Great UI-Design Piece

Bruce Tognazzini's "askTOG" column/blog has been a mainstay of the UI community for some time. Recently, he wrote a piece about an error Scott Adams made editing his blog. In his analysis, he nicely summarizes most of the crucial points about UI design in a very approachable manner.

For no other reason that I found it amusingly direct, here's a snipet:

The only effect of [confirmation] dialogs is to make the developers feel good: “The users may be screwing up, but we warned them, so it is their own fault.”

No, it isn't.

Any time your user loses any work, consider it your fault, and figure out how to prevent it from happening to anyone else.

Cheers,

Geoff S.

Alice's Resturant - Wiki Edition

Being a sporadic blogger (at least here - I do contribute stuff other places more often), it takes something special to get me off my virtual ass, recover my password, login, and type something into the editor here.

I just ran into something that qualifies: the Alice's Resturant page on the WikiWikiWeb.

Cheers,

Geoff S.

Firefox sucks (sometimes)

Just to make sure every Firefox "sucks-rocks-o-meter" gets this right: Firefox sucks

Honestly, this is mostly cathartic; I actually like Firefox - when it's running! I just lost a lengthy blog posting when Firefox locked up all my tabs and windows when I went off to research a link in another tab. I thought "save early and often" was an admonition for crappy software (like that from M$).

What's up with Firefox locking up everything when it's off talking with DNS and stuff? I'm a self-professed moron when it comes to things computing, but I would have thought that things like a DNS lookup might be done in a separate thread...

Serial Comma Rocks!

As a long-time advocate of the serial comma (i.e. the comma placed between the penultimate item in a comma-separated list and the "and" which separates the last two items) on purely syntactical grounds (it breaks the established pattern where the comma delimits items in the list), I've endured continual deletions of serial commas by well-meaning proofreaders and editors (and almost as ardently replaced them) for many years. It was with great glee that I found the a getitwrite blog entry, referencing two legal syle guides recommending the use of the serial comma (many thanks to cafeaulait.org's "recommended reading" section for the link!).

Some less defensible writing traits of mine are favoring run-on sentences and excessive (to the point of incomprehensibility) use of parenthetical comments, but that's another topic...

Cheers,

Geoff S.

"Just Fluff, No Stuff" (or something like that...)

This weekend is the "No Fluff, Just Stuff" conference in Minneapolis. Despite the overly cynical title to this entry (that's just my outlook on life, sorry...), I'm actually looking forward to the event. I went to one either last year or the year before and had a great time. A bunch of my co-workers went to the version in Milwuakee earlier this year and came back with some good information on what to attend (and what to avoid), so I'm hoping to have a particularly enjoyable weekend.

Back to cynicism for a moment. What's up with the "IN THE SPOTLIGHT" block at the bottom right-hand corner of the NFJS home-page? Today it has the Ron Popeil-esq teaser:

What if you never had to write a user interface again? What if you could simply expose your business objects directly to the end user? How would this affect your productivity? Is this even possible?

Following the link takes you to a page totally devoid of any real content (containing just a short blurb about Dave Thomas: "is recognized internationally as an expert who develops high-quality software--accurate and highly flexible systems."). WTF? Not much "stuff" there...

I guess this won't make the NFJS "IN BLOGS" section of the Reviews page...

Cheers,

Geoff S.

The SWT Table is decadent and depraved

After banging our collective heads against the limitations of the SWT Table and it's associated so-called useful classes, I was gratified to read: SWT Happens and Suprised By The State of SWT/JFace Tables along with an article following a popular naming convention: [Rant] SWT Table API is pure CRAP (dev.eclipse.org login required).

If you're looking for a even bigger CF to help you with your "spreadsheet-like" SWT tables, try: http://www.swtworkbench.com/solutions/ed.shtml

Enjoy!

Cheers,

Geoff S.

(apologies to HST: "The Kentuky Derby is Decadent and Depraved")

JPred - interesting way to specify multi-methods (and more!)

Interesting talk by Todd Millstein on his "predicate dispatch" language: JPred
http://www.cs.ucla.edu/~todd/jpred/

On the surface it's some syntactic-sugar for "instanceof, if-the-else ladder", but it's quite a bit more sophisticated. I'm not yet considered it in enough detail, but it looks like a useful tool (FWIW, it was interesting enough that I downloaded it during the subsequent talk).

Interesting comments from the audience & speaker about Subtext and example-oriented programming

After Jonathan Edwards talk on "Example Centric Programming" one of the audience members made a (IMO) fascinating comment. Invoked the metaphor that "code is like DNA" and suggested that the examples/Subtext can be considered as RNA in that metaphor. The role of RNA is the transcription of DNA data into specific protiens, enzymes, etc. (aside: mention was made that the RNA transcription mechanism is Turing-complete - interesting...).

The speaker noted that patterns are really templates, which map into examples very nicely. This observation opens th epossiblity that example-driven languages might have very naturally integrated pattern support.

Onward!: "Programming By Example"

Onward!: "Programming By Example" - Wed. AM

Part 1: "E.G." a tool for building/running examples written in Java under Eclipse. Very interesting new technique to allow people to incrementally write code. Conceptually, very similar to TDD (as the speaker points out), but with perhaps with a more direct connection between the programmer's intent and the app-code and test. Some advantages (maybe) in user-code interaction. Brings an interpreter into the Java development cycle.

Part 2: Speaker's question: "Is this the right way?" Tools on tools on languages, with interpreters, tool-tips, and all manner of "stuff". Simpiler languages? (i.e. examples build into the language) => SubText.

Subtext looks like an interesting graphically manipulated data-flow'ish functional-programming language with protoypes. Much to ponder before passing judgement. Sound bite: "Decriminalize copy & paste" - Cut&Paste is the central mechanism for adding new code.

Tuesday...

Keynote: gotta admire the chutzpah of someone to turn an OOPSLA keynote into a product demo. Ugh. I walked out in the middle...

AM sessions: "Unified Garbage Collection...": Pretty interesting. "Mock Roles not Objects": the JMock guys. Good tutorial on why mocks/stubs/stand-ins are important in TDD. Stuff I learned and internalized a couple of years ago, but based on questions afterwards it seems like lots of other folks are just getting on board with the idea.

PM stuff: Ward's talk was excellent. "Methodology Work is Ontology Work" by Brian Marieck in the Onward! session was a fasctinating mix of philosophy-of-science, vetinary education, and discussion about perhaps why XP is gaining popularity.

Monday Poster Session - what a kick!

Levi and I stood in front of our poster during the Monday welcome session. We were actually expecting a pretty sparse evening. I'm assuming that OOPSLA participants are just extremely polite because both Levi and I were busy talking with people all evening. The highlight of the evening was when Ward Cunningham stopped by and talked with Levi for quite a while. I didn't notice too many nametags, so I may well have talked with the entire Gang-of-Four (but I doubt it). I was really impressed with the interest people had; most folks hung around and asked pretty interesting questions. The JMock guys from ThoughtWorks stopped by. They've got a framework for generating mocks in-situ that has much of the same intent as PseudoClasses.