According to Linus, an operating system should be transparent to users. It ultimately shouldn’t be something that ordinary users have to care about. His view is that much of the hype is around visual elements and effects, not the genuine operating system.
This comment is consistent with Torvald’s own creation, the well-known free and open source operating system, Linux. Although available in a myriad of different packaged distributions – or “distros” – Linux itself is purely the kernel, or heart, of each distribution. The bulk of the software making up each disc is application programs, or developer tools or graphics manipulation apps or productivity suites and more – including, not least of all, a graphical user interface. Yet, none of these ancillary packages is Linux themselves, and while it is true an operating system that has no software is not terribly interesting, the loss of any of these items would not diminish the functionality of the Linux kernel itself.
Indeed, it is easy to see why Linus is disparaging about flashy visual effects being considered an OS feature, given Linux offers at least two major contemporary graphical interfaces (GNOME and KDE). The operating system design does not commit itself to any specific GUI and nor are users locked into these two. There’s nothing in the design of Linux which would prohibit an entirely new window manager being constructed should any person have the inclination and ability.
Torvald’s comments are therefore unsurprising and are consistent with his own work. However, what he goes on to say is definitely insightful.
Microsoft and Apple, he postulates, see the operating system as an important way to control the entire user environment both from a marketing and money-making standpoint. To his mind, proprietary operating systems force people to upgrade their software and hardware or they cannot take full advantage of them.
Linus returns to this theme of control when asked about the future of Linux with mobile devices. A lot of hardware in mobile devices, he says, is being produced in companies in Taiwan where the specifications and software have been pre-designed outside. The Chinese producers are restricted in terms of the hardware capabilities they create because they are working to such a specification. This specification isn’t necessarily any proprietary interface but may simply be the limits that Windows or MacOS can cater for: unless the hardware vendor is going to release their own software to manage the hardware, they must design for pre-existing and known targets, namely proprietary operating systems.
Read on to take control.
If a manufacturer wants to innovate then they must therefore also make a commitment to providing a software stack with their hardware, Torvalds argument runs.
In this case, ASUS has been churning out hardware for many years but decided to take control and implement software to run on top of their hardware. It’s at this point, Torvalds argues, that open source software is really useful. Software, he says, is really expensive to produce and it takes years. A hardware company can’t really afford to make such an investment which leaves fundamentally two options. Either the company must continue to be controlled from the outside influences, specifications and market – or they can take a pre-existing software stack that they may legally and technically change to suit their own purposes.
Such a software stack is open source. By building on top of open source software manufacturers can provide usable code in a much tighter time frame. And, so long as certain criteria are met (usually to make their own source code available in some form) the vendors have no legal burden which prevents them doing this under most open source licenses.
The Eee is but one example; another example is the various satellite navigation systems provided by TomTom, Navman, Garmin and others. Although Pocket PC, PDA and mobile phone variants exist, many of these SatNav devices are self-contained units which do not require any other device or facility in order to run; they just work out of the box. The thanks must go to Linux which commonly is used as the underlying operating system for these devices, along with some tweaks to use the unit’s specialised hardware, such as its global position unit.
These hardware vendors are therefore saved a huge burden; they can concentrate on their area of expertise – namely global positioning and satellite navigation – without needing to be occupied with the work involved in producing an embedded OS. Even should they have the desire, part of Torvalds argument is that software takes a long time to write: it would be grossly unpalatable to a vendor to delay the release of their gadgetry in order to develop their own built-in OS, not to mention the potential loss of getting the jump on their competition.
By exploiting Linux, these companies dramatically shorten the time to market because a tried and tested OS is already available. Additionally, the cost to the end user is reduced because there is neither extensive development time nor the cost of a proprietary OS like Windows Mobile being passed on.
Torvalds argument stands to reason, and it offers exciting possibilities for the future of computing. ASUS has had a runaway success with the Eee PC and this is evidenced by the demand exceeding supply and stores running out of stock.
What does this mean for the future?
ASUS themselves have spoken about how the Eee has been their most successful product to date, with more variations to come.
This is where open source really strikes; after all, it’s possible for a vendor to make a piece of specialised hardware but to provide even rudimentary networking they will require their own operating system.
As said, the possibility is always there for a hardware company to develop their own OS. This is what Nokia has done with their Symbian OS for instance, as used on all their current range of mobile phones. Symbian took a significant time to develop. Yet, by contrast, the OpenMoko project is producing a completely open source mobile phone in a much more limited time period replete with a fully functional OS and a suite of apps – thanks once more to the world of open source. OpenMoko touts their Neo 1973 handset as giving ubiquitous computing and putting control in the hands of the users.
There’s no need to look hard for another example. The One Laptop Per Child project – or OLPC – has lofty altruistic goals of providing low powered and easily manufactured portable computers, primarily for the youth of third world nations, but also for anyone who needs it. The design intentions impose some stringent hardware requirements – not in terms of grunt but in terms of availability and power consumption and other matters.
Once more it is Linux and open source software that helps out: firstly, a pre-written, pre-debugged operating system comes readily available. Secondly, the OS is freely customisable allowing full advantage to be made of custom hardware as well as stripping out items not required. Thirdly, a ready suite of software including tools and productivity apps is right at hand. Fourthly, there’s not a cent to pay to the open source developers; there are no licensing or royalty fees, and this importantly keeps costs down.
There’s no doubt that what Torvalds says is spot on, and that Asian hardware manufacturers, buoyed by the popularity of the ASUS Eee, will begin producing highly usable, customised, specialised and niche products which innovate in their hardware and design choices. It has been and will continue to be, open source software that enables this explosion to take place because of the freedom it gives to such vendors.
Open source software is all about giving freedom and control back to users and inventors. The future of the world’s hardware and technology will be intertwined with the rise of open source software. Bring the revolution to your desktop too. Take up the challenge to make 2008 the year you migrate to open source and liberate yourself from outside control and lock-in.