Via Russell Beattie: LifeBlogger is an interesting little blog app developed by Erik Thauvin that enables Nokia Lifeblog users to post items from their Lifeblog to their real blog - via the Blogger API.

Atom API guide

Ben Hammersly has posted a draft chapter on the Atom API from a new book on Movable Type: A guide to using the Atom API implementation within MT3.

Are you blogging at work?

Weblogs have been appearing all over the intranet where I work and I keep hearing about behind-the-firewall blogs at other companies. I've heard of folks who are blogging their weekly status reports, using blogs as personal notebooks at work, and creating newsfeeds for CVS, bug trackers, and other systems. Still, I get the feeling that most weblog-at-work deployments are experimental - just grass roots efforts to put weblogs in place to improve collaboration and communications and to see what happens. Am I wrong about this?

Traction Software sells "enterprise weblog software" and has put together an interesting list of weblog use cases, summarized below. Are you doing any of these uses cases? Are you blogging at work and, if so, what use cases are you finding most useful?

Traction Software's weblog use cases:

  • Personal Notebook
  • Scientific Research
  • Corporate Communications
  • Program Management
  • Product Management
  • Operations Log
  • Community of Practice
  • Internal News
  • Exception Reporting
  • Human Resources
  • Public, Investor, and Customer affairs
  • Account Team Communication
  • Law Firms and Litigation Support
  • Law Enforcement

Roller roundup.

I've been so busy with other projects, my day job, vacation, and being a Dad that I haven't been able to do much Roller development at all. That's a shame, because we are so close to a Roller 1.0 release. We also need a bug fix release from the Roller 0.9.8 branch, but that is probably not going to happen until things slow down a bit. My development efforts have been at a near standstill, but I've been keeping an eye on Roller. Here's a round-up of recent Roller developments.

John Hoffman of Sun has been giving some thought to the Roller main page and has introduced an interesting sortable interface for John laments the fact that Roller supports only MySQL. With a couple of code changes, Roller 0.9.8 can be made to support PostgreSQL. Roller 1.0 will support PostgreSQL right out of the box.

Lance has been working on improving Roller's search interface and making numerous other fixes.

One blogger tried to install Roller on JBoss and didn't get very far, but I'm not sure why because he does not provide any specifics. Andy Depue is also working on getting Roller to run on JBoss, and trying to get some help on the mailing lists.

I'm not sure why Roller on JBoss is such a pain, but blogger Jeff Sheets observes that Roller on Tomcat just got a whole lot easier thanks to the Roller Demo bundle. I called it the Roller Demo bundle because I'm concerned about supporting HSQLDB for "production" blogging. Should I be?

A couple of JRoller users have complained about the Roller editors recently, so when I heard about the FCK editor I took note. I'm playing with it now in my Roller sandbox, but I'm a little discouraged that it does not support Safari.

It's kinda cool to see a Business Week about a Roller blog entry, even if it doesn't actually mention Roller.

Annotations are not for configuration

Thank goodness for that.

Atom and Cool URIs: dogma, idealism, expediency

Excellent summary of both sides of this important argument.

Tuesday at the RTP-WUG: a big chunk of Webshere for free.

IBM has agreed to release to the open-source Eclipse project key tooling components that until now were exclusively part of the WebSphere Studio product line.  This move is being keenly anticipated as it will greatly enrich Eclipse Development.  Jim Zhang will talk on the new Eclipse Web Tools.
For more information, see the Research Triangle Park Websphere User Group's Web site. Bill Dudney has documented some of the contributions on his Eclipse Live blog.

Java continuations

"the JVM could very easily support continuations"

XML on the Web Has Failed

The Bile Blogger could learn a thing or two from Mark

Atom in a Nutshell (as of Atom API 0.9)

I've been working to learn Atom as quickly as possible and so I've been reading the developing format and API specifications, following the mailing list, watching the issues list and experimenting with Atom4J and the Atom API implementations. Atom is pretty simple, so I'm going to take a stab at explaining it in a couple of paragraphs. To follow this explanation, you'll need to have a basic understanding of the Web including both HTTP and XML.

Atom is a new standard for publishing on the Web. Atom defines both a data format for Web publishing and a protocol for interacting with a Web publishing system, such as a Weblog server, a Wiki, or a Content Management System.

The Atom format is an XML based format. A file that is formatted according to the Atom Format specification is known as an Atom feed. The top-level XML element is <feed>. An Atom feed is made up of a collection of Atom entries, represented as XML <entry> elements. Each entry has a title, a summary, text content, and other attributes. FYI, an Atom formatted file must be served as application/atom+xml or it won't work.

The Atom API, a Web services protocol, is defined in terms of HTTP URIs and HTTP verbs such as GET, PUT, POST, and DELETE. The Atom API specification defines three classes of URIs: Feed URI, Post URI, and Edit URI.

A Feed URI provides a way to get Atom entries. If you do a GET on a Feed URI, you will receive an Atom feed. If the server knows who you are, this Atom feed may contains a series of <link> elements, just like the ones in HTML, that point to the Feed URIs for each of the Blogs, or Wiki namespaces or Content Management document collections, that you are allowed to edit. If the server knows which Blog you are requesting, then the Atom feed will contain the latest entries from that Blog. In that case, the feed will also contain the Post URI for the Blog, Feed URI for fetching the next and previous batch of entries, and each entry will contain an Edit URI for editing that entry.

A Post URI provides a way to create new entries on a server. You create a new entry by using an HTTP POST to write the that entry to the server as an XML <entry> element.

An Edit URI provides a way to edit an entry on a server. To get a specific entry, you do a GET on the Edit URI for that entry. To update an entry, you POST the <entry> XML representation of that entry to the Edit URI for that entry. To delete an entry you do a DELETE on the entry's Edit URI. The use of the PUT and DELETE verbs is problematic for some devices with limited HTTP support and a workaround is being sought.

Currently, authentication is done via WSSE and each Atom API request must carry an HTTP header that contains a valid WSSE token. Mark Pilgrim explained this in his column and Claud Montpetit illustrated it with a nice and concise Java example. However, WSSE seems to have fallen out of favor and may not be included in a future version of the spec.

Did I get something wrong? Please correct me. Leave a comment.

Dive into Python... yum

I ate a big ole slab of Dive into Python cake today (thanks Mark). The cake is probably gone by now, but you can buy the book on Amazon so get to it!

Are Blogs Ready for Primetime?

"blog readers are older and richer than many people suppose"

Blogs Proving Effective Team Integation

"Expect to see more blogs at your site among developers, sys admins and system managers"

Bloggers aren't journalists... really?

"journalists sound like a bunch of insecure cry babies"

RSS Growing Pains

InfoWorld needs to work on their plumbing

Longhorn Through the Open Source Lens

"Avalon is the next Active X"

Atom4J, Subversion, and Mac OS X

Happy news today: Lance accepted me as a committer on Atom4J (thanks Lance). I'm going to be helping him to keep Atom4J in line with the developing Atom Format and API specs. Atom4J, by the way, is not just an Atom Format parser, it is also a Atom API server framework. If you want to add Atom API support to your Java-based Web application, all you have to do is to extend the abstract org.osjava.atom4j.servlet.AtomServlet and provide implementions for a handful of abstract methods. I'm interested in adding Atom API client capabilties to Atom4J and I've already made some progress on that front.

Now, on to the less happy part of the story. Atom4J is hosted at in a Subversion archive. This is my first experience with Subversion and getting set up took some time. I ran into some problems on my platform of choice, which is Mac OS X, but I eventually found some level of success. Here's what I did:

  • I use Eclipse 3.0 so I installed Subclipse, the Subversion plugin for Eclipse. Unfortunately, I could not get Subclipse to support SSL. I was able to get Subclipse to checkout Atom4J by using HTTP protocol instead of HTTPS, but, because of this, I could not do a commit via Subclipse.

  • By default, Subclipse uses it's own Java-based Subversion client implementation, but Subclipse can also be configured to use a command line verison of Subversion if one exists on your computer. So I decided to get the command line version of Subversion.

  • I used Fink to install svn with fink install svn. Unfortunately, the version of svn that it installs does not support SSL and is therefore useless to me.

  • After some googling around, I found that what I needed was svn-client-ssl, so I tried to get that from fink. Fink told me that I needed system-java14-dev. After some more googling, I found that I needed to re-install the Java SDK so I did. I tried find again and this time Fink told me that it was unable to upgrade db42-ssl-shlibs because it conflicted with db42-shlibs. So I uninstalled db42-shlibs and ran Fink again. Fink cranked away fetching and building for about 45 minutes and died with some other error that I can't remember. At this point, I gave up on Fink for the day.

  • Finally, I found a blog entry that pointed me in the right direction. Bill Bumgardner wrote about Martin Ott's pre-built Subversion binaries for Mac OS X. So I downloaded and installed those.

  • I went back to Eclipse, configured Subclipse to use my new SSL ready command-line version of Subversion and presto magico, after several hours of flailing around, everything started working.

UPDATE: In his O'Reilly weblog Brian Coyner shows how to build Subversion and Java bindings on Mac OS X. I wonder if his instructions result in an SSL capable client.

Roller tip: Styling Blog entries according to Blog category

Scott Hudson asks "Is there any way to modify Roller to insert the category information around each entry, so we can do additional CSS styling?" The answer is yes, with a little page template hacking you can do almost anything to your Roller Blog.

To add a div around each entry with a style class that varies with the entry category, you need to add some code to your Blog's day template. Find the foreach loop in your day template that iterates over each day's Blog entries and add a div tag around the contents of that foreach, as shown below. To give each category it's own category, you use a Velocity expression to get the category name $

#foreach( $entry in $entries )
   <div class="cat_$">

     ... display one weblog entry, code removed for brevity ... 


Now that you've got class, you can add some style. You can do this by adding CSS to your main Blog page or to your CSS template if you have one. For example, if you have categories General, Music, and Java and you'd like ot put a red border around your General entries, a green border around your music entries, and a blue border around your Java entries, you would add the following CSS:

.cat_General {
   padding: 3px;
   border-left: 5px red solid;
   border-right: 1px red solid;
   border-top: 1px red solid;
   border-bottom: 1px red solid;
.cat_Music {
   padding: 3px;
   border-left: 5px green solid;
   border-right: 1px green solid;
   border-top: 1px green solid;
   border-bottom: 1px red solid;
.cat_Java {
   padding: 3px;
   border-left: 5px blue solid;
   border-right: 1px blue solid;
   border-top: 1px blue solid;
   border-bottom: 1px blue solid;

Below is a screenshot that shows what this looks like (and yes, I'm aware of the border mix-up on the Music post):

screenshot showing different styles for each category

« Previous page | Main | Next page »