing figure for the productivity hit from all this cheating:$5.3 billion in 1999. Isn't that awful?

Look at the numbers a little more closely. You can restate this report as follows: "On average, every Internet-connected worker in the USA spent $151 worth of work time doing recreational surfing in 1999." Since most connected workers are reasonably well paid, that probably averages out to around five hours per year per person. That's almost six minutes a week lost to the employer: what will those lazy employees do next? Couple that with the notorious fact that American workers spend many more hours per year at work than workers in any other industrialized nation, and it may be just a bit less of a story. Rephrase it this way: "American workers are chained to their desks at least 100 hours more than workers in civilized nationsbut they spend one-twentieth of that time doing the recreational Internet surfing they're not home to do." Aren't numbers wonderful? Perspective Wrong, Wrong, Wrong Before you assume I'm being mean-spirited as usual, be aware of the target here: one Walt Crawford. It's time to catch up with a few more months of "Trailing Edge Notes" from 1995 and my crystal ball was pretty fuzzy during that period. What's the World Wide Web really all about? The new universal model for information distribution? Nonsense. An ideal for text presentationblack on medium gray? Perish the thought. A primary source of reliable information? Uh-uh. The Web is about self-publishing. Not entirely, and not originally, but that's how it's working out. Those were the lead sentences in a Perspective more than five years ago, in "Trailing Edge Notes" for May 1995. I'll stand by the first "nonsense," and we've moved past those awful gray browser backgroundsbut I was at least partly wrong on the last two points. The Web has become a major source of sometimes-reliable information, and many Web users have become fairly skilled at determining which sources are reliable. But then, that wasn't the point of the perspective. I was interested in the Web as a self-publishing medium, both its strengths and its weaknesses. The strengths continue: it's easy to prepare Web pages and increasingly easy to mount them. The weaknesses continue to be those of all self-publishing: lack of editing and filtering. One key sentence in that perspective was wrong, at least in terms of the perceived primary role of the World Wide Web. Today, that perceived primary role is commerce: the Web is about buying and selling. That's an oversimplification, to be sure. I don't know how you could prove this, but I'd guess that self-publishing (in the form of personal Web sites and similar informal sites) still makes up more Web sites than e-commerce, even if it gets a lot less attention. Fortunately, e-mail still represents the most widely used Internet function, with search engines coming in second, and it's fair to assume that many searches are for reasons other than pure commerce. May 1995 was a bad month for my track record in general. A long note suggested that Sony might introduce MD Data (a data version of the MiniDisc) by 1996 as a robust, low-cost 140MB storage medium. That never happened and MiniDisc as an audio medium went into obscurity within the US market (although it's coming back after years of reasonable overseas success). I called for ideas to improve LibraryHi Tech and Library Hi Tech News; the silence was deafening. Not that I was the only seer with cloudy vision. Adrian Mello wrote an editorial in Macworld asserting that Apple should have a 20 to 25 percent market share within five years and that Apple's customer loyalty was twice as strong as any other PC vendor's. The second claim was absolutely false, as Gateway's customer loyalty has been at or above 70 percent since before 1995 (and you can't have 140 percent customer loyalty), and passed Apple's figure later in the 1990s. The first assertion was just plain silly. Apple has returned to profitability and seems to have a solid nichebut that niche is roughly 5 percent of the PC market and hasn't risen above 10 percent in years. Mike Elgan wrote a Windows Magazine piece suggesting that "Internet hoopla" was like the CB craze of the 1970s. Instead, the CB craze was replicated within the Internet through AOL Instant Messaging, ICQ, and chat rooms galore. March 1995, PC/Computing: "The future is flat, if the latest crop of active-matrix LCD monitors for desktop PCs is any indication." What specific units did the magazine have in mind? A 640 x 480 9.4" display offering 12-bit color (4,096 colors) for a mere$2,495; a 10.4" display for $7,995and a 1280 x 1024 true-color 13" display for$11,755. LCD displays have improved and the prices have come down, but they're still too small and too expensive to be good CRT alternatives for most users.

Video by the Numbers

That was the title of a lengthy Perspective in September 1995. I was discussing the general failure of video CD in the USA, despite continuing projections that it would sweep away videocassettes any day now. I went on to discuss the two high-density discs then being proposed by Sony/Philips (on one hand) and Toshiba/Matsushita (on the other hand). The article noted the possibility of a Betamax-VHS war and also noted that "pundits assert that video CDs will totally dominate the marketplace within a year or two of their introduction." I doubted that, but did comment that "libraries should prefer the digital medium for the same reason that audio CDs make better circulating items than did LPs."

I had three major reasons to believe that high-density discs wouldn't sweep away VHS as rapidly as we were told. First, if there was a format war it would stall any adoption. Second, the installed base would make any move fairly slowjust as the move from LPs to CDs was nowhere near as fast as we like to remember. Third, and most important, "people aren't about to give up their videocassettes for the simple reason that they record, for timeshifting and for home video."

The rest of the article was about quality and numbers, noting that 640 x 480 true-color video without compression would require about 1.6 gigabytes per minute, and that the largest capacity proposed (for the two-sided version of the Toshiba/Matsushita disc) was 9GB to store 135 minutes of video. Something had to give: that is, a high rate of compression (23:1 using the best-case scenario). I noted that people didn't always pay for qualitySuper VHS has never achieved more than 5 percent of the VHS marketand that early compression tests showed substantial quality loss at 12:1 compression.

For readers paying attention, that September 1995 article was an early commentary on DVD, since that's what became of the Sony/Philips and Toshiba/ Matsushita efforts when other companies bullied them into combining forces. No, DVD didn't sweep away VHS in a year or two; it's becoming a mass medium but doing so at a leisurely pace. The compression ratio is much higher than I'd expectedbut they've also found ways to make the extreme compression more palatable. Well-mastered DVDs do indeed offer better video quality than any previous medium despite compression ratios ranging from 20:1 to 40:1 or even higher.

A plaintive perspective on that issue's back page suggested that "Trailing Edge Notes" readers should also be newspaper readers. "Local newspapers maintain a community's sense of community and help local merchants to surviveand daily newspapers help keep you socialized." Thoughtful commentators in 1995 concluded that "Daily Me" electronic newspapers should complement print papers, not replace them. For the minority of adult Americans who read and appreciate newspapers, that seems to be the case now and for the medium-term future. Morning newspapers continue to grow in circulation (albeit very slowly).

Perspective

Browser Accessories

Michael Miller poses a poignant question in his March 21, 2000 "Forward Thinking" (the editor's column in PC Magazine): "Remember when a plain old browser was enough?" He describes some of the hot new ways that you can shrink your active browser window. Third Voice adds a side window so that you can comment on the pages you seeand "adds links to the words on the page," mostly giving you even more opportunities to shop. uTOK, the Users' Tree of Knowledge (yes, that's small u, capital TOK: don't you love PC orthography?), opens another side window to share ideas about the page you're viewingand "users rate the people making comments...only those with high ratings survive." Are your comments good enough to make the grade?

There are more, too many to mention here. One of his favorites is Kenjin, which claims to "analyze the content of whatever you're viewing"not only Web pages but your own documentsand then "show you related data from the Web or from information on your hard disk."

Miller thinks these are all "interesting and useful tools" although he concedes that they may seem like overkill. But then, he assumes that the features will wind up in the browsers themselves (or in instant message products), because we all want everything in one toolbar.

I have mixed feelings about much of this. As Eureka service manager at RLG, I'm acutely aware that people have trouble when they're using a 640 x 480 screen and leave on Navigator's entire overhead. I'm not thrilled about the idea that bibliographic searches will pop up suggestions to buy, buy, buy or links to various "related" Web sites. Some of these notions are amusing and could be useful, but they seem aimed at further reducing the already short attention spans of Web surfers and other computer users. If I'm reading one document, do I really want a popup suggesting other places I should be?

I may show my age when I say that, in the end, much of this stuff just gives me the creeps. As always, your mileage may vary: you may think these items are the best developments since Alexa.

Postscript: The May 9, 2000 PC Magazine includes a "first look" at Kenjin, a free download that needs 20MB hard disk spaceor 200 to 500MB for the "personal" version. According to Edward Mendelson's review, "the idea is ingenious, but the execution is erratic, and the toolbar has an annoying habit of displaying an ad when you want a list of links or files." Surprise, surprise: free software pops up lots of ads. The claim is that Kenjin uses "Bayesian inference logic to interpret the context of the target" (a Web page, word processing document, spreadsheet, or text that you've dragged to the Kenjin toolbar) "and zero in on appropriate links." PC's tests found the results much less precise than search engine results. When they opened the browser to the Society of Neuroscience, the first link was the page currently being browsed (now that's useful!); the second was the Society for Historical Archaeology. Why not? They both start with "Society," don't they?

Need I mention that Kenjin's servers will have a complete record of not only your Web browsing but also your writing? If that doesn't bother you, think about it a little longer. Unless, of course, you're a McNealyite ("You have no privacy. Get over it.").

A Keyboard for the Palm

This one actually makes sense, if you're one of those who relies on your Palm device. The Palm Portable Keyboard costs $100, folds up into a case roughly the size of a Palm III, and unfolds into a full-size keyboard. It takes power from the Palm's batteries; you dock the Palm on the keyboard. According to PC Magazine, the 8.2oz. keyboard works well although it doesn't feel quite as sturdy as a typical desktop keyboard. Wretched Excess Strikes Again What can you buy for a mere$350 these days? Not much, when you think about it. Perhaps a Palm V and accompanying keyboard, a high-end inkjet printer or a good laser printer, a 16"-viewable flat-screen Trinitron display, or a 40GB hard disk. Or, for those who believe there's life beyond computers, a 27" TV, a few dozen mass-market paperbacks, a case of excellent wine, a few good restaurant meals, or a significant tax deduction and good feeling when donated to The Nature Conservancy or your own favorite cause.

Or you could buy the Philips Pronto, a palm-size device "designed to replace that pile of remote controls on your coffee table." It's not just a universal remote (which you can buy for $50-$75): it has a touch-sensitive backlit screen! And you can log on to Web sites that "contain libraries of icons and command codes" for all sorts of devices. A surprisingly enthusiastic First Look in the May 9, 2000 PC Magazine gushes that "you're not just buying an advanced, programmable remote control; you're also buying into an online support community." Right.

Review Watch

Desktop Computers

Morris, J. (2000), "1 Gig gets real," PC Magazine, Vol. 19 No. 8, pp. 30-43.

The original IBM PC ran at 4.77MHz. My first PC (a CP/M system) ran at 4MHz. That was roughly fifteen years ago. Now we're at a gig: 1GHz (1,000MHz) PCs are on the marketand at reasonable prices. This review includes three 1GHz desktops, all from big-name vendors, together with a handful of systems that run at a mere 850MHz. The article also includes careful discussion of how AMD's Athlon and Intel's Pentium III compare and why one might do better than the other in certain benchmarks. That's a little exotic for most of us. The main message: these are seriously fast machines. You pay a premium for that superfast CPU, but that means that the premium for 700MHz CPUs is disappearing fast.

There's no Editors' Choice (rarely awarded in First Looks roundups), but only Dell's Dimension XPS B886r and B1000r garner five-dot ratings. These systems aren't cheap ($3,948 and$4,299 respectively) but they are loaded with all the extras and perform extremely well. Gateway sells a similarly configured Select 1000 for $3,310, but it doesn't run 3D graphics quite as rapidly (and some buyers will prefer Dell's Pentium III to Gateway's Athlon). Somers, A. (2000), "The 800Mhz power elite," Computer Shopper, May, pp. 150-8. This review includes six fast, well-equipped systems using either AMD's Athlon-800 or Intel's Pentium III-800. The minimum configuration specified is 128MB RAM, 13GB hard disk, graphics card with 16MB RAM, DVD-ROM drive, 18" display, and V.90 modem. The Best Buy goes to Systemax'$2,549 Excite PJM-800A, an Athlon system. It's the cheapest in the roundup but includes a CD-RW drive, a five-speaker Altec Lansing surround-sound system, a 30GB hard disk, and a graphics card using today's best graphics processor (the nVidia GeForce 256) and 32MB RAM.

Also marked "noteworthy" is the $3,457 ABS Multimedia System 1, with 256MB RAM, a 34GB IBM hard disk, the same graphics card as the Systemax, a different five-speaker surround system, and both a CD-RW drive and an Internet camera. The most powerful systems were from Dell and Gateway, but Computer Shopper weights price and configuration more heavily than benchmark performance, apparently. Digital Cameras Derfler, F. (2000), "Broadband," PC Magazine, Vol. 19 No. 8, pp. 164-77. This article isn't just an equipment roundup. It's really an overview of how a fast Internet connection can "reshape the way you work," including quick comments in a number of areas. A significant portion of the article is devoted to digital video cameras, emphasizing their use as Webcams. I was surprised to see an even dozen models in the$40-$150 price range, from tiny names like iRez and known quantities such as Logitech and Kodak. Editors' Choices go to Ezonics'$150 EZDual Cam and Intel's $130 PC Camera Pro Pack. The EZDual can serve as a Webcam, but it can also be used as a portable, taking low-resolution snapshots. Intel's unit comes with a long cable for flexible use and a strong bundle of software, while providing good image quality and accepting composite video as inputso that you can feed camcorder footage into your computer. Newman, H. (2000), "Can-do camcorders," FamilyPC, Vol. 7 No. 4, pp. 110-11. Ready to produce your own videos? The four cameras in this roundup cost$1,200 or less and record digitally (but on tape), so that you can edit the videos and add transitions on your computerwithout the loss of video quality caused by each generation of normal video editing. Typically, decent editing software comes with the camera. You will need quite a bit of storage space on your PC, and unless you have a FireWire (IEEE 1394) interface data transfer will be fairly slow.

The highest-rated camcorder in this roundup also has the lowest resolution: Sony's $1,099 HandyCam DCR-TRV310. It gains points for offering the best optical zoom (20x), the "NightShot" feature that lets you film at night using infrared sensors, and recording digitally on regular 8mm tapes rather than Mini-DV tapes. It's bigger than the competitors and a little heavier, but still only about two pounds. Highest rated of the MiniDV camcorders is Panasonic's$1,099 Digital Palmcorder PV-DV910, which also limits resolution to 460K (the other two camcorders capture more than 600K pixelsbut regular TV won't even display 460K). The Panasonic is "the size of a thick paperback" and weighs less than 24 ounces; it's also simple to use and offers very good image stabilization circuitry.

Home digital video was supposed to be the "killer app" that made FireWire standard and caused an enormous increase in Apple's market share, while making Sony a contender for US desktops. But how many people actually edit their home videos? So far, apparently not enough to make a major market. Some libraries may have uses for digital camcorders, however, and the prices are getting plausible.

Internet Appliances

Brown, B., and Brown, M. (2000), "The Internet in every room," PC Magazine, Vol. 19 No. 8, pp. 210-22.

What's an Internet appliance? Based on this review, it can be almost anything that isn't a PC but connects (somehow) to the Internet. The six production units reviewed range from InfoGear's iPhone and Netpliance's i-opener through WebTV to the Sega Dreamcast game unit.

It's a fascinating story and it's hard to fault the Editors' Choice box: "Not quite there yet." The i-opener seems to be closest, but I think there are reasonable questions about its business model (how are they selling 10" LCD screens and all the surrounding circuitry for $199 and hoping to make money, except by restricting you to a slightly overpriced and probably under-featured ISP?) and it won't handle most Web multimedia features. WebTV units may be OK for occasional surfing, if you can cope with the coarseness of TV images. Or you could buy a PC. Projectors Larson, B. (2000), "Lightweight projectors," Macworld, May, pp. 34-5. This absurdly brief roundup covers seven "lightweight" digital projectors, where light has been interpreted as less than ten pounds. You won't find Compaq or Sony units here; what you will find are medium-priced units that cost more and aren't as well equipped as full-size LCD projectors. Prices range from$4,999 (Proxima's UltraLight L51) to $8,995 (Sharp's Notevision7). The highest mouse rating (4.5) goes to Epson's PowerLite 710c, a relatively light unit (5.8 pounds) and the brightest of the bunch at 1,000 lumensbut also the second most expensive, at$7,999. There's an oddity that a longer review might have explained: the features table shows "native resolution" in one column and "maximum resolution" in anotherbut for most LCD and DLP units, the native resolution is the maximum displayable resolution. I'd hoped we were past the point where manufacturers could claim a maximum resolution that units were physically incapable of displaying (except by scaling), but maybe not.

Removable Storage Devices

Lunsford, K. (2000), "Hold everything," Macworld, May, pp. 85-9.

Ready for DVD-RAM? This article includes a roundup of six early DVD-RAM drives suitable for the Mac, although most of the article is about DVD-RAM and the Mac. DVD-RAM currently stores 2.7GB on one-sided discs or 5.4GB on two-sided discs; 4.7GB one-sided discs (the same capacity as single-layer DVD-ROM) may arrive by the time you read this.

According to this writeup, DVD-RAM is much more useful than CD-RW because of the way it works. The article claims that the only way to replace anything on a CD-RW disc is to rewrite the entire disc. That isn't the way I've heard it, but I can't (yet) speak from personal experience. DVD-RAM, at least with these drives, basically works like a fat diskette or very slow hard disk.

The article draws an appealing case for DVD-RAM's cost-effectiveness by comparing it to 250MB Zip disks or 2GB Jaz disks. In that comparison, DVD-RAM looks pretty good: about $30 for a 2.6GB DVD-RAM as compared to$15 for a 250MB Zip disk and $100 for a 2GB Jaz disk. Compared to CD-RW, of course, it isn't so hot:$30 will buy at least ten CD-RW discs offering 6.5GB storage.

As usual for Mac peripheral roundups, most of the names here are companies PC owners will never have heard of, such as LaCie, ProMax, and the Editors' Choice, APS Tech's $500 DVD-RAM External SCSI. Speaker Systems Labriola, D. (2000), "Sound choices," PC Magazine, Vol. 19 No. 8, pp. 254-6. The good news is that decent-quality three-part PC speaker systems have come down in price: most of the systems reviewed here cost$100 or less. Unfortunately, the best ones are still expensive: the Editors' Choice goes to the Cambridge SoundWorks DeskTop Theater 5.1 DTT2500 Digital, $300 (but you should get a discount for remembering the whole name!). It's a six-speaker system with separate control box, designed for full Dolby 5.1 surround. Also unfortunately (if all too typical for PC magazines), PC Magazine doesn't seem to feel that speakers deserve the kind of analysis and objective measurement lavished on most hardware and software. There are no measurements of any sort, and you get the manufacturer's exaggerated power claims. If you're interested, read the article: several other systems deserve consideration, particularly if you're unlikely to mount five speakers around your home office. Utility Software Gottesman, B. (2000), "The ultimate utility guide," PC Magazine, Vol. 19 No. 9, pp. 154-234. While Windows 98 provides some of the utility features that previously required separate programs, it doesn't do everythingand separate utilities frequently do it better. This roundup breaks down into eight sections: suites, retail antivirus, corporate antivirus, e-mail filters, Internet hard drive, disk utilities, desktop migration, and "toolboxes"a catch-all for miscellaneous utilities. Most of you don't need products from every category, but every computer user should have a few utilitiesand the more varied your computer uses, the more you're likely to need some of these. As always, this is one to read in its entirety. That may be more true this year, since the functions incorporated within suites generally don't have their own sections (except that every suite includes a decent retail antivirus program). There are now three strong utility suite contendersand this year, Ontrack's SystemSuite 2000 gets the Editors' Choice. While it incorporates some code from the Mijenix suite that it replaces, it's the most integrated suiteand it finally includes antivirus software. I love PowerDesk, which comes with SystemSuite 2000but you can get the basic version of PowerDesk 4 as a separate free product. Norton SystemWorks 2000 scores second best and gets an honorable mention; Norton's CleanSweep uninstaller, Norton AntiVirus, Norton Utilities disk maintenance and defragger, and Registry cleaner are all top-notch. SystemSuite's integration is tighter, it's easier to use, the system load is lower, and it's the only one of these suites that works on Windows 2000 and Windows NT. McAfee Office 2000 is a distant third. I believe that most users should have either SystemSuite or SystemWorks. Do you need retail antivirus software? Not if you have SystemWorks, since Norton AntiVirus 2000 is part of itand is also Editors' Choice for the separate category. There's only$19 difference between the two products: why not buy the suite? Guess who gets Editors' Choice for corporate antivirus software: Norton AntiVirus Enterprise Solution, $3,100 for 100 users. Good e-mail filtering isn't cheap, and these are strictly corporate solutions. The Editors' Choice, CommandView Message Inspector, costs$10,995 for 500 users (sitting between your router and firewall on one side and your PCs on the other side).

What's an Internet hard drive? A service that lets you store files at a remote site, accessible over the Web. Some of these services offer small amounts of free disk space (and you get to watch ads while you're reaching your files); others offer a little more space at a price. One Editors' Choice, Connected Online Backup, claims to offer "unlimited" space for personal systems at $14.95 a month; it comes with background backup software. Another award goes to X:drive, which provides 25MB of free space. I don't quite understand these services except for mobile professionals, but that may be my problem. A year or two ago, "disk utilities" referred to Norton Utilities and its competitors. This time around, the reviewed products offer rollback capabilities, partitioning, or imaging. Editors' Choice for rollback software (to restore the system's state as it was before a software crash) is GoBack 2.1, a$50 product from Wild File Inc. that's been consistently well reviewed. PartitionMagic continues to be the partitioning software of choice, while Norton Ghost 2000 Personal Edition is the best imaging software. But then, Norton Ghost ($70 by itself) also comes with Norton SystemWorks. What's desktop migration? That's the problem you face when you move to a new PC: keeping all of the choices that make the PC yours. None of the products in this section earn an Editors' Choice. Web Information Managers King, N. (2000), "Harness the Web," PC Magazine, Vol. 19 No. 6, pp. 127-36. Web harvesters have been around for a couple of years, offering varied approaches to saving and organizing stuff you find on the Web. It's an odd software category that can certainly be useful for some users. This review covers six productsand notes that none of them can handle all eight options that PC Magazine identifies as useful. That list of options may suggest how complex this category is: capturing full pages, full pages but without graphics, highlighted sections of pages, selected graphics, selected frames, an entire site, a partial site (but more than a page), and the URL only. Two of the products become icons on Internet Explorer's toolbar, including the Editors' Choice, iHarvest One (www.iharvest.com). It's free, but it also offers the "most well-rounded set of features for collecting and organizing information from the Web." Honorable mentions go to askSam's SurfSaver 2.0 (also free), particularly well suited for users grabbing large quantities of text, and Webforia Organizer & Reporter ($100). If you love Navigator, go for Webforia: the others require Internet Explorer.

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. CD-ROMs and DVD-ROMs for review should be sent to Walt Crawford, c/o RLG, 1200 Villa Street, Mountain View, CA 94041-1100; Windows only. Visit my Web site: http://home.att.net/~walt.crawford.