Features

iPaq scandal reaches boiling point with user petition launched

by Guy Kewney | posted on 25 July 2002


Compaq's "arrogance" has been attacked, in public, by a body of users. Can HP really carry on with its plan to kill the Jornada name and stick with the tainted iPaq brand? - Analysis ...

Guy Kewney

I recently had to go to bat for a user of the iPaq. He tells me he's spent the last two years trying to get Dixons and Compaq to admit that his model has a duff circuit; it only works when absolutely fully charged.

In the course of his two year battle, he's been persuaded by tech support that the machine isn't working because he has an out of date PC, and has been talked into buying a new one. He's been talked into buying new software, downloading "fixes" which don't fix, and generally, being miserable about lost data, lost business caused by failures, and wasted time.

I got onto a senior Dixons executive working in quality control, and gave him a hard time. He seemed surprised I was bothering. "The iPaq, you know what the problem with it is, surely? It doesn't work."

The user petition provides "a comprehensive list of the issues that these devices suffer from. A quick read of this list should be evidence enough that it is time HP/Compaq admits that these devices are defective and takes bold steps to fix them," it says.

It's a list which the trade has known about for ages, and has kept quiet about. The problems are hard to understand, because the iPaq has been so popular - so popular that when HP and Compaq merged, executives agreed that the iPaq brand was, obviously, the one to keep going.

Our analysis suggests that there is a perfectly simple explanation for the iPaq's success. It has been a wonderful prototyping tool for the Pocket PC development community, because it is possible to plug a full-size PCMCIA standard PC Card device into it.

Developers and pioneers don't mind that. But real users of a Pocket PC who want a wireless LAN, don't go for an iPaq because to use an iPaq with its huge, heavy PCMCIA jacket, you have to be dedicated. Real users aren't; and they have gone for the Jornada, with a CompactFlash wireless card - neat, small, and much less power-greedy.

The iPaq has been kept alive by one thing: its ability to connect to Bluetooth devices, with a built-in Bluetooth wireless. Again, this arrived in time to help people do some prototyping work at a time when most Pocket PC size devices - PDAs and the like - don't have Bluetooth, nor any way of getting it attached.

The truth appears to be filtering through to Hewlett-Packard at last.

Yesterday's announcement of the new Jornada 928 punctures the complacency of the pro-iPaq lobby. It wasn't going to happen, officially; when the merger was finalised, HP/Compaq executives were very clear about one fact - that the Jornada line would be ending.

Officially, the story about why the 928 has appeared is something like this: "We needed to have a Pocket PC Phone Edition, and the Jornada design was ready, and the iPaq design wasn't ready."

That's part of a disaster-story, verging on the farcical. As near as we can tell, what really happened is that Compaq was working with HTC on a "Wallaby" project which would have been the first iPaq Phone Edition of the Pocket PC. When HTC finished the design, Compaq didn't like it.

There were some good reasons why Compaq didn't like it. It didn't have Bluetooth for example. There are reasons why Bluetooth can't be integrated into the current generation of Phone Edition devices - and they are all perfectly understandable - but more of that in a minute. The thing that matters is that HTC found itself with an expensive design, and nobody to pay for the costs of developing it.

So they approached NEC (sources say) with the Phone Edition box they'd designed, and NEC said yes, they'd buy half the machines. And NEC showed it to someone in mmo2.

Who this was, I can't find out. But whoever it was, understood that mmo2 needs to make money from mobile data. It was almost certainly the same person who set up mmo2's deal with Handspring for the Treo, and who ordered lots and lots of them (over 100,000 say some sources) and with BlackBerry - there, the order was for 165,000. And they did a deal with HTC, ordering another 100,000 machines, which have been branded XDA.

There is simply no way the UK market can absorb that many phone edition PDAs. Even if they were simply nice phones, what mmo2 needed was 100,000 users all spending fifty pounds a month on GPRS data. There simply aren't that many rich idiots around.

It isn't even clear that this amount of traffic would make mmo2 the money it needed - the spectrum it would swallow up would be enormous, and certainly, would require more base stations and mini-masts than mmo2 can afford to install in the next year. But it doesn't matter; the figures have been re-assessed, and, sources say, reality has penetrated. The 100,000 XDA (remember, this is the ex-Wallaby) order has been cut to 20,000, and BT has bought the rest. And the Blackberry order has been similarly cut back - this probably explains why RIM's share price has collapsed, since it isn't in a position to sue mmo2.

While all this was going on, the battle inside HP was raging aobut the design of the 928. Some wanted to have Bluetooth in it. But at the time HP was designing the 928 - about 12 months ago, say sources - the Bluetooth stack was not stable.

"We can't create an embedded system with an unstable stack," said one of HP's Grenoble designers. "We really wanted to have Bluetooth, but the only way we could do that, was to have it in a CompactFlash plug-in card. And we couldn't do that."

The reason they couldn't do that, was explained by HP's pre-sales boss, Christian Chaffard. "The problem is that the phone needs a hardware link for voice to the CF card. This pin isn't provided in a standard CF interface. So we were trying to make the connection over software, using the Pocket PC OS - and when we did the figures, we found that we couldn't guarantee the latency timings."

To qualify for approval as a GSM phone, the device has to guarantee not to introduce more than about half a second of delay to the voice. As anybody will know if they've made a cellular call to a friend in the same room, the network, compression, decompression, and error correction, all add up to about a third of a second in good conditions; adding any software-induced delay to that, means that the device simply isn't a GSM phone.

At Compaq, putting Bluetooth into the latest iPaq was less of a concern, because - as you can see from the petition - Compaq simply didn't seem to care.

HP staff say they are appalled at what they are now finding about Compaq's attitude to customer concerns. One executive, asking not to be quoted by name, said: "We don't do things like that in HP. If a product has faults, we withdraw it, we replace it, we fix it, we produce a new rollout. It's terribly expensive! - so we don't launch a product till it's ready. But Compaq was just producing faulty designs, and replacing them with untested updates."

Faults in the iPaq range are so many, it's not always possible to be sure which of them is actually causing any particular malfunction. One of the worst problems has been the power supply, say resellers, with most returns due to simple failure. But there have been so many revisions and so many updates, that it's never really feasible to diagnose what is going wrong with any customer at any time, they say.

And, they add, Compaq simply wouldn't replace faulty ones. They would promise, say resellers, but when it came to acting, they just didn't.

Today, we have the new Jornada 928 announced. It is a Jornada - not just in name, but designed by HP. And the next edition will come out in Spring 2003, says Chaffard - and that will be branded Jornada too, and not iPaq.

HP is starting to understand that its original excitement over the iPaq brand was a hasty judgement, and wrong. Whether it's ready to acknowledge that it has to kill the iPaq yet, and go with Jornada and the Jornada tradition, remains to be seen.