Apple or Ubuntu

[Colin Watson]

On Tue, Sep 18, 2007 at 07:58:25AM +0200, M. Fioretti wrote:
> On Mon, Sep 17, 2007 18:00:25 PM +0100, Colin Watson
> (cjwatson at ubuntu.com) wrote:
> 
> > > Absolutely right (see #6 of http://digifreedom.net/node/56). A
> > > very interesting and important corollary of this fact is that it
> > > makes no sense today to promote Free Software with the "you can
> > > study and fix the source code yourself!" argument. It's
> > > counterproductive, actually: you are telling people they can do
> > > (or give the feeling they may _need to do) something most of them
> > > couldn't care less about and would probably pay to _not_ do.
> > 
> > The best approach isn't to say "*you* can study and fix the source
> > code yourself", but "you can get somebody local to study and fix the
> > source code for you rather than having to wait for a faceless
> > corporation to do it".
> 
> This is a very good argument for local politicians or small/medium
> businesses, as it also creates local jobs which cannot be outsorced. I
> do point it out in the Family Guide, actually.

Definitely.

> I had not mentioned it only simply because I understand that this
> specific discussion is about individuals, i.e. single end users who in
> practice are almost never going to do something like that.

Oh, I don't know, my parents might well do if only by asking me.

One of the things we're trying to create with this whole free software
thing is an environment where knowledge about how to fix software is
decentralised. That means that you might well have the kid down the
street who's good at puzzling out computer problems, in the same way
people sometimes go to their neighbour who's good at fixing cars or
building garages. Aside from free software, software is an anomaly in
our society in that it has very little in the way of DIY culture; you
put prefabricated bits together, IKEA-style, but you don't build things
from scratch. Looking at behaviour in the rest of society, I don't
believe that's an intrinsic failure of people to be able to cope. I
don't believe in this dichotomy between programmers and "end users" who
just consume whatever they're given. Nobody talks about the end user of
a hacksaw, and yet they're on sale in the supermarket a few miles from
here. (Analogy is a great advocacy tool, and there are so many great
analogies to choose from here.)


I also think it's tremendously important to nurture the next generation
of programmers. I started out with a Spectrum when I was seven, which
was great; there were lots of magazines you could buy for pocket-money
prices that as well as talking about games often included source-code
printouts that you could type in and play with. When the Spectrum wore
out and we moved to a PC with DOS, that was still OK because there were
batch files and QBasic; not ideal by today's standards but I did plenty
of interesting things in that environment. Then we moved to Windows and
I was basically stuck except for the dodgy copy of Turbo Pascal I'd
acquired (sorry, Borland). Later I won some development tools in a
competition, but the environment was really drier for a young programmer
than it should have been. If there had been a system around at the time
that was free, provided a complete desktop with office tools and such,
and yet also included full development tools and the ability to use the
entire operating system as example code, I'd have jumped at the chance,
and it wouldn't have taken much to persuade my parents.

Sure, the majority of kids won't be interested, but a substantial
minority will. I don't believe that children have somehow got stupider
in the last 20 years; they can cope with it just the same way as we
coped with typing in BASIC programs out of magazines when I was young.
Somewhere out there is the next Linus Torvalds, and they're probably
dependent on their parents - i.e. the people we're marketing to - for
their computer. With any luck their parents might even care about the
quality of their education, and get something that sounds like it would
make a good learning tool rather than something marketed as an
appliance. Why do we have to assume that end users only want black-box
appliances and never want to learn?

> > > it may be useful for the OP to try answering it from the above
> > > point of view, that is end-user data ownership: in other words,
> > > after you have heavily used and customized Apple OS for one year,
> > > doing everything it lets you do... are you Free to move all the
> > > files you produced with that Mac (text, video, anything) and all
> > > your *configuration* data (email provider, bookmarks...) to
> > > another computer and OS without lots of reverse-engineering?
[...]
> > Psychological OS lock-in is certainly a major factor; changing
> > operating systems can be hard work.
> 
> Sorry, but I'm not talking of psychological lock-in,

Fair enough, I misunderstood your point; though I think mine is still
important.

> > heck, non-Debian-derived would be hard work - just because I've been
> > using it so long and have built up so many of my habits around
> > it.[1] This isn't because Linux is doing evil proprietary lock-in,
> > but because habits are hard to break. I think that aspect of it is
> > at least as strong for Apple users as any application lock-in
> > effect.
> 
> This aspect of the problem is relevant only if you can really start
> from scratch, that is if you never did anything relevant with a
> computer before, or you don't need anymore all the *produced* with the
> old environment.

This seems like a non sequitur. Both data formats and environment
conversion are important, and I never said otherwise. However, I'd argue
that data formats are much more important in a commercial setting, where
I entirely agree with you that interoperability is essential. My
experience of converting small-time home users is that they don't care
all that much about many of their documents, and only a small number
(e.g. electronic address book, the odd holiday video, old e-mails, maybe
some letters) are really important. It's great and much easier if they
all just work, but often some kind of jury-rigging is acceptable here as
long as creating, printing, and sending new documents works well. I've
done just this for home users by simply setting up the old application
to run under emulation when I couldn't find any way of importing its
data, and there was never any problem with this solution.

By contrast, in a commercial setting you can simply give your users a
short training course in whatever they need from the new environment,
but home users are both more inclined (and permitted!) to fiddle around
with odd things off the beaten path, and more likely to have built up
little edifices that nobody knows how to support. Converting these over
is a hard problem. I'm inclined to say that the best answer may be to
convince people that it's so much easier in the long run to customise
their computer to their taste with free software, but it's not an easy
sell.

> This is a common fallacy of many pro-FOSS migration
> talks: they more or less implicitly assume that the audience popped
> out of nowhere, that people are much more locked by past or external
> documents than by interfaces.
> 
> There are many private and small business users who _want_ to change,
> but cannot do it only because of formats. I can be eager to migrate to
> FOSS as much as I want, but if, in order to eat, I have to exchange
> with customers and suppliers large spredsheets with MS Office macros
> or Publisher files (just to name two in thousands of practical
> examples)...

I agree with you in the business market, but I don't know many private
users with customers and suppliers they have to exchange large
spreadsheets with. Home accounting is probably a better example there.
It is, at any rate, a very different setting and I don't think private
and small business users should be conflated.

Cheers,

-- 
Colin Watson                                       [cjwatson at ubuntu.com]