Tim Berners-Lee

Date started: December 19, 1997

Status: personal view. Editing status: As good as it will get for a while..

Up to Design Issues

Philosophy of the Web - Filtering and Censorship

Filtering and Censorship

Information is powerful stuff.  The world has been enthralled by the power which the Web, the universe of accessible information, gives to people and to groups working and playing together.

Information about information is powerful not just as information, but because it allows one to leverage one's use of information, to benefit from that which is relevant, accurate, stylish, unbiased, or timely, -- whatever one regards as being of "quality" -- without being enmired in that which is not.

Powerful tools are often usable for constructive or destructive purposes just as paper and ink be used for truth or lies, and metal for ploughshares or swords. The Web's power stems from its universality - for example that a hypertext link an point to any information out there, not just a subset.  People have asked whether I regret that the Web has been used for some uses, but I have to reply that if somehow it had been built to control the material which was placed in it, then that would be the technology controlling society, rather than the other way around as it should be.

True, one could take the view that our society is not strong enough to be trusted with a powerful information system.  One could take the view that society does not currently have the wherewithal to prevent the Web from being abused by destructive forces to an extent that the overall pain is greater than the gain.  I do not believe this is true. In the western developed world, at least, I believe that the democratic process will have sufficient control over governments and the judicial process sufficient control of criminals, to continue to defend the health of the evolving society.

We should be very careful, by constant inspection, to ensure that this continues to be the case.

Filtering and Censorship

One of the threats which posed itself in 1994 was of government censorship over information on the Web.  In general, there are information acts which societies regard as legal, and those which are illegal (such as fraud).  The problem which arose was that in the very subjective question of what information is deemed suitable for children, there was a threat that, in order to "protect" children, seeing no other alternative, governments were contemplating making draconian legislation for example prohibiting the transmission of "indecent" material. The problems here were many.

First of all, the concept of "indecent" was being enforced as a central single concept, quite against the distributed subjective nature of its definition in society.  The Web works as a decentralized system, with no hierarchical or other structure to force society into a shape imposed by technology.  This works.  Centralization of such an idea would [prevent the Web from being an accurate mirror of society itself.

Secondly, the problem being solved was the reading of such information by children, not its transmission. Thirdly, the question of "transmission" seemed to include intermediate parties who were not responsible for the content in an editorial or authorship sense. And one could list other problems, but this is enough for the present.

Information Quality

The basic problem being addressed was that of subjective information "quality".  This is the same problem reported by newcomers to teh web who find (typically after a search engine search) too much "junk".

It is unreasonable to ask for information delivered from the web to be of consistently high "quality" if you can't define what "quality" is.  There is a need, then, to be able to represent "quality" in a completely subjective way.

This is what the PICS project was all about.  PICS was specifically aimed at demonstrating that individuals could obtain their own subjective notion of quality without the government having to try to "protect" them by enforcing some centralized notion.  Politically, PICS is a system necessary for the preservation of free speech on the Internet.

The system needed a few different sorts of documents

a "rating system"
which defines a scale or scales on which one might judge a document.  The fact that anyone can create one of these is a strong force allowing decentralization of concept, breaking the problem of the global, centralized definition of for example "indecency".  PICS allows communities of any size (from one up) to establish their own criteria.  Agreement over a large community enhances global harmony, but threatens diversity. Agreement over a small community does the reverse. So in fact some balance is necessary
a "label"
which is a statement about something in terms of the schema.  This can be made by any party, not just author or reader, and certainly not just central government.  These can be created and exchanged in all manner of ways, so the PICS standard for interoperability is essential.
A "profile"
which describes for a given person the particular rating systems and levels on those scales which represent "quality" information at a given time and in a given context.  This sort of information can either be input by a person using a graphic interface (such as a set of sliders in a dialog box), or can simply be transferred from someone they trust, whether family, organization, or friend. Inability to transfer this would prevent people from building their own communities with common standards of trust: hence the importance for this (picsrules) as a standard.

These are all subsets of a general metadata language, designed to be easy for people to use.   In particular, by being limited in their power, they allow graphic interfaces to be built.

On social responsibility of technologists

The argument has been made that PICS technology should be suppressed as the power it gives may be abused by governments.  (There are even those who have suggested that the whole scheme is a government inspired plot to promote censorship and limit free speech.  This is certainly not the case, as neither in the idea,  the funding nor the intent.)  Whereas most readers may find this far fetched, it is worth a response on principle.

As I pointed out when closing the first International World Wide Web conference, speaking to (then a mere 350) geeky web enthusiasts, I firmly believe it is the task of scientists and technologists to be aware of and responsible for the social implications of their work.  This cannot just be left to "professional socially responsible people", as each engineer and scientist is often  best aware of the potential of the work.  Uttered in the auditorium at CERN, whose particle physicists trace their roots through nuclear physics, I don't think the message went unheard, even though it may have sounded strange in such a new field.  Now, (1997) the World Wide Web Consortium has one of its three domains dedicated to the relationship between Technology and Society.

So what about PICS?

The question basically is whether the potential danger of the technology outweighs the freedom and positive good it accords.  You can certainly argue this for nuclear fission, and you can certainly argue it for the wide distribution of firearms in populous countries. Can you argue this for PICS and metadata?  Is there anything about PICS specifically or metadata in general which makes it more of a danger than a boon?

The specific types of document in PICS are very general. As a system, it is quite generalist, and extremely decentralized.  It does as good a job as it can of leaving policy up to others to set, although it does (compared with other systems one could imagine) tend to favor by its nature cultural diversity, and freedom of speech, including freedom to endorse other's work.

 The specifications of communication protocols enable, but do not enforce, what manufactured software will or will not be able to do.  One cannot, therefore, at this level say what individuals will be able to do.  The technology can leave the policy up to others, which leaves other groups to ensure that the values which they hold dear are not lost in new legislation, industry practices, or public apathy.

A metainformation system allows one to talk about information. It enables all kinds of uses of information

It is not the place of a technical metadata system to try to limit the statements one can make with metadata, or the laws if any which are made.  That is the role of the democratic process and whatever government the people trusts. The W3C as an industry consortium can act for industry in promoting standards, but cannot act to create laws.  What we can do is explain to lawmakers and others the effect and intention of technology.  That is what this article attempts to do.


So Metadata, PICS and otherwise, is powerful, as is information in general.  Constant vigilance by concerned members of the public, industry and government is a very important part of the system of controls which keeps society healthy.  The PICS technology was created specifically in order reduce the risk of government censorship in civilized countries. It was the result of members of the industrial community being concerned about the behaviour of government. The indications are that in this it will succeed,  but that does not remove the need for such vigilance.  

To conclude, out of fear or ignorance, that PICS is more of a danger than it is a boon would be throw the baby out with the bathwater.  Metadata is not just a new tool, it is the start of a machine-understandable web (a "web phase 2") of information whose impact should be as empowering  to humanity as the human-understandable web of today.  We must understand it as we build it.

Tim BL,  December 1997

Last edit $Date: 2009/08/27 21:38:07 $