alt+f2 and type application (my favorite launcher is katapult and the keys for that are alt+space)
and in general, hotkeys it up… win+w is web browser, win+f is file manager, win+e is text editor, and win+d is doom3 (always a pleasure to accidentally open it between the two and get side tracked for an hour :)”

That is nowhere near as efficient and easy as it could be. There is a much better way, but I will reveal all soon…

I’m just amazed nobody has thought of it already – it shows how unable to think outside the box the entire world of computer users is…

Using folders for your music is a GOOD idea, unless you’re one of the masses of idiots who pretends to ‘like’ music, and has to pretend to ‘want’ to listen to ‘rock’ today, or ‘jazz’ tomorrow, thus needing iTunes and its bloody irritating ‘filing’ system to find them what they pretend to ‘want’ to listen to.

I know where every single song in my 4,000 track MP3 collection is, all filed in folders, either by artist, or about 10% in ‘genre’ folders, such as ’80s’, ‘Love’, ‘Dance’, etc. I don’t want a long list of 4,000 tracks, thank you, that’s for idiots.

Your ‘long list’ idea doesn’t work when you are dealing with thousands of files. I remember the position of some of the 4,000 songs SPATIALLY, i.e. I know it’s in the first, top level folder,
A-F
not in
G-M
or whatever, and then within
A-F
it’s within
Dance
etc.

Your ‘huge list’ idea is stupid. Microsoft’s ‘Ribbon’ is equally stupid.

Spatial memory is where it’s at. I will reveal all very soon…

http://www.youtube.com/watch?v=ggoTQqNXOQs

Perhaps some of these questions can be clarified by thinking about the interaction between user and computer as a sort of ‘communication event’ embodying a ‘language’. Implicit in a lot of discussion is the idea that there are two kinds of language of interaction: textual as implemented via the keyboard, and visual as implemented via the mouse. Similar can be said about the information/files stored on the computer: that there are two basic kinds of stored data: document files of various kinds (readable text, source, etc) and image files (everything else — applications, pictures, videos, music, etc).

As such, we can reformulate our questions/thoughts about effective UI design as a question about language: what is the most natural language for human interactions? Is it a visual language or a textual one?
Note that these last questions are not about computers per se — we could be considering a mutual dialog between two people or people and community as much as with any sort of hardware.

However, in observing actual interactions between real people, we notice immediately that while there are lots and lots of very well established textual languages (English, Spanish, French, Hindi, ect, etc), there are very very few visual languages in common usage. Examples of visual languages include traffic lights, some types of road sign, symbols indicating bathrooms, danger, and various sorts of mapping symbols.
Beyond that, it is fairly much admitted among professional linguists that a true visual or “symbolic” language is something of an oxymoron at best. What visual language elements do exist in common usage tend to be brief, sparse in meaning content, lacking in connotative richness. Moreover, even the most well developed symbolic/visual languages tend to have vocabularies several orders of magnitude smaller than the average vocabulary of most textual languages.

The simple fact of the matter is that human bodies do not have an image projector built in at birth. While we can easily receive visual images at high bandwidth, we cannot express images with similar bandwidth. We do, however, have a voice — an auditory projector that can generate information as well as ears to hear with. Hence, it is *much* more natural for human beings to communicate complex information via voice (textual) than it is to communicate (interactively) similar information via any sort of image media. The expressive bandwidth of bodily physical motion (ie, to use a mouse pointer) is always going to be much lower than the expressive bandwidth of text (voice commands and/or keyboard usage). The linguist perspective about actual languages in common usage merely confirms this.

Visual languages find their niche in situations where the complexity of the information to be communicated is low. Selecting single items from small sets of available/potential items is basically the *only* situation where visual representation is preferred. Otherwise, we see visual language used as indicators (traffic lights, stop signs, bathrooms, etc).
However, this format of communication does not scale. As applications become more complex (as we use computers to work with ever larger amounts of data in more and more complex ways) the disparity between the carrying capacity of a visual language and a textual language grows ever more evident. A visual language works well when there is a fairly limited set of contexts (or modes) that need to be identified. Yet when the number of possible contexts (application states) grows without bounds, the meaning of any single visual element quickly becomes washed out (meaningless). Attempting to identify commonalities that span all possible applications becomes an ever more difficult task as the variety of applications continues to increase, and is best avoided from the onset.

Therefore, while there will always be a niche for visual languages, it is evident that textual interactions are here to stay and are ultimately more natural for _ongoing_interactive communication_. In situations where there is little need for back and fourth dialog/communication, visual elements work well (think presentation mediums: billboards, printed magazines, television). For situations that require user interaction and participation (social organizations, CAD software, etc) textual representations will always dominate (either voice or keyboard).

The reality of the above paragraph is evident when you consider that the people who most vehemently promote visual models of interaction (icons, flash, 3D effects, etc) are generally business types interested in promoting (selling) a product. Their interest is to generate only one type of interactive user response/expression: dollars spent. The use of flashy visual elements is to entice new users into spending money, not to generate lots of continuing and ongoing dialog with the user (communication is expensive for a business). Optimizing a UI design to get a new user to start using an application is a far different activity than optimizing a design to facilitate ongoing user interaction. Presentation only oriented things will be visual — interactive oriented things will ultimately be textual at their root, regardless of whatever else gets integrated with it.

Would you rather teach your children a “pigden” language that they can ‘get up and running with’ right away, or a hard language (English) that will last them their whole lifetime? What sort of job opportunities would you really expect to get if your entire working vocabulary consisted of only a few hundred words (a illegal alien laborer) vs. someone who had real mastery of a language (doctorate students typically have vocabularies of at least 50K words). How obvious does it need to be that people who can type are going to be far more effective with a computer (and more employable) than people who can only operate a mouse?

Where complexity is increasing and interactions are ongoing and long (computers are here to stay) it is clear that textual interactions with a computer are going to be ultimately better (voice commands/dialog or keyboard). We should design with the end in mind (re “7 habits”) rather than continuing to even worry about the illusionary short term gains of purely visual computing. The evidence is in and the verdict is clear: prefer textual interactions in GUI design than visual ones. Arguments/discussion of the “grandmother newbe user” largely miss the point that the effort to hide complexity is itself complex and in the long run, make the situation worse for everyone. This sort of mindset does not scale and this becomes more an more of a problem as things DO actually scale larger.

As far as the ideas of this Blog are concerned, I largely support them and have even done some work/experimentation along similar lines with varying levels of success. I encourage further principled dialog along these lines and look forward to seeing these ideas developed further.

Forrest Landry,
San Diego, CA.

Are we expected to upload all our documents to “the cloud” where a computer can compare their topics based on their summaries and tag them for us? Would we not worry about the security of our private data?

Do we need hierarchical ontologies after all?

In your linked Clay Shirky video he talked about Flickr users tagging photographs which could be retrieved by the filter: “cats in sinks”. You implied that this was superior to a hierarchical filing system which would have be categorized (no pun intended) as either cats/sinks, or sinks/cats. Yet, as soon as I read this I wondered about:

“kittens in sinks”

Surely the system should know that a kitten is a kind of cat and whilst placing a priority on Flickr photographs tagged with “kitten” and “sink” it should also include “cat” (which is a conceptual generality of the specific ‘kitten’ class, in other words a “superclass”, ontologically-speaking) as it may otherwise yield zero hits for being too specific.

‘Yes, we particularly want kittens, but we would accept cats, some of which may turn out to be kittens in sinks tagged incorrectly as cats in basins.’

Who gets to decide upon the structure of this ontology? What are its political ramifications? Perhaps, the imprecision of any such interface could be helped by making it conversational… approximating iteratively to a point of common understanding between man and machine. Rather than expecting it to know what we mean on the first attempt.

Having multiple routes to discovery enables users to find the path they feel most comfortable with, and more quickly discover what they can do in the OS/application. As some have mentioned, some users prefer the mouse, or thinking categorically or visually. My favorite thing about items like the OSX spotlight vs. finder/dock/shortcuts or Vista’s co-compatibility with searching or sorting into folders is that users can choose what they feel comfortable with. Even though I have had the spotlight search for years on my mac (and it is beautiful and sometimes I use it), my spatial recognition is stronger and I often access applications or items by remembering where they are on the desktop or dock. Accomodation of spatial/visual thinking is just as important as intelligent “keyword” searching/filtering.

I agree that needless categorization sometimes plagues our electronic efficiency and ease of use. But even an intelligent command line is more of a programmer’s solution. I can teach my mother to use Google in only a few seconds, but how long will it take for her to search by keyword? Multiple avenues of discovery are valuable. Please don’t leave the spatial thinkers out in the cold. :)

Loved the article!

http://quicksilver.en.softonic.com/