LarryRoth.net

There is a helpful post from the folks at “The Authentic Fake” that has various MacBook Pro shortcuts. I found several I didn't know about. My favorite was:

10. Cool shortcut combo
To finish off, I’l give a cool (relatively) shurtcut combination. Clicking Control+Option+Command+8 will turn the screen in some sort of high contrast mode, but more like an X-ray screen![1]

Pretty cool!

David Churchville has a nice post that poses the question:

Does the agile software development mantra of “Do the simplest thing that could possibly work” lead to functional, but unremarkable software? [1]

He continues by pointing to an example of when functionality might be chosen over usability.

It’s an interesting thought, but I feel the main benefit of agile development is to test early and test often. This testing, if done properly, should help to find usability issues. And then, agile development really helps out by not only finding the issues, but by fixing them before they are released.

Said another way, in traditional software development, it seems all to common that when the deadline slips, one of the items that is often discarded is usability testing. Any early testing—while perhaps not true usability tests—will help out.

[1] Is Agile Development Killing Usability?, http://www.extremeplanner.com/blog/2006/07/is-agile-development-killing-usability.html, posted July 26, 2006, viewed, July 28, 2006

Fintan Darragh posted a tip for a quick way to change the toolbar’s icons and text.
From the article:[1]

Secondly, and this is certainly a hidden feature within Mac OS X, when you Command-click on the top-right-hand button which appears in windows, you can set the toolbar’s icons and text to different sizes and combinations.

Well, he certainly is right about it being a hidden feature—I don’t think I would have stumbled across this functionality.
[1] Mac Tip: Window Dongle, http://www.dech.co.uk/2006/07/mac-tip-window-dongle/, posted July 27, 2006, viewed July 27, 2006

I have been working on setting up new project management processes in our company and one of my colleagues passed along a link to Trac. After playing around with it for a couple days, I must say that it is quite an impressive app. It’s easy to setup (apt-get in Debian), easy to tweak (via it’s command line admin tool and even by tweaking the python based templates and code), and easy to use.

Trac combines a Wiki, bug tracker, and Subversion source repository browser in one Web app. It’s a nice mix of keeping project knowledge in one place as well as viewing and managing changes. I really like the simplicity and customizability of the tickets in the bug tracker.

Not only is initial setup a breeze, but ongoing project setup is easy as well. I was able to quickly add a perl script that not only setup the Trac project–with our particular customizations, but also created the source repository, and added the Trac project to the Apache config file.

If you are looking for project management support for your software projects, you should check it out.

Ebrahim Ezzy and Richard MacManus over at Read/Write Web put together a nice summary of Search 2.0 and how it compares to traditional search. They define Search 2.0 as the third generation search technologies and that they are:

…designed to combine the scalability of existing internet search engines with new and improved relevancy models; they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities that make information more productive.

[1]

In their post—which by the way is part 1 of a 2-part series—they examine 5 different Search 2.0 companies. They point out a key feature of each and also compare it to traditional search (e.g. Google).

What was interesting for me is that when I went to try the search engines they mentioned, 2 of them were down, 2 required a setup process, and several of them were simply returning Google results with their own "value add", such as clustering.

Does having to train a search engine or relying on an unknown communities’ rankings really further search? I have always felt that technology should help automate things for us, not require more effort on our part. The lure of Google was—and is in my opinion—the simple page with only one thing to do, enter a search term. And you could be reasonably certain that you would find the information you wanted within the first couple result pages.

I must admit, I did like some of the added functionality in Clusty—although as part owner in a Branding firm, it feels wrong to mention anyone named Clusty.

I noticed a couple of comments to the post with which are worth highlighting:

• …I think the people that want all this Search 2.0 nonsense are the people having trouble coming up with a comprehensible list of phrases to describe what they’re talking about…
• Part of being a good search is… uptime! (and speed).

Check out the post, it’s a good read.

[1] Search 2.0 vs Traditional Search, http://www.readwriteweb.com/archives/search_20_vs_tr.php, posted July 20, 2006, viewed July 21, 2006

An interesting news item disclosed a discussion between Tim Berners-Lee, inventor of the World Wide Web, and the Director of Search at Google, Peter Norvig. Mr. Norvig presented what he felt are very tangible road blocks to the usefulness of the Semantic Web. His concerns are:

• Incompetent content providers
• A lack of support for standardization
• Deliberate deception[1]

These are all valid points, and specifically, I think we are already seeing quite a bit of deception. Some content owners deliberately provide false metadata in order to improve their ranking. Tag spam and keyword stuffing are just a couple examples of this.

What was Berners-Lee’s response? He feels that the Semantic Web can increase relevance by identifying the content’s originator and why the content can be trusted.[2]

[1] Google exec challenges Berners-Lee, http://www.zdnet.com.au/news/software/soa/Google_exec_challenges_Berners_Lee/0,2000061733,39263931,00.htm, posted July 19, 2006, viewed July 19, 2006
[2] ibid.

There is a good article at IBM Developer Works about using the FindBugs static code analysis tool. Test cases are important, but when you inherit code, a good static analysis code tool can really help you out. The article is a couple years old, but still quite relevant. I thought this comment summed it up nicely. From the article:

Much of the knowledge of how a class works (or is supposed to work) evaporates shortly after it is written, in the form of implicit assumptions, invariants, and expected use cases that are clear in the developer’s head but never get written down in the form of design notes, comments, or documentation. Existing code is always harder to work with than new code.¹

Even with strong test cases, you can still miss more complex bug patterns. FindBugs can help you out.

This article at StreetFish provides a summary of the ongoing research with brain implants at Brown university. It shows some (early) exciting results for enabling people with disabilities and perhaps increased HCI for everyone.

From the article:

Braingate it its name, and it was created by scientists at Brown university and already it has been used to give computer control to a 25-year-old Massachusetts paraplegic, allowing such functions as controlling the tv, his prosthetic arm and drawing shapes using the cursor.
The system requires an implant of a 4mm square chip into the motor cortex, connected to the brain by 100 electrodes, each measuring the electrical signals of the cells they connect to. The signals are then calibrated by thinking of particular movements, and the computer works out the rest.

I found additional information about this project on the Brown Web site.

• Controlling Movement Through Thought Alone
• BraingateTM *: Developing a Device to Provide Motor Impaired Patients withA Novel Neural Output (1.2mb pdf)

If you are in the market for a new Web server, Sun has announced a new 4U Web server, the Sun Fire x4500 that you may be interested in. WARNING: it is only capable of housing 2 dual core x64 AMD processors and can only accommodate 24GB of storage. Think that will be enough?

Ok, to be fair, it does start at ~$33,000. That’s a little more that I paid for our last XServe, but—MAN—it sure looks sweet…

Eyal Lupu has a nice (short) primer on Web service support in Java 6. It’s an easy read, and it really gets you excited about the possibilities of Web services. Since the concept of Web services was first mentioned, it seemed like an ideal way to foster code reuse and componentization (in the Web domain)—ideas that are core to Java development.

However, in the Java space, there has always been a fairly steep learning curve and/or a significant amount of code and configuration that went along with using Web services. But now, thanks to the power of annotations and Sun listening to the Java community, it is now quite simple.

Here is a before and after snippet. See if you can spot the difference:


Before

package ws;

public class CalcWS {

   public int add(int a, int b) {

     return a+b;

   }

   public float div (int a, int b) {

     return a/b;

   }

}

After (from the article)

package ws;

import javax.jws.WebService;

import javax.jws.soap.SOAPBinding;

@WebService

@SOAPBinding(style=SOAPBinding.Style.RPC)

public class CalcWS {

   public int add(int a, int b) {

     return a+b;

   }

   public float div (int a, int b) {

     return a/b;

   }

}

See the difference? Just 4 lines of code two imports and two annotations. Pretty cool! Now if we can just get Tor Norbye to abandon Creator and Project Semplice and focus on enabling the following annotation:

@TurnThisObjectIntoACoolWebAppLikeRailsDoes

That’s not too much to ask, is it?