Daily Archives: August 19, 2008

Instant-on OS: a bad solution to a legitimate problem

Laptop manufacturers have recently started marketing a new technology which will soon be available to costumers (if not already available). There are a number of different approaches but the general idea is to provide the user with a BIOS which contains an OS and applications tailored for speedy access. So the user can, instead of booting into Windows or Linux as usual, boot into this “instant-on” OS (which is usually Linux). The advantage is that this “instant-on” OS boots in a few seconds which is quite fast compared to a regular Windows or Linux boot. Depending on the specifics of the technology used there can be other advantages like extended battery life, etc. I think the problem is real and legitimate: if a user just wants to check something quickly on the web or read email, booting up can introduce a significant delay. However, I think the solution of providing a customized OS in the BIOS is a bad one. I really wish the engineers would have spent time making the boot process faster for regular OSes. At this moment, I’m convinced that the “instant-on” OS feature is going to be a failure. Here’s a list of reasons in no particular order:

1. Boot times can already be mitigated by avoiding the boot process in the first place. Put the laptop to sleep instead of turning it off. Sure, the idea of the “instant-on” OS is to turn off the laptop completely and rely on the ability to turn on the laptop instantly. So in theory the user would keep the laptop off more often. So this “instant-on” OS should in theory be more environmentally friendly. But are the trade-offs of using the “instant-on” OS be worth the savings?

2. Missing features in the “instant-on” OS are going to be an irritant. Browse, browse, browse and bang, you can’t read a web page properly because the browser is not up to date or lacks a plugin? Or if someone sends an email with an important attachment which can only be read with software not in that “instant-on” environment? Users will say “screw it” and boot into their regular OS right off the bat rather than wait and see whether they’ll hit a brick wall again.

3. This is yet another environment to babysit. Yes, users of smartphones already deal with this. I myself have a Treo. The bookmarks on my Treo are not the bookmarks I have in Firefox in Linux. I’ve had to configure email access again on my Treo and then I stopped using it because the email software was just too old.

Someone somewhere will scream “but you can upgrade or install another email reader, you idiot!” Sure, but that’s my point: this “instant-on” OS thingy represents yet another environment which must be managed. Is it really worth managing that thing besides my regular OS and my smartphone? Users of smartphones will turn to their smartphones rather than their laptop with “instant-on” OS when they want “instant-on” capabilities with all the downsides which come with it. People who do not already have smartphones will have to make the mental shift towards managing multiple environments. For most people, we’re talking about dealing with a different OS than the one they are used to: Linux is not Windows.

Of course, some of this babysitting can be eliminated by relying on web-based applications. For instance, a user who has Gmail as their only email application does not have to reconfigure email readers everywhere. Bookmarks can also be managed online. But there is always a minimum of configuration which must be replicated across environments.

4. How secure is this feature going to be? Are updates addressing security issues going to be released frequently? How is my data going to be stored? It is probably not going to be encrypted. Is this going to be the low-hanging fruit for people looking for sensitive data on laptops?

5. What about added costs? Nothing is free.

6. A laptop is not the best vehicle for “instant-on” computing. People need computing devices which will fit into a pocket and that can be used instantly. They understand that such small size comes at a cost. There is a monetary cost. There is also a technological cost because these devices are not full-fledged computers. People also need portable full-fledged computers. But the “instant-on” capability on laptops fulfills neither needs. The laptop is too big to fit into a pocket so it does not replace the need for pocket-size computing devices. On the other hand, the “instant-on” OS does not give access to all the capabilities of the laptop. That is, a laptop with a regular OS already fulfills the need of having a portable full-fledged computer. The addition of an “instant-on” OS does nothing towards fulfilling this need. What I mean to say here is that it is unlikely that customers will think that such a feature is a must.

I need to reiterate here that I think the problem is legitimate and should be solved: it would be desirable to have laptops be able to go from being off to full functionality in a matter of 2-3 seconds. However, engineers should aim to achieve this by booting the regular OS users want rather than a BIOS-embedded special environment.