Lennart Poettering: BSD Isn't Relevant Anymore

Polytropon freebsd at edvax.de
Mon Jul 18 20:31:45 UTC 2011

On Mon, 18 Jul 2011 15:48:46 -0400, Jerry wrote:
> On Mon, 18 Jul 2011 16:58:08 +0200
> Polytropon articulated:
> > Here the circle closes: Without STANDARDS, you wouldn't
> > be able to view the digital pictures you took with a
> > camera 10 years ago because the manufacturer decided
> > to use a proprietary image format without any documentation,
> > as you should only use the software supplied by the
> > manufacturer. Dropping program version X and advertising
> > version Y with the new models of the digital camera,
> > and everything you'll have is a bunch of files nobody
> > can read anymore. You can also see this in computer
> > media, although with a lower half-life period.
> > 
> > If you want to get into the future, rely on established,
> > open and free standards.
> > 
> > In my opinion, there is no alternative. Everything else
> > would just increase costs (e. g. migration costs). But
> > there are fields of use where costs simply doesn't matter
> > (as it seems).
> I apologize for cherry picking this; however, your analysis is so
> faulty that I was force to. You camera analogy is simply absurd.

I wanted it to be understood as an analogy.

> You were aware that Kodak dropped the C22 development process decades
> ago which effectively make all films designed for that process useless.
> It also spelled then end of GAF, but that is another story. KODACHROME
> Film was discontinues after a 74 year run. Actually, it was created due
> to Kodak's inability to properly stabilize the layers in the color film
> it was trying to create; but that is another story. I still have
> several collector's grade cameras that used films such as the 116 and
> 616 designations. These films were discontinued in 1984.

You're talking hardware (film material) here, not software.

Your analogy illustrates how technology does disappear. It
gets more and more complicated working with film material,
as digital cameras allow you to do all the things that you
could do with expensive cameras only in the past. Even
professionals have switched (of course to expensive and
therefor professional camera models), both for photographing
and for movies.

In software, see "planned obsolescense" and "digital medieval
times" ("digital middleage") and movements that want to keep
witnesses of our today's culture.

This means you will _always_ have to judge: Need a short-term
solution that is "the best" for a short term, or need a long-
term solution that is "good" (or even just "good enough") for
a longer period of time.

Sloppily engineered and halfway done solutions can - by means
of marketing - be sold for the first kind of products quite
easily, and "constantness" is not an important topic for the
main markets (home consumers).

> Should I sue
> Kodak, or any other manufacturer for their failure to continue support
> for these devices? When wan the last time you purchased a new Polaroid?
> News Flash: It was discontinued. Now, can you guess why? Perhaps you
> have noticed people using cameras that don't apparently use any film.
> You might want to investigate that further. You will find that newer
> technology supersedes and eventually obsoletes older technology.

It's _always_ that way. Interestingly, some oldest technology
still prevails. There are still books made of paper even though
there are alternatives. In the last year, more paper was used
and printed than in the year before, and the trend continues.
Even if you can argue that the use of actual paper is less and
less _required_, it's more and more _performed_.

We know paintings in caves older than 2000 years, books
older than 1000 years, paintings older than 500 years.
What will be present of our _today's_ digital culture
when the encryption codes are lost? When there are no
drives to read the media, or the media simply dissolved?

Of course you are right that newer technology will _always_
supersedes and eventually obsoletes older technology. But
you will also have to agree that technology will be used
as long as it's possible to make money from it, just see
petrol-driven cars as an example, and oil-based technology
in general.

> The point is, time moves on and technology advances.

Advances - yes.

Improves - not implicitely.

Fast and with best intentions for whole mankind and
environment - debatable.

Time moves on, and it's hard _not_ to move on.

I may point you to the "Matrix" movie trilogy. When mankind
finally looses interest in what it creates, because industry
tells us "It's all okay, just buy, just consume, it's the
best for you", then we will be unable to control our own
future. Just voting with the wallet seems to be insufficient.
It IS important in a market, but as the market isn't free
(as per definition), it's hard to see 100 percent control
in here. Free alternatives must be present in order to keep
the commercial products "on track", so they follow the needs
of the customers instead of _defining_ them. This would only
make technology its own purpose, and finally, in the end of
the ongoing obsoleting, it obsoletes man.

Maybe a strange approach, but do _you_ want your life to be
fully defined by others who create a golden cage for you,
which still would be a cage?

> To continue to
> keep an industry shackled to an out dated protocol simple because some
> user, somewhere, sometime, may actually use it would only serve to
> enervate the software and hardware industry.

There are (admittedly not obvious) branches of industry
where "old fashioned" stuff is still used, often because
it "just works" and there were massive costs in making it
work, and there are expected much more costs to migrate.
An example is MAN Diesel, using proprietary VMS software
on a SimH emulation basis. Or look into banking, into
insurances or fiscal administration where "old fashioned"
mainframe stuff is used. Those branches do _not_ advance
in a way economy would want it. Just find out what COBOL
is and why it's still important, as well as RPG, even if
it may be older than you. :-)

> Further more, this would
> serve to invigorate a cottage industry based on creating applications
> that could be used to "convert" such files to a newer format. Actually,
> several such programs exist now.

Yes, OpenOffice for example. :-)

Of course, formats _not_ requiring continuous conversions
are better and more efficient. But after all, it's not
about efficiency in fields of use where money is not
important (or can be "re-gained" from customers or

> I really hate the way "standard" is used by so many FOSS users. They
> use it as a shied against innovation. Rather than embrace newer
> technologies, they throw up the "standard" shield and claim that
> product "A" (product being anything your want it to designate) is bad
> because it doesn't follow some arbitrary standard.

It doesn't have to be that way. As an example, see HTML 5.
This standard obsoletes the "how it looks" of traditional
HTML and replaces it by "what it is", separating semantic
description from style description. Sounds familiar? It's
what LaTeX does for many years now.

Another field where standards are needed is when content
should be brought to the widest audience. Proprietary
formats simply can't do that. Patent & IP problems are
mainly - in my opinion - the biggest stop sign for inventions
and improvements. How can a developer be sure to release
something new or make a product from it when he has to fear
to be sued into ground next day?

THIS is the big danger for developers.

Standards are the key to market share, they must be free and
open to be implemented. A camera that can be attached to
_any_ Mac or PC will sell better than a camera that works
with just _one_ selected operating system version (and also
requires a plug similar to USB, but mechanically slightly
different, and electronically fully different). A cellphone
that can connect to _any_ provider will be of higher interest
to customers than one that just works with one provider in
a specific region of Europe. And a printer that can be fed
normal paper will be more comfortable than one you have to
buy specific-sized special paper for.

A good example are network-connected devices, either by
Ethernet cabling or even wireless. Just assume there would
be no standards in plugs and protocols - network printing,
NAS and special devices such as networked surveilance
cameras would be the total chaos.

> A product will
> stand or fall on its own merits. 

Those merits are mostly defined by the customers buying that
product, so market share is all that counts.

Is it? For economical reasons - for sure. For improving the
society, and if it's just by adding something to the diversity
of available options...

> To insist that any product follow any
> strict guide lines effectively removes the developer's ability to
> improve upon or create new or better products.

Conforming to standards is exactly what brought the PC to
where it today is. In the beginning, complying to industry
standards (e. g. serial port RS-232) was the alpha and omega
if you wanted to sell any computer, as the primary target
group insisted on it. Over time, demands have changed, but
target groups also have. Today, home customers seem to be
in the focus of software makers, and you can clearly see
this in the platforms used and programs created.

Sadly, history of program development has shown that the
market leader defines standards, just the same way as the
winner of a confrontation rewrites history afterwards
according to his own requirements. So what's left for
FOSS development? Mostly to re-implement what's already
present on proprietary systems. There is even no need for
improving it, no need for "better", as "the same" is what
users are interested in, of course "the same for free".
You'll even find inferior products in fields where better
free alternatives do exists, but due to marketing those
are out of scope.

> In my own country, we had the basis for HD TV back in the early 80's. I
> know individuals who were working with RCA at the time. Yet, it took
> 30 years for the industry to finally dump the existing framework and
> basically start over, You see Poly, sometimes you do have to change,
> unless you want to go the way of the dinosaur.

With the development of FOSS in mind, you have the chance
NOT to "swim with the swarm", with all advantages and
disadvantages it implies.

Your TV example is very good. I've recently read a text
that predicts the future of CDs - a text from the late 80's.
When we consider what we are _currently_ using, the text
predicting "no important future for CDs" looks quite funny.

> Now, if this had been a
> FOSS project, we would still be watching B&W TV on a big 19" screen.

Sometimes, the old fashioned technology is _everything_ you
can rely on. For example in regions where you have no internet
connection and no cell phone access, for broadcasting you
use middlewave (AM) radio, and maybe CB for communication.

Coming back to the TV example, it's still funny how industry
continues the ongoing renewal of technology, making "images
sharper than reality", obsoleting things by newer things that
are supposed to be the nonplusultra, just to obsolete them
within few years. Abandoning analog satelite TV is a good
example, as well as terrestrial TV. There are many other
approaches sold as nonplusultra that have vanished today.

As I often say: The top technology of today is the garbage
of tomorrow. It's just _your_ decision in how far you do
participate in an ongoing renewal of all your stuff - and
if you can afford to do so.

Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...

More information about the freebsd-questions mailing list