Friday, 30 October 2009

Farewell, desktop? Personal background

There's a lot of work going on at the moment rethinking the desktop. Google's best known for this, and its new OS will take this further - though it's hard to know exactly how, as it's keeping tight-lipped about what it will actually consist of. In the meantime, other ideas are being developed, so I thought I'd document a couple I've looked at.

First, some background
I've had my own PC for the best part of 20 years now - which makes me feel very ancient - and I've never been a great fan of Microsoft's software. When I started, I was using a DOS program for my business which would not run under Windows (version 3 as it was then) because of the way it used memory. So I looked around for an alternative. I found it in OS/2, which was not only able to run the DOS program; it could also run other things at the same time. This was an eye-opener: a proper multi-tasking operating system in a PC! Although Windows became more capable over time, it remained technically far behind OS/2. Sadly, Microsoft realised early on the importance of getting the operating system pre-installed, so it cornered the market and OS/2 never got the market share it deserved. In addition, Microsoft always had the marketing savvy to ensure that Windows was 'good enough' for most people's needs as they (the users) saw it at the time.

By the late nineties, IBM had ceased to support OS/2 - they had other things to worry about, such as escaping bankruptcy - but fortunately by this time a new alternative to Windows had emerged, namely open-source software based round Linux. So, around 2000 I started to investigate what was available. There were 2 basic types of desktop: simple window managers, and more integrated desktop environments. The former were small and fast, but required rather fiddly configuration in text files; they were though ideal for old PCs without much oomph, so I used them on some old hardware as backups etc. For the main machine, I wanted a full desktop, and there were 2 main competitors for this: GNOME and KDE. These were similar in what they could do, and even at that time were able to do most of what could be done on Windows. KDE seemed to be more tightly integrated, whereas GNOME gave the impression of being more a bundle of not particularly related programs; so I went for KDE.

I used a variety of distros, starting with Red Hat, and using Mandrake (now Mandriva) for a while. However, whilst this was easy to install, I also like to know more what's going on behind the scenes, or 'under the hood' as Americans call it. When I discovered some program was sending out requests over the internet, and I had no idea what or why, I decided it was time for a distro where I had more control over what exactly was installed. I tried Linux from Scratch, and learnt a lot from this experience - though the main thing it told me was that compiling everything needed in a modern desktop system takes an awfully long time, and you're much better off leaving the compilations to those nice people in the distros who do it all for you. Eventually I settled on Slackware, which has the great advantage of simplicity and gave me more or less the complete control I wanted.

The disadvantage of Slackware, though, is that it only has a limited number of packages, and whilst there are independent orgs which package up other binaries, if you need other programs, you have to compile them yourself.

In the meantime, I also started using the internet in the mid-1990s - one of OS/2's advantages was its good networking support. Seems hard to believe now, but I started off with a 2400bps modem. OS/2 included some examples and, although at first I didn't really understand what the WWW consisted of, one of those examples was a link to the Botanical Gardens in Canberra, where you could see what was currently in flower. I am not particularly interested in either gardens or Australia, but the light went on regarding the potential of the web, and I became an early adopter of many web technologies, introducing them to some of the many voluntary groups I belong to, particularly the open-source ones that can be used for free.

There have been many claims over the years about how internet technologies would replace the desktop. I didn't understand Sun's slogan 'the network is the machine' when they first introduced it, but it later became a good metaphor for the interconnected world of computing. Sun also introduced Java, and applets were at one time supposed to revolutionise desktop computing; however, when I saw the speed at which they operated over my 2400bps modem, my main thought was: 'they have got to be kidding'. Oracle too unveiled a thin client which was supposed to put Microsoft out of business, but it disappeared without trace. More recently, there has been a lot of hype about 'Web 2.0', and this time it seems to me the many claims about revolutionising the desktop are actually becoming realisable.

Like many others, I recently decided to invest in a netbook; however, unlike many others, I was not interested in running Windows on it. I chose an EEE (primarily because of price), and this came with a version of KDE pre-installed. However, whilst this worked well enough, it was rather limited in what it could do, and the distro didn't have much in the way of additional packages either. So, time to install another distro. The best-known for these netbooks at the moment is Ubuntu's Netbook Remix, but that required me to download a large CD image (pretty daft, as few netbooks actually have a CD reader) and I'm not a great fan of Ubuntu anyway - particularly its lack of a root signon. So, I decided to go for Debian, which gives more control and has the great advantage of a vast library of binary packages.

This gave me the opportunity to have a serious think about whether I actually need a 'desktop' any more, or whether I could go over more to the web. What programs do I want or need nowadays anyway?

But that, as they say, is another story, which I'll put in a separate post.

No comments:

Post a Comment