Target: You

I hate ads. I mute the commercials when I watch TV and use software to block ads on the Web pages I browse. (Don’t tell my employer, whose revenue comes largely from ad dollars.)

The problem with advertising is that it’s a broadcast game — advertisers want to reach the widest area possible. I’m a narrowcast person — there are certain things that interest me, and many more I either couldn’t care less about or that actively repel me. If someone could come up with a way to show me ads only for electronic gadgets, peanut-based foods, and businesses in my neighborhood, I’d be inclined to hear them out.

Well, someone may be on the right track. The iCAP Forum, a group of about 70 vendors, is promoting the Internet Content Adaptation Protocol (iCAP) that aims to give ISPs the power to modify content dynamically, which among other things would allow them to run locally targeted advertising.

ICAP lets a cache server send content to an application server for processing before forwarding it to the user who requested it. The application server can pass or modify the content before it returns it to the cache. One of the ways it can modify the content is to check it for viruses and strip the virus code out if any is found. Because the protocol is based on HTTP, it wouldn’t have helped last week’s “Anna Kournikova” virus outbreak, which was spread by e-mail, but it could filter out malicious JavaScript. And of course, if it can screen out viruses, it can also screen out content that doesn’t meet your corporate acceptable-use policy.

If you’re an ISP, the attraction of iCAP is obvious. Advertisers should be willing to pay more for targeted ads. If you’re a corporate IT shop, that benefit won’t excite you much.

But here’s one that will. Suppose the iCAP application server recognized whether your user was working on a PC or a wireless device like a PDA or cell phone and could server the appropriate code for each kind of user. That would eliminate the need to maintain redundant URLs for equivalent content.

What if you have multinational clientele? Let your iCAP application server decide the user’s native language, and serve up the appropriate page.

The possibilities are exciting. The iCAP Forum last month released version 1.3 of its iCAP specification and submitted it to the IETF’s Web Replication and Caching working group for approval as a standard. Some Forum members aren’t waiting for the completion of the standards process — they’re delivering pre-standard iCAP-enabled products already.

Many organizations are just beginning to explore the benefits of adding caching servers to their Web infrastructure. If you’re among them, be sure to ask your caching vendor whether its product supports iCAP.

Picture imperfect

Last week, PC Magazine, one of our partners, published a review of Polycom’s ViaVideo videoconferencing device. The reviewers gave the product three stars out of five — an average rating.

I too have been playing with ViaVideo, but my opinions differ from those of my colleagues. I can’t recommend this product, because its flawed interface and high price make it a poor value for organizations seeking to establish a foothold in desktop videoconferencing.

Don’t blame Polycom’s hardware, though. ViaVideo includes a multimedia processor in the camera housing that offloads image processing from the host’s CPU. I was pleasantly surprised at the quality of video images, even when my calling partner was moving around.

The software, however, is another story. The application uses a non-standard Macromedia Flash-based graphical interface that takes a bit of getting used to when configuring the product. Worse, though, is the way it presents directories of potential conferencing partners. ViaVideo uses Microsoft Internet Locator Servers to find compatible users. Unlike Microsoft’s free NetMeeting, however, you can see only limited information from each directory entry in a tiny window on the screen.

You’d think you could get around this limitation by using the hardware with a different videoconferencing package, but you can’t. Every time you launch ViaVideo it uploads software to the camera that makes it work. The hardware and software are tightly coupled.

Installation is simple — you don’t have to open up your PC to install an add-in card. The unit connects to a USB port on your computer, the line-in port of your sound card, and an electrical outlet. You must install the driver software before you hook up the hardware. The camera, with built-in microphone, sits loosely, one might say precariously, on a plastic base. I set mine up on top of my monitor; if you have a laptop, the company will sell you a mounting kit.

Polycom provides only a razor-thin Getting Started manual with the product, but more documentation is available on the installation CD-ROM and the company’s Web site.

At a street price of about $550, ViaVideo is a poor value. If you need desktop videoconferencing, I suggest using a less expensive videoconferencing device such as the 3Com HomeConnect PC Digital WebCam (about $115 street) and a PC headset ($40 street) with NetMeeting.

Wireless service world out of whack

Last week, we reported that Metricom, purveyor of the high-speed Ricochet service, was hurting for cash, while OmniSky, the wireless handheld modem maker, was flying high.

You can attribute part of this difference in fortunes to service startup costs. Metricom is building its infrastructure from scratch in 25 cities, which requires a huge capital outlay. OmniSky instead partners with ISPs to provide services to its 34,000 subscribers. OmniSky’s service has been available longer, and any business can lower its costs once it has recouped some of its initial investment from the early adopters. OmniSky’s still not turning a profit, mind you, and doesn’t expect to for almost two years.

The question of Metricom’s financial health aside, Ricochet is by far the better service. Who wouldn’t choose 128Kbps access, if they could get it, over OmniSky’s measly 19.2Kbps? That higher level of connectivity would let me use my regular notebook PC for all my Internet browsing and e-mail needs wherever I went — a far more attractive option than carrying a separate mobile device that I have to synchronize with my primary system. By contrast, OmniSky’s service, which runs at the very fastest modem data rate available in 1991, all but requires you to look at a very condensed version of your information, and thus is suitable just for the Palm, Handspring Visor, and HP Jornada 540 handheld devices with which it runs.

It just seems wrong. Using OmniSky and handhelds to access Internet information seems like an interim solution, just a step above the virtually unusable WAP-based Net-enabled phones. So why is the OmniSky service doing so well?

In reality, it’s doing well only in relative terms. Thirty-four thousand subscribers is not exactly a groundswell of enthusiasm. It’s lower than the number of support calls a company like Dell gets in a single day. The fact is, slow speed and a tiny screen fail to make a compelling case for connectivity for any but the most compulsively connected.

Personally, I’m looking forward to trying out Ricochet. I’m hoping an angel investor that wants to get in on the ground floor buys in to the company now, when it really needs the cash. Barring that, perhaps someone will pick up the infrastructure intact if the company liquidates. But that’s premature — the company is proceeding with its plans for now. Don’t let the shadow of doom keep you from testing out the service if it’s available in your area. Just don’t pay up-front for any long-term contracts.