Doug's Inner Net News

Reification of Attention

Most of us have never thought of our "attention" as a tangible thing. Advertising executives surely do, though. And so do a group of people led by Steve Gillmor who created a non-profit organization called AttentionTrust.org.

According to the thinking of this group, your "attention data" is valuable, and you should have the right to control it. Attention data is the record of your past online activities that indicate where your attention has been. Attention data has predictive power, which is why it's valuable to advertisers -- they want to know what will grab your attention and what you will immediately dismiss. Above all, AttentionTrust.org believes you should have the right to control how your attention data is used, and even negotiate the terms for its use.

Recently, one of the directors of AttentionTrust.org has launched a new company called Root Markets, with plans to build a business around attention data. The concept of the business is abstract, but the basic idea is to act as a broker of attention data.

This all seems incredibly abstract to me. And so, I have to ask this question: Can AttentionTrust.org and Root Markets gain any broad acceptance if the "attention" of the average person does not include an interest in attention data? The point is this: the ideas of "attention" are far too abstract to be of any interest to the average person.

Is Control-S Necessary?

How many decisions do we make just because everyone has always done it that way? You write a desktop application, and you add a File menu, and that File menu has a menu item labeled "Save," and you bind Control-S to that "Save" menu item, right? Of course, we've always done it that way!

Why?

Control-S, and "manual save," are so natural, we never stop to think about whether it really makes sense. But if you do stop and think about it, I think you'll agree with me that we really don't need it.

A world without Control-S is a different world. When you make a change to a document, the application stores that change immediately. The upside is that you never lose your work because of a crash or power failure. At first, it may seem like the downside is that you can't undo, but that's not the case. All modern desktop applications support multiple undos. The downside really is just that you may have to perform a lot of undos to get back to the place you want. So, in the end, it comes down to what is more important: never losing any work or having to do multiple undos to go back to a previous document state. I think it's time to rethink Control-S. A good alternative might be a "Save a Copy" menu item, which allows you to save a snap shot, or a "Create Checkpoint" menu item, which saves a named checkpoint somewhere, possibly even in the document itself.

For more reading, check out Michael Feathers article entitled I'm Tired Of Saving! Michael suggests that the IDEs used by software developers could not only do away with Control-S, but could also run test suites in the background and provide instant feedback about any problems in the code. This is a step beyond just checking the syntax in the background.

Oreilly's Open Books

O'reilly is a popular publisher among techies, and deservedly so.

One of the valuable services O'Reilly provides is the Open Books Project. This project provides books online for free. There are many reasons O'Reilly has made these books free. Some books have outlived their economic viability, meaning that O'Reilly would not recover the expense of another printing. Some books are free because the author(s) negotiated a deal that required that they be free online. Some books are free for still other reasons. Even though these books are free, you may buy the printed book versions from O'Reilly.

Some of the most interesting books include:

• DocBook: The Definitive Guide
• Version Control with Subversion
• Creating Applications with Mozilla
• MySQL Reference Manual

The Open Books Project web site is worth knowing about, as it can save you a few bucks. However, if you are like me, sometimes you will still pay for the printed book because you prefer reading on the living room couch, or on a plane, or someplace other than a computer.

Where Do You Find RFCs in HTML?

Have you noticed that some RFCs are available in HTML and plain text?

Every few months, there is a discussion on the IETF general mailing list about the format of RFCs. The result is almost always the same: we'll continue to publish plain text RFCs, but we'll keep an open mind.

Well, a few years back, Marshall Rose and friends created a tool named xml2rfc that converts a certain XML format into the standard plain text RFC format. In addition, the tool converts the XML to HTML. (There are similar tools that convert from nroff and from Microsoft Word into plain text RFCs, but not HTML.) This made the text-only side happy, and it allowed the plain-text-sucks side to be happy, too.

The problem is, not all RFC authors use the xml2rfc tool, and they continue to create only plain text RFCs. For those of us who prefer to read HTML in a browser instead of plain text in a text editor, it's helpful to know which RFCs are available as HTML, and where to find the HTML version. It turns out that the developers who created xml2rfc wanted the RFCs linked together, so they created a database for RFCs. The xml2rfc tool checks the database to find the link targets for referenced RFCs. You can check that database yourself if you want to see if a particular RFC has an HTML version. You can get the database as a zip file or just visit the directory.

As a side note, there are a lot of RFCs converted to pseudo-HTML, which are available at http://www.faqs.org/rfcs/. These "HTML" RFCs are basically text with <pre> tags around them.

Do You Speak XML Schema?

XML Schema is commonly used in technical documents, including Internet RFCs, to express the syntax for XML in various documents and protocol messages. It's also the basis for SOAP, WSDL, and UDDI.

I have felt for a long time that XML schema is ugly and useless. It's ugly because it is so difficult to read. Most published documents, even Word documents, don't publish the XML Schema with syntax highlighting, which would certainly help. But if it's hard to read, at least it's machine readable, right? True, but it's rare for anyone to use validation tools with a published XML Schema. The published schema is often just for human consumption.

Now, however, I feel that my feelings have been affirmed. The following is quoted from an email by Tim Bray:

Many people in the XML community - I hesitate to say consensus, but would claim at least a plurality - feel that [Relax NG] is substantially technically superior to both DTDs and W3C XML Schemas. ... W3C XML Schema is a profoundly bad design which exhibits very poor comprehensibility and interoperability.

So, at least in Tim's opinion, Relax NG should be the way forward.

Read Tim's complete email in the archive. It's a short but interesting read.

Who Should Pay Whom?

It's been around the blogosphere. SBC's CEO Ed Whitacre (pictured) calls the application providers "freeloaders." See, for example, the post at Techdirt. The idea is that the application providers should pay the access providers -- SBC for instance -- for access.

But isn't this backwards from the other popular distribution model? What about cable TV? Imagine Comcast telling Disney that it must pay to use Comcast's cable lines. No, in the cable TV business, the cable company pays the content company. Comcast's cable lines have no value without the content. So, perhaps someone should ask Mr. Whitacre why his situation is any different.