I’d love to see a really simple time tracker application, something like TimeTracker, that ran on Tiger’s dashboard. Just a little box for each client or project, click when you begin, click when you end, maybe a comment field, and a way to extract results for a given time period.
I’m happy enough with my local video store. I don’t want to go to the trouble of dealing with Netflix when I’m three blocks from an all right video store.
But I’d like to manage a queue of movies to watch. I love Netflix’s Web interface. I just want a faux Netflix that I can use to keep track of what movies I intend to watch…
(Extra points if you write it as a Dashboard widget for OS X Tiger…)
I’m an RSS addict. If I follow your site, and you’ve got RSS, that’s how I do it.
I like that RSS can tell me when a post is updated. Unfortunately, sometimes it’s not clear to me what was changed: was new information added? Was whitespace added? Just copyediting? It would be great if I had an RSS reader that could show me what changed, with two side-by-side views and color coding. Or even just highlighting the edited parts in red.
For RSS readers that cache lots of content, this should be easy to implement, and would save a lot of brainpower on my part…
Lots of sites have algorithms that predict what books or music you’ll like, based on other things you’ve liked and disliked.
The next generation of those algorithms should focus on predicting which new items are extremely popular, but which you will despise. It would help vociferous readers avoid things like The Da Vinci Code, which would otherwise be recommended due to their near-universal popularity.
In the Mac’s Finder (the equivalent of Explorer, for you windows people), a window looks something like this:
The area on the right displays files and folders in the current folder. The area in the upper left (above the line) lists all of the available disks on the system. The area in the lower left (’Desktop’ and below) is the sidebar.
The sidebar is nice, because you can drag discs, folders, and even individual files there. If you click on a folder, the finder changes into that directory. If you click on a file, it will open in the appropriate application. To remove an item, drag it out of the window and it disappears with a poof — literally: there’s an animation and a sound effect that go with it, same as removing an item from the dock.
It would be really handy if it was possible to have contextual items on the sidebar. If you’ve selected an MP3 or WAV file on the right-hand window, your designated ‘music’ folder should appear in the sidebar; possibly iTunes should appear there as well. There can be a neat fade-in effect, as in iChat when contacts come online.
If you select an AVI file, a ‘television’ or ‘movies’ directory could show in the sidebar, along with a video player. (I use vlc, but presumably QuickTime would be the default.) Drag a file to the folder to move it to that directory; drag it to the player, and it will open.
This should be relatively easy to implement, its interface shouldn’t be too complicated, and it would be one step closer to having the computer think for you.
I’ve finally turned up a product suggested here, a program to change process priorities on Mac OS X. Process Wizard does the trick, and beautifully, too!
It’s unfortunate that Two Ideas didn’t inspire Process Wizard to be written, nor did this site help me find that program. But it’s the first thing written about on this site that I’ve seen in the real world.
Yeah, yeah… there are a lot of computer-related posts here lately. So sue me. I’m really bothered that there doesn’t seem to be any way to change process priorities that’s part of the Mac OS X GUI, though: I frequently have to drop down to the command line to give my video player the highest possible priority, and to drop the priority of my peer-to-peer clients and suchlike. A process renicer GUI shouldn’t be difficult at all to write, and is a darned useful thing to have around.
Update: Process Wizard does this, sits unobtrusively in the menu bar, and is even free. (I found this program via a page of super OS X menubar items while I was looking for something else.)
One of the advantages of using Unix-based operating systems was the opportunity to try any number of different solutions for managing windows in a GUI. (This led to a disadvantage of incredibly inconsistent user experiences between and sometimes even within applications, but what the hell.) One idea I’ve had for a while, which used to be impractical but is no longer so clearly problematic, is to have a real 3D interface for window management.
Apple’s Expose demonstrates that it’s technically possible: though the windows have been reduced to arbitrary percentages of their original size, they continue to update and behave normally.
What I’d like is to be able to move windows forward or back, and for their size to change with their distance. Maybe a control key and the scroll wheel when on a window’s title bar, and the window moves forward or back on the stack. As it does so, it becomes larger or smaller.
Each window would have its own notional pixel dimensions that assume a minimal distance, but the windows could be resized arbitrarily without changing that notional pixel count. One could keep many windows open but shrunken around the edges of the screen, then find a particular window visually, center it, and move it to the front.
I would find this both intuitive and practical: I could select windows visually, watch progress of operations, and still not suffer from a lack of screen real estate. (Even my 20″ cinema display is just not enough. If I had a larger desk, I might consider a larger display already…)
Several nights ago, Emily and I were discussing what a To-Do list manager she wanted would look like. I thought about it some more, and realized that it would be a neat generic interface to a Wiki back-end.
I see a split screen: the bottom part, at least half but possibly more, would be the content of the current wiki page, and the top part would be a 3D representation of the ‘nearby’ wiki pages. It might well look like a chemistry-class model of a molecule. Front and center would be a large bubble that represented the current wiki page. It would probably list the table of contents for that page, if it had one.
Behind the bubble would be other bubbles representing ‘nearby’ wiki pages, with frontlinks and backlinks represented by different colored ’springs.’ Multiple links could be represented by thicker springs. The size of the other ‘atoms’ (pages) could be determined by ‘distance’ (number of links from here to there) and the size of the actual page.
This visual representation might be useful for ‘flythroughs’ describing and diagramming type and density of knowledge in a wiki. It would be useful for rapidly navigating between related concepts to find the most appropriate material. Most important, it would be fun and pretty eye candy.
It would be nice if there was an XML schema for restaurant information.
It should include a name, a street address, a type of cuisine, the hours, information about delivery (available? minimum charge? distance?), and an external URL, at the least. This would be useful for building restaurant-hunt applications, neighborhood web sites that can list many restaurants with a consistent appearance, and so on.
An easy extension would be an additional list of URLs for reviews. These could include newspaper reviews, but could also include pointers to resources such as the King County restaurant inspections Web interface.
A more complex version could include information about menu items: names, descriptions, and prices, at a minimum. This could be used to provide a common interface for Web food-delivery front-ends, either operated by the restaurant or by aggregators who deliver for a number of restaurants.
One could also build a lunchtime-negotiation Web interface, whereby people could share their preferences and places they would refuse to eat, and the back-end could negotiate a list of agreeable choices. Though the XML wouldn’t be necessary, it would help if peoples’ preferences included price ranges and distance from their present location; also, such a service would benefit greatly from not requiring employees to update information about available restaurants.
The OS X address book is a wonderful thing. It integrates with lots of other Mac software via an API, and it will sync with my Nokia 6600 via iSync. I’m even able to use my address book for my Unix terminal-based mail client (Mutt) via a Little Brother’s Database plugin.
Unfortunately, the glorious world of Unix software has no way to get at the address book, though a patch for Thunderbird exists.
A partial solution exists with AddressBook2LDAP, which converts an Address Book so that it can be served by an external LDAP server. In fact that application only points toward the correct solution: someone should write a self-contained miniature LDAP server that directly serves a user’s OS X address book. It could restrict itself to the local system, and it should be configured to export the address books either of a single user or all users on the system.
It would be better still if Apple built this directly into their Address Book application.
Such an application would serve as another bridge between Unix and OS X. The fact that it would make my life easier (at least until I see if Tiger’s Spotlight-enabled mail client is good enough to become my primary client) is secondary. (I have a workable solution with lbdb, but zillions more Thunderbird-using Mac addicts do not.)