Not having felt enough abuse, Atwood revisits the Angle Bracket Tax
I just can't escape the notion that he's attributing his symptoms to the wrong problem. That is, I agree with him that that if you are trying to figure out an XML file, that there is a problem. I just don't accept quite so readily as he the idea that the problem is the XML.
Parallel notion - when I find myself trying to figure out java byte codes, I have as yet somehow managed to avoid the conclusion "Wow, java byte codes suck - they really should be replaced with something more easily consumed by humans". I've likewise failed to reach that conclusion when starting at windows executables, or jpeg files, or zipped archives.
I'd instead call attention to the facts that (1) the normal flow for a use case should never involve presenting "raw" XML to the user, and (2) the use case should include the state of mind of the user, requiring that it at least be addressed by the solution. Part of the problem with XML configuration files is that the user is already unhappy (if the software were doing what he wanted, he wouldn't be trying to modify the configuration) and then you make that worse by filling his nose with angle brackets.
Does your software distinguish data that represents opinions from data that represents facts?
Similarly, with logs, we can improve our search results: if we know that people are clicking on the #1 result we're doing something right, and if they're hitting next page or reformulating their query, we're doing something wrong.
http://www.wizards.com/default.asp?x=mtgcom/daily/mr11b
Filed for future reading
Comes courtesy of Excel, which was proposed as a model for the "correct" behavior of software I'm currently delivering.
Open a new Spreadsheet in Excel, and set the value of the cell A1 to
=MIN()
Min is a function that requires one or more arguments, and Excel even reminds you of this once it recognizes that you are intending to call min. But for this exercise, we are testing the behavior of excel when there is an invalid formula in the spreadsheet edit control.
First test: Attempt to save the document. You get an error message: the formula you typed contains an error. Yup, it does. Bravo, have a biscuit.
Second test: Attempt to close Excel (File.Quit). Oops, you can't quit: the formula you typed contains an error. And then, just to rub it in, you get a second error message: Cannot quit Microsoft Office Excel. And this message is so important that it sits on the top level of the windows hierarchy; when you tab to another application, the error message sits in the foreground. So I can sit here merrily blogging with the message box sitting in front of my browser.
Third test: Attempt to close the spreadsheet, using the close button on the document window itself (x in the corner for the win). <fx: crickets> No error message at all, no behavior, not even a dial tone.
If you maximize the Navigator window or any process window in Oracle Workflow Builder, you will not be able to access the menu from your keyboard using the Alt key.
Uh huh. Here's your sign!
I'm putting together a bunch of bar charts to illustrate some data. How should I choose the colors I use?
One resource which I think will help some is VisiBone's color chart. But I'm not sure whether I want matching or constrasting colors.
If you are a windows user:
1) Launch notepad
2) Invoke the find dialog
3) Bring some other application into the foreground
4) Bring the notepad session into the foreground. You now see notepad, with the main window active and the Find dialog visible but inactive.
5) Without using the mouse, dismiss the Find dialog
Why are you supposed to guess that is what works?
Tukey's suspended histogram, lifted shamelessly from Some Graphic and Semigraphic Displays.
Jakob Neilsen: "I've long been predicting that Internet Explorer 8.0 would be the first good Web browser"
While I can see as Microsoft might finally get it right on their eighth try, the idea that they will be the first to get it right seems like a reach.
Bruce Tognazzi normally is a pretty reliable guide in the jungle of user interfaces. He is also the subject (or perhaps I should say target) of today's challenge.
Look at the results of this search for recent columns. At this time, the search page lists the home page, and four articles.
The challenge is to navigate from the home page to each. The first is a gimme: From your starting point at http://www.asktog.com, you can click directly on the link to Multiple Mistakes Drown Interface.
Find a route from the home page to each of the other three. I consider spidering the website to be cheating (although tempting), but using the resources available in Google is allowed.
I don't promise that there is a legal solution - got fed up looking.
A 404 Page. Very pretty.
But I think it fails - the first pass through, I didn't recognize it for what it was (It appeared to me to be a blog entry titled 404 - yes, I didn't read it carefully, that's the point).
Communicate first.
Over the past few months, I've invented some interesting ways of displaying data to a computer user - mostly mechanisms for adding multiple dimensions to graphs.
This past weekend, I read Tufte's trilogy, and threw all those ideas away.
Instead, I'm looking at the way my user agent displays web pages. One of the things that I noticed is that my priorities are completely backwards.
Headings, H1 H2 and so on, are not content - they are structure. Like the lines on a sheet of graph paper, they should be relegated to the background.
Think of a contour map. One color, in different shades. Grey is the easiest, so that's where I've started. Very light grey background (CCCCCC), with darker structure elements (808080). I'd rather be using earth tones - the earth/sea maps in Tufte were visually persuasive.
Also, I came away with the notion that using colors that H. sapiens has been using for survival the last mumble thousand years are probably a good choice.
It slowed me down somewhat that I misread the cascading rules. So I lost motivation, thinking that J Random Webhack was going to be able to insist on his colors. But that's taken care of by making my rules !important. Fair enough.
Then comes the question of whether IE correctly implements the standard in question.
"In games, the software designer chooses the goal and the users job is to achieve it. In (eg) a word processor the user chooses the goal and the designers job is to help them achieve it. The linear tutorial form doesn't work because the designer can no longer predict what the user is trying to achieve."
Brian Ewins, at the Wiki.
I don't expect to write a lot about UI, because I'm not an expert. I know somebody who once worked with an expert, and I do a lot of reading on the topic, but that's as far as it goes. I wouldn't be at all surprised if I end up ranting a lot about UI, however, so into the category bin it goes.
