By James Green
Many people have commented and marketed, that GNU/Linux in it’s
various distribution flavours, is beginning to hit mainstream, and
most are pointing at servers more than desktops as the main growth
area. Whilst this is true, it is perhaps worth a thought or two for
the other side of the equation – while GNU/Linux is getting ready
for the mainstream, is the mainstream ready for GNU/Linux?
Unix has, from the very early days of computing, been the
popular and justified choice of Operating System for the server.
The nineties has seen Microsoft become masterful in the art of
product research and marketing. Microsoft has taken on Unix in the
server market with it’s NT product and indeed ground has been made
up, but now, of course, if the latest figures are to be believed,
confidence in NT and Microsoft in general is waivering and
GNU/Linux, the modern Unix, has begun the assault on Redmond’s
hopes of Microsoft-everywhere.
The Desktop is an area where Microsoft has excelled in both
research and marketing. The idea of the Windows95 desktop may not
have been Microsoft’s, but the implementation was good and uptake,
driven by this and falling hardware prices/higher demand for
computing, meant high sales of the OS and associated applications.
Hence we now have an enormous Company that is an effective monopoly
and that is bad for competition.
GNU/Linux is growing up, the kernel has been suitable for
mainstream server deployment for some time now, and projects such
as GNOME and KDE are bringing it’s power to ordinary users via
desktop interfaces quite similar to that of Windows95. But with the
effort well underway to take Linux to ordinary users, what of the
users themselves?
Mainstream is, to most people, the majority of computer users.
It is deployment overall rather than on servers or desktops
primarily. There are, of course, many more desktops than servers,
and this is possibly one of the main sticking points for calling
GNU/Linux ‘mainstream’. Mainstream users are home, office and
mobile users, servers are a specialist area.
Given that most computer users are familiar with the Windows95
desktop GUI, it seems an obvious choice to make the upcoming
GNU/Linux desktops compatible with the ideas introduced in it, the
icons, the drag-and-drop and associated actions. But even with
similar functionality in GNU/Linux desktops, will that make the
average computer user switch? No. Why? Because most people choose
familiarity and compatibility over what GNU/Linux has to offer –
technical superiority. They don’t care if they click a button
labeled ‘Start’ or ‘KDE’ or even a foot. They don’t even care how
to access the spell checker in their word processor, or even what
word processor they use – they just want to get the job done –
productively.
It is often said that the computer-generation is that which has
been taught using computers at school. Those who haven’t are either
taught computers by children or trained by specialists, or
sometimes both. I see kids coming into the workplace without a clue
how to do a job until they see the Microsoft Office menu entry in
their Start Programs menu. Familiarity and memory kicks in. The
same applies to adults who learn from courses in Office
applications.
Microsoft, therefore, has an important head-start. Bar the
noticable exceptions such as Mexican schools, Microsoft has a
monopoly on Operating Systems deployed in the education sector and
to a lesser degree in the larger world of the office. Users won’t
voluntarily format their C: drive (how many could?), it would have
to be up to the admins to make the change and then give their users
the training needed. In schools, a problem exists – do you switch
to a technically-superior OS which has a good chance of becoming
the future desktop in the office and home, or do you choose the OS
that parents know and is used by the mainstream? A catch-22 quickly
emerges.
It is likely that Microsoft will (barring disaster) dominate on
the desktop for some years to come, with GNU/Linux becoming a
commonplace secondary OS, and only then will users bother to spend
time in the new unknown. Skills learnt on Windows will be re-usable
on GNU/Linux with the deployment of GNOME and KDE, but there still
remains a major problem, one that is seen as a bonus by the current
Unix community and which could significantly hinder deployment of
GNU/Linux in the mainstream.
Choice and GNU/Linux go hand-in-hand. The development potential
on the platform is enormous, with libraries freely available and
source code to hand. On Windows no source code is available, and
the coder must go through the Windows APIs to get many tasks done.
On this note then, it is obvious how GNOME and KDE, Window Maker,
Enlightenment, fvwm and derivitives, and all the other standard
choices we make come about express the freedom to develop according
to your own ideas. But this leads to a complete lack of standards.
The Windows95 GUI is a standard in it’s own right, one which has
been learned by countless individuals as the way you use a
computer. Of course, Macs are avaliable, but not in significant
numbers.
When an office clerk comes to use a computer, s/he will probably
use a Windows95-style GUI. If the admin switches to GNU/Linux,
choices over UI on various levels need to be made. Ill-informed
admins may choose components that are incompatible with company
partners’ software choices and problems may emerge. The clerk may
become familiar with using Star Office with GNOME and
Enlightenment, then the admins may decide to use KDE and suddenly
the clerk will become confused. S/he’s house has been re-arranged
and now the washing machine has a different set of symbols for the
various loads.
The fragmentation may be GNU/Linux’s main point of frustration.
A home user may offer a friend help in setting up their printer,
but he knows the KDE method, not the GNOME method, and since the
friend’s parents have both just got to like GNOME they aren’t keen
on messing about getting KDE to work just to set up a printer.
Unless normal users are prepared to look at and learn the
choices they may find incompatibilities that lead them back to the
world-standard Windows packages. As a community, we need to be very
careful when creating and choosing options so as not to confuse and
frustrate the world at large.
Build GNU/Linux with the masses in mind, else risk the standard
Windows packages being used forever.
James Green
1999/06/18