Software errors

[Reinhold Rumberger]

On Sunday 23 May 2010, Mark Greenwood wrote:
> On Sunday 23 May 2010 14:06:47 Reinhold Rumberger wrote:
> > On Sunday 23 May 2010, Mark Greenwood wrote:

<snippety>

> > > I'm not criticising, just saying it's an inevitability. I
> > > understand the complexities of software testing only too well.
> > 
> > I got that. You were just making it sound a little too much like
> > the problem was unique to the Linux world for my taste.
> > 
> > > The difference with Windows or Macintosh is they have a lot
> > > more money to throw at testing,
> > 
> > You'd think so, but it sure doesn't seam like it sometimes.
> > 
> > > and a lot more central control over
> > > the development process of the code that goes into an OS
> > > release,
> > 
> > That is just plain wrong. The distributors have as much control
> > as they want, due to the open nature of the code. In fact, go
> > ask the kernel devs how many distros they know that don't patch
> > their software... ;-)
> > 
> > > as well as the fact that third party apps are the
> > > responsibility of third parties, who also do their own
> > > testing.
> > > 
> > > Uniquely with Linux, every single application is "released" by
> > > the Linux distro. They actually have more testing to do than
> > > MS or Apple, with about a millionth of the resources. The
> > > collaborative and open nature of Linux development allows
> > > person one to go off and change his library code as and when
> > > he pleases. Whether person 2 who is using that library in his
> > > application actually updates his application in time for
> > > Kubuntu's LTS release is up to person 2, not Kubuntu.
> > 
> > As explained above, they are free to apply any patches they
> > please. Sometimes distros even have errors removed that are
> > still present in the original software...
> > 
> > > Updates are frequent, collaboration is
> > > wide-ranging and flexible, and the number of potential
> > > pitfalls and incompatabilities is simply vast when compared
> > > to the uncollaborative closed-source world where app
> > > developers work in a vacuum with a pretty much stable (in
> > > terms of updates) OS. Hence, sometimes, Linux releases will
> > > contain software that doesn't work. Hence, sometimes, Linux
> > > updates will break software that was working before. Yes it
> > > happens in Windows-land too, but it's much more likely to
> > > happen with Linux for all the reasons I've mentioned. I call
> > > it the price of free software.
> > 
> > Well, you kinda comparing apples and oranges because you include
> > all the software present in the repos in a Linux distro but
> > only the software present in a typical install of Windows.
> > 
> > I would like to put forward the thesis that if you evened that
> > out (by either including more software on the Windows side or
> > just a default install on the Linux side), you would find that
> > there are more problems in the Windows world, actually *due* to
> > the closed nature. Somehow, the developers don't seem to feel
> > enough pressure to develop cleanly and securely since nobody
> > can see their code and criticise it. Ever tried using a
> > non-mainstream device that didn't my response was simply meant
> > to say that it could be almost anything, come preinstalled in
> > Windows? In Linux it will often just work while in Windows you
> > will have to spend days tweaking (not to mention rebooting).
> > Add all the faults due to bad documentation and no code
> > access...
> > 
> > I seriously don't understand why people assume that software
> > quality in a commercial project is automatically higher than in
> > a free one. My experience is that there is no difference on
> > average.
> 
> I understand what you're saying but you've sort of missed my point
> because I've pushed a button with you about open vs closed source
> software.

No, you didn't - but you've obviously missed my point... :-P
(I don't have any "buttons" about closed source software - it's how I 
earn my living. I have one about MS, but that's different)

> Forget about comparisons with Windows and Macs, those
> are not even relevant to what I'm trying to say. Forget about
> 'software quality' because that's also not part of my argument,
> it's a meaningless bullshit phrase invented by people with
> spreadsheets.

That isn't true, and anyone who ever touched a line of code written 
by someone else knows that. Well-written software follows some design 
and has a decent architecture which means that problems can be fixed 
with relative ease and the handling, etc. is consistent.

> What I'm talking about is the nature of, let's call it 'Free
> Software Development',

How about "open collaborative software development"? ;-)
Just since it has nothing to do with price and we might want to try 
to emphasise what you think is causing problems... :-)

> and the inevitable outcome of that
> process.  A distro release is a snapshot of many different
> threads of development at a point in time.

It isn't. It is a snapshot of many different threads of development 
at many different points in time with some changes by the 
distributor.

> This may easily mean
> that the release of one thread of development has not yet caught
> up with the latest changes in the release of another thread. Now,
> the distributors have the control to fix bugs, and to choose
> which versions of applications to take, but are you seriously
> suggesting that if some fundamental library completely changed
> its API that Kubuntu would fix every application that depended on
> it? If you are, you are a fantasist my friend.

Why are you assuming that they would ship this new lib if it broke 
half the system? Also, most libraries will be backwards compatible 
and/or installable side-by-side with older versions. That's what the 
version numbers in the name are for.

My point is that the same problems occur in the CS world, Windows 
being an example. Sure, not in the base release, but then that is 
unusable because it contains little to no end-user software. You 
tried upgrading to Vista when it came out? I was unfortunate enough 
to be forced to work with it, and believe me, that kind of breakage 
is *not* unique to open/free software development.

> That's the way it is, and that is why things that worked in one
> release are often broken by the next. It's just the way it is -
> lots of individuals working to no particular plan with no
> resources for testing.

That, too, is incorrect. IME, testing in popular open source software 
projects is better organised and more effective than in most big 
software houses. And your idea of the chaotic development in OSS is 
rather strange to me - no software can grow in size to a point where 
it seriously matters without some pretty well-defined kind of plan 
behind it.

  --Reinhold