The first public source code release of Sakana is available, and it turns out Brian implemented RSS 2.0 support before I had a chance to. Oh well. Not that there's any information in there yet, but you can subscribe to the Rileys wiki changes if you like. “qbullet.sidesmiley”
6:44 PM | Radio
No Comments
Noticed tonight that Dave Bort (Danger employee) has an RSS feed for his very funny “drarings“. I really wish some of my other favorite web comics would do that, especially those which are rather less frequently updated as of late.
By the way, I got back to Champaign late Saturday afternoon. Almost missed my connection in Chicago. Freak ten-minute snowstorm in Champaign reduced visibility to zero, so the plane had to fly around a bit before landing. Had not one or two, but three, welcome packages from Airborne Express on my doorstep (I only registered one account, and didn't even ship anything from it!).
Haven't stopped working since I got back, with the exception of a Warcraft III break tonight. Research is going really well, but I still need to focus more. Played my cello today, for the first time in over a month! Need to do more of that. Worked a bit on Pester on the plane; adding system-wide hot key support.
6:39 AM | Radio
No Comments
Why is Mac OS X USB so broken? I realize I probably have a more complicated USB setup than 99% of people, but inevitably I have to perform various weird procedures to get my USB devices working after I've been away from my desk. Sometimes my machine crashes in the process; I file bugs, sometimes they're fixed. This time, after finally getting both my keyboard and trackball to work properly, I see this message several times a second:
USBF: 542976.918 Microsoft Trackball Explorer\M-.[0x56eba00]: error setting feature. err=0xe000404f USBF: 542977.142 Microsoft Trackball Explorer\M-.[0x56eba00]: error setting feature. err=0xe000404f USBF: 542977.366 Microsoft Trackball Explorer\M-.[0x56eba00]: error setting feature. err=0xe000404f
Every second or so, the trackball resets, so if I'm dragging something at the time, it's as if I released the mouse button. Unplugging and re-plugging the trackball doesn't help. Even better, the message appears whether or not the trackball is actually connected. [Logging out and back in fixed the problem.]
Mac OS 9 USB was not this bad; in fact, it seemed downright elegant when I wrote a USB driver for it. It had some neat features like auto-downloading drivers which OS X doesn't have (yet? ever?). This is progress?
On the other hand, having to spend 5 restarts upgrading the NVIDIA drivers for a dual-display GeForce4 on a Windows 2000 box tonight made me appreciate my Mac. When you upgrade drivers, you're warned to uninstall the old ones first or bad things might result. After the upgrade is complete, you've lost all your display settings—back to displaying on a single monitor, in this case. When you change one setting, a dialog box informs you that the control panel must close because it might flake out otherwise (I wish I were joking). The default dual-display mode is to make Windows think your monitors are both one gigantic screen, which means among other things that the taskbar won't sit at the bottom of the top monitor. There's an option to make them appear as separate monitors, but it's somewhere different from where you set up the dual display features.
NVIDIA seems to think if you cram every imaginable option in there, someone will find use for all of them. They have something akin to an OS X-like drawer in their interface. It pops up when you click a tab, animating its appearance. Except the drawer isn't wide enough to hold the tree view full of option panels, and it's not resizable. You can close and open it manually (why?) with some bizarre looking controls, and there's also an animated pushpin that seems to do something but I'm not quite sure what. Now that's great UI.
5:56 AM | Radio
No Comments
The Sakana home wiki now contains some useful information if you want to set this up. It's the first wiki software I've seen or used that I absolutely had to have now. Other people may feel the same way. “qbullet.sidesmiley”
The aforementioned RSS support probably won't come until next weekend; I need to spend all day tomorrow on research.
4:16 AM | Radio
No Comments
Photos from my birthday are up. Most of the photos we had taken at the restaurant turned out to be unsalvagable.
6:00 AM | Radio
No Comments
Danger's hiptop developer wiki is up now, for real—there's a “wiki” tab on the main developer page to replace the previously-inactive forum. The software behind both it and the Rileys wiki is Sakana, which I discussed yesterday.
I've started hacking on Sakana—RSS support will be my first project, likely, so I can subscribe to the recent changes on developer.danger.com as I already do with CocoaDev. An initial source release of Sakana should be coming soon.
12:29 AM | Radio
No Comments
iAppViews: Apple's iApp Style Tables and Outlines. “Some of Apple's iApps, such as iPhoto, and Safari, have tables and outlines with alternating blue backgrounds. This appearance is not trivial to duplicate.” I spent many hours getting my own NJRTableView to do this properly, and a recent email I received from Nat recounted his problems doing the same for NSOutlineView. (Thanks, Nat, for sending me a follow-up email with a pointer to this site.)
3:11 PM | Radio
No Comments
The Rileys wiki is up and running, though mainly content-free for the moment. Its appearance is pretty much the way I want it, and the functionality will continue to improve. I showed it to my father after dinner tonight and he took to it immediately—not that that guarantees he'll contribute in future, but it's a good sign.
The wiki is about 2500 lines of Python with Medusa as its HTTP server. In order to provide a sane URL, it runs behind Apache using mod_proxy_http. I had some problems initially because of a bug in Apache httpd-2.1-dev, which causes ProxyPass to not work properly if the proxied server supports connection keep-alive. There's no problem with 1.3.x, the version which every other sane person uses. “qbullet.sidesmiley”
(Just discovered developer.danger.com is a Mac—since I saw an error message from port 16080, Apple's traditional WebPerformanceCache port. Of course I could have checked the HTTP headers, but that’s too easy. Neat!)
9:25 PM | Radio
No Comments
rileys.us is up, sort of. At some point I'll install a real content management system, probably something lightweight and wiki-like, but this is better than nothing, and I'll be able to reuse my work on the style sheets. Trying some prototyping in Dreamweaver and Fireworks MX, since I had copies sitting around—definitely these versions are a lot less annoying to use than earlier versions. Integration between the tools still leaves a lot to be desired, and I can't get Fireworks to output transparent PNGs no matter what I do. (Not that it matters since WinIE6 still doesn't support them). Fireworks is a lousy Mac OS X citizen, consuming 10–15% of my CPU time even with no windows open.
5:23 AM | Radio
No Comments
Looks like some good CPU usage improvements are coming to Radio. I've had to set the upstreaming time to a large number to avoid Radio's monopolizing my G4, so this will be great. Wish I had time to try out all the cool Radio tools that are in development, but here I am stuck with a half-broken activeRenderer design for a year or more. Perhaps after I get rileys.us up and running and our photo album fixed… the writing is what's important, anyway.
1:07 AM | Radio
No Comments
