W3C home > Mailing lists > Public > www-talk@w3.org > May to June 1995

Re: KidCode: Next steps

From: Terje Norderhaug <Norderhaug.CHI@xerox.com>
Date: Mon, 19 Jun 1995 15:35:46 -0800
Message-Id: <ac0b9c21070210048a70@[130.191.9.113]>
To: Darren New <dnew@sgf.fv.com>
Cc: Nathaniel Borenstein <nsb@nsb.fv.com>, Wes Morgan <morgan@engr.uky.edu>, rating@junction.net, www-talk@www10.w3.org
At 1:27 AM 6/19/95, Darren New wrote:
>Well, since a "document" is the smallest thing to transfer, you never
>really want "part of a document".  What you seem to want is to get
>different documents depending on your preferences.

I replied:
> I refer to document as what is transmitted from the author/content provider,
> not what is viewed by the reader.

Thus, while a browser may receive the a "document", parts of this document
might be filtered by the browser to better fit the needs of the reader. The
result is of course also a document, although not the same as the one
transferred.

You may view "Document" as a recursive structure (i.e. a part of a document
is also a document). However, it also make sense to just transfer parts of
documents, and compile these into a document presentable document in the
browser. It allows for some very interesting re-uses of information.

At 7:13 AM 6/19/95, Darren New wrote:
>>
>If what you're saying is that someone writes HTML then gives it to a
>service provider who then modifies it, I think that's something entirely
>outside the control of HTML or HTTP.

This does not match with what I am trying to say.

>If it's changed before it goes on
>the web, nothing in the web is going to be able to "fix" it.

You seems to have a censorship perspective on this issue. Giving the right
meta-information/markup of the document, a browser would be able to "fix"
the document to adapt to the preferences of the user.

>> No, advertisers would probably not want to label interspersed commercial
>> as such. However, the flexibility of digital technology/networks and
>> associated social factors support the creation of a Web commercial-zapper.
>
>Um, I'd love to see even the hint of an outline of how to automatically
>determine whether something is a commercial or not.  How do you
>distinguish a page of "neat links to photos of automobiles" from "neat
>links of photos of Chrysler's latest models"?

Just by using the same technology as groups can use to sensor indecent
content. A database is created, which stores references to content that a
group of people don't like (in this case, commercials). Then, the browser
can check with this database whether the document (or part of it) is a
commercial, and either avoid it, avoid reffering to it, or filter out the
commercial. Alternatively, you can build commercial-free proxy's. Nothing
new here. Of course, part of the determination is manual, but we assume
that people actually would pay to avoid ads (or group together do share the
effort).

I would guess that the counteraction from the ad people would be to make
the commercials occur randomly in documents. However, the more that can be
done to give those that use interspersed commercials a hard time, the
better. More trouble for them means the facilitation of alternatives.

>> My suggestion is that rating technology should be designed flexible
>> enough also to support zapping of commercials. Especially support
>> for rating separated from authors are interesting for this purpose,
>> as well as ways of coding content that facilitates automatic analysis
>> to classify the content of a document.
>
>Great! Propose something.  So far, all you've said is "KidCode" isn't the
>best thing for all purposes.   That's fine.  I know that already.

Such considerations should be an integrated part of a proposal. Thus I am
proposing something, not just bashing KidCode. If you look back at the
discussion, you should already have an early framework for alternatives.
I haven't really bashed KidCode, but rather tried to come up with some
constructive and normative ideas. Hopefully you might find something you
didn't already know if you look for it.

>Why not propose how to distinguish commercials from non-commercials.

Any suggestion can not be based on that advertisers would support it.
Here is one:

1. Start by building a database of links to commercials (blacklist).
2. Create a way to identify ads, for example, by a clearinghouse or
volunteers.
3. Build a filter that can be added to browsers, or tailor proxies so
that links to blacklisted documents are removed from documents.
4. Follow closely any development in the general standards that allows
exact addressing of parts of documents without requiring authors consent.
5. Enhance the databases, filters, and proxies to be able to filter those
parts of the documents.

I encourage you to build on this to propose other ideas that might further
the purpose of defeating the advertisement based revenue model for the Web.

>Please take into consideration that some people might want to see the
>commercials.

Of course. Note that I am not negative to product information or
commercials themselves. I just don't like them interspersed in my
content (personal reason) and realize the negative consequences this
revenue model has on a medium (ideological reason).

>While you are at it, remember that 1) I might want to see
>computer commercials but not car commercials,

This falls under the more general problem of adding meta-information to
facilitate identification of interesting information. It would to some
extent complicate the architecture of a proxy that acts as a filter of
commercials. I am somewhat sceptical to that people in general would prefer
to get specific types of ads interspersed in unrelated content. However,
there might be a wish for seeing _related_ commercials in content, just as
many read trade magazines partly for the ads.

>2) that if you make it
>possible to reliably zap commercials, *you* will pay for the web sites
>rather than the advertisers.

Exactly! The revenue model of the Web will be changed to the better.
However, it might be worth considering to support users choice in way of
paying. Those that rather view ads should be able to. On the other hand, it
would be a bad idea to protect advertisers as web technology is furthered.

>> power to try to keep status quo. This suggest that any system that help
>> people to avoid ads must be designed flexible enough to adapt.
>
>Sure. What do you suggest?

Decentralize as much as possible. Do not favor the blacklist created by a
particular institution or company. Provide easy ways to create personal
add-ons so individuals can themselves be innovative in how to avoid ads,
and stay ahead of the ad business. Make sure that many alternatives are
proposed as the technology is designed - so what do you suggest?

>> A commercial-based revenue model on the Web gives a reason for
>> content-providers to avoid doing any self-rating, as this might decrease
>> the number of people
>> actually seeing their ads. This might be particularly interesting applied
>> to "adult" content. The metaphor of local porn shops (Borenstein) might not
>> be very appliable when the "shops" are not a part of a community but exists
>> only to generate an optimal revenue from advertisement.
>
>Or the optimum revenue from customers.

Right.

>One or the other.  I've seen very
>few porn shops advertising anything on the *outside* of the shop.

Then, as I argued, the metaphor of a porn shop might not be relevant.
Thus, we cannot assume that porn sites will behave in the same way as
a physical porn shop.

>Besides, in the US, it's already essentially illegal to put porn up without
>charging for access to it, in some sense.

The medium reaches outside US, and so does its users. Thus what is locally
illegal might not restrict what will appear.

>> This would be a bad decision, as content providers might have a different
>> opinion about appropriated age than various people (in various cultures).
>
>So who has a better concept?  Do you want each country or state or
>community to have to build a rating authority?

I propose that instead of tagging the content with "ratings", they should
be tagged with criteria that can be used to calculate a rating. Thus,
instead of coding the document with "Not appropriate for kids below 7",
it should be coded with something like "Intercouse, visible penis, blood,
rape, no nude women" on which an algorithm can be run to calculate a rating.
This allows different rating algorithm, in the same way as logical markup of
web documents allows tailored presentations. It is less cultural dependent,
and more open for various individual preferences.

As a start, it might be an idea to look into the criteria used to evaluate
movies in various countries. I am sure that it would be possible to create
some form for categories based on this. Then, pick a random selection of
movies, identify which categories are present, and run this with the actual
rating into a fuzzy system. Voila, you have an automatic rather that
follows their norm...

>Where's the money for running a rating server going to come from.

Well, you have some potential revenue models. Pay-per-use, subscription,
not to mention - commercials :-)

>Remember, some of these sites get hundreds of thousands of hits a day. If
>every hit from every adult web site had to route thru a rating server,
>you're going to be paying big bucks to run the rating server.  The
>solution *has* to be distibuted. You'll never keep up with any sort of
>centralized rating.

Agree.

>> Implement the technology is in any case the easy part. When we know what we
>> want, implementing it is not a big deal. A number of social and political
>
>Right!
>
>> decisions are made as such technology is implemented, or should preferable
>> be
>> made in advance and as the technology is implemented and we learn more
>> about its larger scale consequences.
>
>I have seen very few social consequences predicted in advance of the
>widespread deployment of a technology.

There are all to many "media determinists" out there, saying "technology
will..." and then some determined suggestion. However, it doesn't work that
way. The future isn't predetermied, but depend on the choices we made. Some
of these are made as we design technology, and choose between alternatives.
The social consequences we want to come doen't come automatically. Choices
needs to be taken based on what we know and believe potentially can happen
- and sometimes we just cant know that, either.

>And by the time you get around to
>getting the Catholic Church up to speed on how to put digitally signed
>ratings on other peoples' web pages, we'll have moved on to a completely
>different technology and interface mechanism.  (I mean, imagine things
>were rated in the FTP protocol 5 years ago; would it help now?)

Too bad for the Catholoc Church. Too bad for slow legislators. No reason to
take their role, unless (as might be the case) it would be strategically
beneficial.

>> That something can be done doesn't
>> mean that it should be, everything is possible with digital technology
>
>OK. Write up an RFC saying how you would build a commercial-zapper.

I would be happy to collaborate with you and others in doing this. I
believe considerations should be taken also in areas where I do not have
the expertise. I believe an appropriate strategy would be to identify how
currently evolving rating technology can be used. Are you ready to join
with your mind in the effort?

-- Terje <Norderhaug.CHI@xerox.com>
   <URL:http://www.ifi.uio.no/~terjen/>
Received on Monday, 19 June 1995 18:42:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:17 GMT