Tom Halfhill’s column in Maximum PC is usually one of the more informed pieces of technical journalism I get to read in any given month. This month is an exception. His topic is the ever-popular USB 2.0 vs. FireWire (IEEE-1394) debate, and he comes down firmly on the USB 2.0 side on the basis of compatibility. What’s disappointing is the points that he leaves out. Here’s an excerpt:

[USB 2.0] has special features to support slow and fast devices without sacrificing performance…
…It reclocks the slower-data traffic to avoid bogging down faster devices…

Well, how about not having slower devices in the first place, which is the case with FireWire? The very next statement is even more egregious:

…and it guarantees the delivery of time-sensitive audio/video data – a feature called isochronicity.

True enough, but misleading. As written, in the context of a comparison between two technologies, this could easily be interpreted as an advantage of USB 2.0 vis a vis FireWire…except that FireWire has had isochronous transfers for years, and practically no products are out yet that use the feature in USB 2.0 to prove that it works.

Nowhere does Halfhill mention that FireWire provides symmetric connections between devices and/or computers acting as peers, while USB requires a computer as a “master” polling “slave” peripherals. This shortcoming of USB is supposed to be addressed in USB On-The-Go, but that’s even newer and less supported than USB 2.0. In addition, it’s a nasty choose-the-master kludge rather than a truly symmetric protocol version. Absent is any mention that USB 2.0′s supposed 480Mbps to 400Mpbs speed advantage over FireWire is bogus because of additional protocol overhead, or that 800Mbps 1394b products are practically upon us with speed upgrades up to 3.2Gbps likely to follow rapidly. Strangely missing is any comment that USB 2.0 is an Intel baby, designed from the ground up to benefit the chip monopolist, while FireWire/1394 is an open IEEE standard supported by many vendors.

Halfhill’s prediction of success USB 2.0 is not unreasonable. USB 2.0 does have advantages as part of the USB “brand”. What’s disheartening is that Halfhill doesn’t even mention the technical issues that consistently favor 1394. I for one expect better technical perspective from him; I can get the marketing viewpoint from Intel’s PR department.