We'll see | Matt Zimmerman

We are a large community, but only a small number of people can travel to attend UDS in person.  So, over the years, we’ve experimented with different approaches to enable remote participation in UDS.  If you participated in UDS remotely (for example, using the audio feed, IRC, Gobby, etc.), please tell us about your experience by filling out this survey:

http://www.surveymonkey.com/s.aspx?sm=1HYiY7yv7f3_2b2EXcFhsecg_3d_3d

If you registered to attend in person in Barcelona, you’ll be receiving an email with a (different) link to the (same) survey.  Please use that one instead, so that we can easily sort feedback from local and remote participants.

I recently attended a talk given by Dan Lyons. It dealt in part with the economic and professional impact of digital media on journalism, and spawned a lively discussion.

Reporters have traditionally provided access to information and points of view which are otherwise unavailable to the reader. With the explosion of digital content, from individuals and publishers, access to information and points of view is rapidly approaching a price point of zero. News, commentary, reference material, humor and fiction are all plentiful. If information wants to be free, information is getting its wish.

This means that selling access to information is a rapidly shrinking business. The variety and quality of fact and opinion available for free on the web is far beyond sufficient for me, and I expect the same is true for many others.

In fact, there is too much of it. The challenge is in finding the good stuff. I rely on a network of friends, aggregators, and search tools to find the material which is of good quality and relevant to me. I need to invest a significant amount of time and attention to make it work, and in some cases it requires specialized technical skills. There are other people who do this much better than I do. That seems to me like an opportunity to offer a valuable service.

I think such a service would require a strong human element. Perhaps eventually, science fiction “intelligent agents” will be developed, which will scour available content for items we might want, learning and applying knowledge of our preferences. Until then, humans, with the use of technology, seem much more capable than automata of identifying good and relevant content.

So, how about a personalized service, which would provide me with fresh, good quality, relevant content, from many sources? Not a simple tag/keyword/recommendation engine system, but something truly personalized. That is something I would pay for. Could this scale so that one human could service many others in this way? Is anyone already offering something like this?

The idea is similar to the way in which Canonical derives revenue from free software, by providing the content (software) at zero cost and providing services around it, including packaging and certifying other software.

An unlikely duo spotted in Barcelona:

I’ve written briefly before about how our lives are digitized at an increasing rate and wondered about the impact this has on our social behavior.  One particular arc which interests me is how we engage in discussion.

Usenet news

In the early 1990s, newsgroups were the norm for discussion.  They were multiplying daily, and for almost any topic, chances were good that there was already a Usenet newsgroup, whether you were a recovering system administrator or a kibologist. The killer feature of newsgroups was their universality: anyone could get access to a Usenet feed via their ISP, and in one shot be connected to the full spectrum of discussion groups. Their weakness was that, despite the vast namespace, there wasn’t room for everyone. There were already too many groups, and ISPs grumbled about the time and expense of shipping all of this data around. Why should you have your own newsgroup? Newsgroups existed everywhere and nowhere, and authority was questionable at best. To create a new newsgroup, you might be expected to discuss it in a certain place, ask a local administrator, participate in a voting process, or all of these and more. Even if you did it all right, someone who disagreed could delete your group at any time by posting a control message from anywhere in the network.

It took some time for a new message to propagate through Usenet, and most people didn’t read their news all that frequently, so discussions progressed at a moderate pace. A typical discussion lasted for days, and a popular thread might go on for months as it spiralled away from the original topic. News reading programs developed advanced features for filtering out irrelevant discussions.

FAQ editors distilled the accumulated wisdom of newsgroups into documents, and posted them to the group periodically as they were revised. People worried about the signal-to-noise ratio of popular groups, and predicted that Usenet would become worthless when it dropped too low.  This was a constant perceived danger as more and more people began to participate, having no idea how to behave appropriately, and were rebuked.  Some groups appointed moderators to filter out inappropriate content.

Mailing lists

The next dominant pattern was that of email discussion via mailing lists. These were hosted in a definite “place” within an Internet domain, and any system administrator could set up a new one without stepping on anyone else’s toes. They were a bit harder to find, as there was no complete index of mailing lists, but this too was an improvement. If “newbies” couldn’t find your mailing list, they couldn’t post irrelevant content or otherwise misbehave. The barrier to entry provided a useful constraint on growth.

Mailing lists propagated messages more quickly, and people read email more frequently, so the pace of discussion increased.  A misstep on a mailing list could get you “flamed” within minutes.  Content posted to the list was only sent to individual members, and (unlike Usenet) was not copied to every ISP’s news server. The cost of a message, instead, was measured in terms of the attention of its readers. People worried about the signal-to-noise ratio on their mailing lists, and appointed moderators.  The FAQ pattern became decoupled from the discussion group as the web became a more convenient storehouse for this information.

Keeping up with multiple or high-traffic mailing lists required sophisticated software to sort mailing lists into separate folders, set appropriate headers on replies, scan or read many messages quickly, and so on.  Joining a mailing list required sending specially formatted control messages to a certain address.  People with only basic email capabilities found mailing lists difficult to use, and the people on mailing lists generally considered this to be a good thing, because they would probably just make a nuisance of themselves anyway (not knowing the appropriate etiquette).

Web forums

Forums filled a vacuum for people who wanted to participate in discussion, but found mailing lists and the associated software too complex and cumbersome.  Setting up a new forum was still moderately complex (a task for system administrators or webmasters) but anyone could participate in a forum with very little effort or technical expertise.

Without the underlying standardization of news or mail, forums vary widely in their capabilities, social patterns and content.  In general, they seem to be less linear than newsgroups or mailing lists, where the usual pattern is to scan all of the new messages since the last visit.   Visitors to forums instead look at “what’s newest”, or “what’s hot”, or search for a specific topic.

Forums have evolved fairly sophisticated mechanisms for managing the signal-to-noise ratio.  Very active forums tend to be moderated by visitors, who vote content (or other visitors) up or down according to relevance.  Participants maintain a visible and persistent identity within the forum (an avatar), in contrast to the simple email addresses used on Usenet and mailing lists, and forum content can be deleted (an impractical task in mail and news systems).

On the other hand, some types of noise (like “me too” messages) seem to be tolerated well enough in forums, while this is considered disruptive on a mailing list.  I suspect that etiquette varies widely depending on the capabilities and culture of the particular forum.

Blogs

The key advancement of blogs was that anyone could set one up, without any technical expertise, and make it available to everyone.  Originally, blogs weren’t discussion-oriented at all, which I think contributed to their slow growth early on.  Over time, through comments, trackbacks and web-based aggregators, they have grown some discussion capabilities.  Most people seem to read blogs through an RSS reader of some sort, though, and rarely visit the originating site at all.  Few blog posts attract more than a small number of comments or responses on other blogs, and these often receive a disproportionately small amount of attention compared to the original post.

The blogosphere moves quickly, but is largely incoherent as a whole.  Bloggers read each other’s content, and it influences their opinions and their writing, but there is very little direct response or feedback compared to the earlier examples above.  Etiquette is all but inapplicable to individual blogs, as readers can be assumed to be “tuning in” to the author’s intended content.  If they don’t like it, they simply don’t return.

It’s even more difficult for third parties to observe blog discussion.  It can be challenging to find responses, and inconvenient to display them on-screen at the same time.  The present infrastructure still seems much better suited to a singular “speaker” or “listener” role than to multi-directional discussion.  It will be interesting to see whether it becomes more discussion-friendly in the future.

Micro-blogging

I don’t think I fully “get” micro-blogging yet, as I haven’t used in enough myself.  It seems much more discussion-oriented than blogging, more like a very fast-paced mailing list than a small blog.  Short, unstructured discussions can be held, two-way or around a topic, and broadcast content is acceptable as well.  It would be nice to see more of a threading model, to make discussions easier to follow.

The real-time nature of micro-blogging seems to show obvious promise for discussion, but I’m not sure that the current formats support this very well.  I’d be interested to hear from more experienced micro-bloggers about this.

What next?

I think a “best of all worlds” online discussion system would combine:

• the clear threading of mailing lists
• the universal accessibility of blogging
• the mobile/real-time nature of micro-blogging
• collaborative filtering (something better than what I’ve seen in forums)
• a balance between “pull” (everything since I last checked) and “push” (whatever you want to tell me right now) models
• decentralized infrastructure, which can be hosted by anyone anywhere, and improved upon through open development

Taking things even further, I’d like to have all of the advantages of a face-to-face conversation (time-efficiency, rich expression, personal connection, well-understood etiquette, etc.),  and the advantages of online discussion (spanning long distances, interacting with large numbers of people, careful organization of thoughts, flexible starting and stopping, background and context).  As long as I’m making a wish list, how about a seamless connection between public and private discourse?

If you know of experiments happening in this area, please post them in comments here.

We all witness bad behavior at some point or other.  For many of us, the most common examples are provided by men misbehaving toward women.  Whether it’s in public at a conference, on an IRC channel, in an errant wiki page or two, or in a private conversation, how we respond to it is critically important.  This is particularly true where the behavior undermines the security or agency of another person.  Perhaps most of all, it applies where someone is speaking up about it.

If I’m standing with a group of people, and one of them behaves badly, I think that they’re a jerk.  If no one else seems to notice or object, then I start to wonder if they’re all jerks.  If someone speaks up, and is attacked, ignored or discredited, then I’m certain that I’m in a den of wolves.  Feelings like these are toxic to communities, and I don’t want anyone to have to feel this way in one of mine.

Managing one’s own behavior, although it is an essential first step, is not enough.  We must also critique the behavior of others, and signal to our peers that we object to bad behavior.  Furthermore, we must support those who speak up, particularly when they are doing so on their own behalf, or as a member of an underprivileged or under-represented group.  It may be difficult to speak up when you are an observer, but it is much more difficult when you are a target.  This isn’t about coming to anyone’s rescue, but openly accepting their objection and their right to voice it—even if it’s directed at you.

I will not trivialize the effort required to do this.  It is not easy to “break ranks” and stand as (or with) an objector. It is, however, often the right thing to do, and justifies the application of will and the taking of risks for the sake of integrity.  I will also not profess that I have always made the right choice myself.  Indeed, too many times, I have stood by, and I am ashamed for it.  I have made excuses for myself and rationalized my choices, explaining to myself why I couldn’t do what was right in a particular situation.

That is why, in the title of this article, I am addressing myself above all.  I am calling myself out, and calling on my peers in the Ubuntu community to do the same.  Don’t accept bad behavior.  Stand behind those who object to it.  Hold yourself and others accountable for the well-being of your community, and let others know that you are doing so.

There seems to be a significant misconception around Ubuntu 9.04, that it uses the (comparatively new) ext4 filesystem by default. It doesn’t. Ubuntu 9.04 still uses the tried-and-true ext3 filesystem by default. If you install Ubuntu 9.04 and use automatic partitioning, you will get ext3, which we consider to be the most stable option at this time.

ext4 is available as an option in the manual partitioning screen for people who wish to try out the latest stuff. There are some known bugs which affect this configuration, though, and filesystem bugs generally carry the risk of lost data, so this is not for the faint of heart.

A Linux Weekly News article included an error (now corrected) in this regard, and even our own Ubuntu Weekly Newsletter reprinted a BeginLinux.com article which implies that 9.04 carries the risks of ext4.

We included the manual ext4 option for people who want to help test this new technology, to help its developers isolate and fix any remaining bugs so that it can become the new standard for Linux and the default for Ubuntu. Many people are interested in it, have opted to try it, and written about the experience (which is probably how the misconception started).

In short, don’t panic. Ubuntu 9.04 isn’t any more likely to eat your data than any previous version.