Mozilla

Posts Tagged with “data”

Final (??) 2010 goal for data

December 18th, 2008

Here’s a slightly revised version, based on the Tuesday Air Mozilla / IRC discussion.

Goal: Make the explosion in data safer, more useful and more managable for individuals

  • Products offer people realistic options for understanding, managing, combining, sharing and moving data created by or about them
  • People expect the ability to understand,access, manage, combine, share, and move their data

Revised “data” goal for 2010

December 14th, 2008

My proposed revision of this goal is:

Goal:  Make the explosion in data safer, more valuable and more managable for  individuals

This would be followed by some subpoints, along the lines of those below. They need some work, but I want to post the general idea for reaction before I spend more time on the subpoints.

  1. products offer people realistic options for managing data created by or about them
  2. people EXPECT access to their data, ability to combine it, move it, manage it as theirs

The change is because I don’t feel we have a solid enough consensus on the original proposal. This was:  provide leadership in

  • helping people exercise better ownership and control over their data
  • making anonymous, aggregate “usage data” more of a public resource

The idea of making any data available to anyone has generated concern. Some of this I think is due to a lack of concrete examples, or to a misperception that this would involve Mozilla software tracking people’s behaviors, or to a concern that it’s hard to anonymize data. But some of the concern is a basic discomfort with the currently invisible generation and processing of so much data, or the idea of a public “data commons,” or a concern about what’s happening “to” me through my software.

The data explosion is only just beginning, and it’s powerful. New forms of data can help us understand new things and solve new problems we can’t even see the shape of yet. But there’s a risk that each one of us will end up at the mercy of others who control the data. This risk affects our privacy, the degrees of choice and innovation available, and the degree of centralization of our online lives.

We should have a goal that reflects both the potential benefits and the risks of the data explosion.

2010 Goals: Data Discussion

October 27th, 2008

On Tuesday, October 28 at 1 p.m. Pacific Time I’ll be holding a session to talk about the privacy and data section of the Mozilla 2010 goals. This will broadcast on Air Mozilla. We’ll also use the IRC channel (#2010goals) for people who want to ask questions via IRC. As usual, Asa will be the contact for Air Mozilla aspects of this.

I’m doing this Tuesday because a big set of the platform team will be in town and available at that time. That creates an opportunity to have a session that includes face-to-face discussions with a set of people who are rarely together. This follows the goals discussions in Barcelona this weekend, and will proceed other goals discussions among smaller groups where hopefully people will feel more engaged and comfortable speaking up.

I’ll get a more complete schedule for other goals discussions together shortly. If you want to participate in a session on data but can’t make the Tuesday session, let me know. If there are a lot of such people then I’ll schedule another session. Or you might contact anyone who is part of the other Mozilla contributor groups I listed — they will be participating in goals discussions.

Disconnect Regarding Data

October 6th, 2008

I’ve read the comments to my last post a number of times and I think I now understand what’s happening.

There are a bunch of comments along the lines of “if Firefox starts to include something that tracks my behavior and automatically sends that information off to someone else, then I don’t want to use it anymore.” Absolutely. I don’t want to use that kind of product either. That’s why I’m part of Mozilla — to build products that don’t do this sort of thing. To be explicit: Firefox and Mozilla will remain intensely focused on privacy, protection of personal data and user control over that data. The Mozilla community won’t build or support products that do otherwise.

There are also some comments that discount the examples I used because they are “server-side.” Yes. Absolutely. The examples are server-side because that’s what I mean.

The kind of data I’m trying to talk about is more like census-data: how many people are using the Internet; what are the broad patterns of Internet development and usage. In our physical lives, the basic demographics of our population collected in a census are a valuable shared resource. In understanding the Internet aggregate, anonymized, server-side census-like data can also be a valuable public resource.

This kind of data can of course remain a private resource, held by those websites big enough to generate their own understanding. My point is that moving some of this census-like data from the private to the public realm could have great benefits.

I’m wondering if this distinction, which is so clear in my mind, has not been clear in my writing. The term “usage data” may have made this worse. I explicitly do not mean using the browser to collect individual usage data. I mean looking at broad usage patterns that can be discerned from aggregated, server-side data, such as the examples I gave before.

Basic Examples of Usage Data

September 28th, 2008

In past posts I’ve said that I believe there is a need to make basic, aggregate, anonymized information about Internet usage more widely available. If everything that is known about the basic usage of the Internet is closed and proprietary then the Internet as an open platform will suffer. Here I’ll try to describe the kinds of data I’m talking about. For now I’ll call it “usage data” though that’s just a term of convenience.

There is a set of usage data that we’re quite accustomed to seeing in aggregated, anonymized form. Unconsciously I think many of us have come to realize that without public availability of this data we cannot understand even the basics of how the Internet is working.

One familiar example is the amount of bandwidth a site serves. Bandwidth data is critical to planning capacity and making sure the website doesn’t “go down” when spikes of traffic occur. Bandwidth usage is also tracked quite carefully by the ISPs (Internet service providers) for their planning and billing purposes. As an example, here’s a blog post showing bandwidth usage when we brought our facility in Amsterdam online. In addition to the data, there are also a series of posts about what was involved in making this happen, which we hope helps others who want to do similar things.

Another familiar example is the amount of “traffic” to a website in a day or a month. This is one important method of determining how popular a website is. Changes in these numbers can reflect trends and changing behavior. A specific page view might be associated with a particular person, and thus be sensitive personal data. But the total number of page views is not related to a specific person. It tells us overall how popular a site is.

A third familiar example are download numbers, which can be very informative in specific settings. For example, we had a real-time download counter during the Firefox 3 Download Day event. We were able to provide automated counts of downloads and current number of downloads per minute, each broken out by language, during this event.  Here’s some basic analysis of download locales, showing how global a project Mozilla is. And here’s a post showing the effect on download rates caused by a popular talk-show host. This information can be useful without any personal or individual data being disclosed.

These examples are clearly very general. I use them precisely for this reason — to demonstrate that we already understand the usefulness of this type of data and that it can be presented in an aggregate, anonymous form. There are other forms of aggregate, anonymous data that can be equally useful in understanding how the Internet is being used and ultimately, understanding what the Internet really is. I’ll describe some of those in a subsequent post; this one is long enough for now.

The types of data I’ve described above are carefully tracked, analyzed and used in planning and decision-making across the industry. It’s often not publicly available. We’d like to see more of this sort of information publicly available. We hope to start publishing more of this type of information about Mozilla. To do this, we need to be confident that people understand this is not publishing personal or individual data, and this is not Mozilla changing.

This is part of our effort to make the Internet accessible. At the same time, Mozilla will continue to be at the forefront in protecting individuals’ security and privacy.

Data — getting to the point

July 24th, 2008

I’ve received a couple of emails from people saying it’s hard to comment on the data issue without some idea of where I’m heading or what I’m thinking. So here goes. I’ll come back to some of the topics I’ve written about already. And I’ll continue with the other posts as well; I think we need some depth of analysis to make good decisions.

In the meantime, here’s the basic message.

I would like to see Mozilla provide more leadership in helping people manage the collection and treatment of data related to them — what I’ve called “Associated Data.” I don’t have a specific plan of what leadership would look like, or what features or capabilities this means our products, services or websites should implement (or block). There are a lot of different types of Associated Data; the desired treatment of different types may vary. This is something I’d like to see us figure out.

I would also like to see Mozilla provide leadership in treating some basic aggregate, anonymized usage data as a public asset. To do this, we need to develop a sense of what data this might include and what aggregation and anonymizing techniques make the Mozilla communities comfortable. Some data — like public disclosure of bandwidth use, website rankings, etc. seem to be areas everyone is comfortable with, but we should make as few assumptions as possible. Sometimes it can be hard to get truly anonymous data and so this is an area where great care — and therefore  leadership — is required. But if everything that is known about the basic usage of the Internet is closed and proprietary then the Internet as an open platform will suffer. I don’t have a specific plan as to what Mozilla might do here; that’s the point of the discussion.

These are difficult and sensitive topics, it would be easier to ignore them. But both of these areas are critical to building the Internet that is healthy for the individuals using it. The Mozilla mission is to keep the Internet an open platform, and to promote the values in the Mozilla Manifesto. It will be hard to do this if we ignore the effects of data.

Data Relating to People

July 23rd, 2008

In my last couple of posts I’ve described why I believe Mozilla must pay attention to data in order to help individual people deal with  data about them.

There’s a lot of data about people being created.  I’ve listed below some of the basic kinds of this data  that I think we need to be able to distinguish in order to speak meaningfully the effects.  I’m calling all of these categories “Associated Data” for the reasons described at the end of the post.

Is there a type of data about people that’s of interest or concern to you? If so, take a look and see if it fits into one of the sections below.

  1. “Personal and potential personal data.”  These terms are already in reasonably wide usage to mean specific information that identifies an individual, such as name, address, email address, credit card number, government-issued identification number, etc.   In some cases it’s used to include other information that can be combined to create personal information, such as an IP (Internet Protocol) address.
  2. “Intentional Content.” Data intentionally created by people to be seen by people.  When we post to social networking pages,  blogs, photo sites, product review sites, create wishlists, send gifts and other online markers we intentionally create content about ourselves or associated with us.   Sometimes this information is in big chunks, like a blog post or photostream; other times the information is in small bits like a recommendations, “pokes,” etc.  Sometimes we want this data to be public and sometimes we may not.
  3. “Harvested Data.” Information gathered or created about an individual through the logging, tracking, aggregating and correlating of his or her online activities.   It’s possible today to record just many of the actions someone takes online (the “clickstream”) and then to harvest patterns and other useful facts from that data.  For example, an e-commerce website you visit regularly will know a great deal about your shopping patterns, what kinds of items and what price ranges you look, how many times you look before you buy, the average purchase amount, the average time between purchases, etc.   They’ll know which ads you respond to and which you ignore.
  4. Relationship Data.  Our relationships with other people, such as our “friends” or followers at various sites.  This can  be either Intentional Content or Harvested Information.  I call this out specifically because a relationship always involves at least two people.  And so the treatment of this information — is it public or private, how is it used — always affects at least two people.  I’m not yet positive this is a useful topic, but (obviously) I think it likely enough to include it here.

“Associated Data.” It will be helpful to have a term that describes all these types of data.  In a vacuum “Personal” would seem the best because this is all information that somehow identifies, is related to or associated with a specific person.  But I think “personal” is understood as item 1 already.    I’m using the term “associated data” to mean all of the types of data listed above.

Are there other broad categories of information about people that would help us think clearly? Are there different categories altogether that would be more helpful?  And are there examples of this kind of data you’d like to make sure we think about? If so, note them in the comments or somewhere where we can find them.

Why focus on data?

July 22nd, 2008

I’ve said in a previous post that I believe Mozilla needs to pay attention to the amounts and types of data that increasingly define the Internet experience.    I’ve even created an outline of  different topics relating to data that I think should be part of the discussion.  Why is this?

Principles 3, 4 and 5 of the Mozilla Manifesto state that:

3.  The Internet should enrich the lives of individual human beings.
4.  Individuals’ security on the Internet is fundamental and cannot be treated as optional.
5.  Individuals must have the ability to shape their own experiences on the Internet.

These principles are at risk if individuals have no  control over the creation or use of the data that describes us.    These principles are at risk if we sit back and hope someone else addresses them.  We need to build them into the Internet.

What can and should Mozilla do to help people be safe and in control of their online experience in the midst of this rising sea of data?

Framework for discussing “data”

July 21st, 2008

Here’s my starting framework for the data discussions.

  • Outline the different reasons data is important/ different ways in which understanding data is important to building the type of Internet we want to live with.
  • Develop some shared, reasonably specific terms about the kinds of data that related to people.
  • Develop some shared, reasonably specific terms about other kinds of data, in particular the types of aggregated data that tell us  how people are actually using the Internet.
  • Identify a range of overall approaches consumers might take, or want to take, regarding data.   I’m thinking this is at a very high, general level, such as the approaches of not caring, to a considered trade of data for services, to absolute control of all data.
  • Identify approaches Mozilla  could take to data, at a very high level.   Everything from avoiding the collection of any data to making the trade-off between convenience and data clearer to consumers, to providing tools to help consumers with these trade-offs.    Each and every approach Mozilla might take must be one that is based on our stated principles of safely, control and benefit for the individual human being, and on promoting the Internet as an open platform.    There’s a range of possibilities there and I imagine we’ll have some lively discussions.  This should not obscure the fact that there is an entire set of activities that we will not consider.
  • Identify what Mozilla might/ should do with our products, our websites and the product related services.  Some of this discussion is underway already of course, with the anti-phisihing, anti-malware services we offer in Firefox and the discussion of website analytics that occurred via blog and discussion groups earlier this year.   Setting these within a general framework will be very helpful.

The last couple of topics are discussions where our values and goals are critical.  These are areas where Mozilla actions — if any — regarding data will be distinctly Mozilla.   In other words, actions based on our mission, and designed to bring the principles of user safety and control to life, and to promote the I health of the Internet as an open platform.  I suspect the temptation to jump to this last couple of topics right away will be high.   And we’ll probably jump backward and forward a  few times.

It’s important to have the earlier discussions, and to do so with a focus on developing shaved concepts and vocabulary.     Let’s develop a shared understanding of the kinds of data that exist, and then we can talk more intelligently about whether it “should* exist or if and how it should be regulated or controlled.  Similarly, let’s develop a shared understanding of the high level approaches consumers could take with data before we discuss what approach each of us thinks they *should* take with data.

After we have enough shared vocabulary we can talk more effectively what Mozilla can and should do regarding data to manifest the principles of openness, innovation, and user safety and control in our activities.

Thinking About Data

July 21st, 2008

Our online lives are generating increasing data about us as individuals and about how groups of people are using the Internet. At the dawn of the World Wide Web 15 years ago people “surfed” to websites and viewed information. Today Internet life is more participatory and people create more information. In addition, a range of tools have been developed for tracking and generating data about people and our activities. The existence and treatment of this data is important to our online security and privacy. The treatment of this data also affects the public ability to understand how people use the Internet.

I believe Mozilla must think, talk, and respond to this new level of data somehow. I recognize that any discussion of what Mozilla should do regarding data may be perceived as Mozilla wanting to use data to make money, or otherwise changing our nature. This is not the case. Our goal in thinking or doing anything regarding data will be to improve the safety and control of individual people, and to improve the overall health of the Internet.

I’ll say this in many different ways, but I expect some will remain suspicious. The good news is that people are sensitive to this topic precisely because they recognize that the treatment of data is important. I’m hopeful that people will give us the benefit of the doubt as we have these discussions. And if that’s not possible, at least keep an open mind.

Because the topic of data is so complex, I’ve put together an outline of the different facets of this conversation that are important to develop a shared understanding of the landscape. From there we can integrate this understanding with the Mozilla mission and the principles of the Mozilla Manifesto. I’ll post that framework in a separate post.

Skip past the sidebar