Java 6 is now available for Leopard but it's 64-bit only and I've confirmed that Eclipse 3.3 won't run using it. So it's appropriate only for running 64-bit ready Java software (like Tomcat) for now.
If you still want to set it as default VM, first try the Java Preference app in:
/Applications/Utilities/Java
The app is apparently sensitive to command-level profile changes so it didn't work for me. What worked for me is the following:
cd /System/Library/Frameworks/JavaVM.framework/Versionssudo ln -fhsv 1.6 CurrentJDKsudo ln -fhsv 1.6 Current
To undo above changes, simply repeat the command back to where CurrentJDK and Current symlinks were pointing to. For me, following worked:
sudo ln -fhsv 1.5 CurrentJDKsudo ln -fhsv A Current
Are there any programmable game chairs out there? I mean one with force feedback motors instead of chair on woofer kind. Chairs from Ultimate Game Chair looks good but I've read that they are driven by sound output instead of specific force feedback signals because game console makers started to encrypt the signals.
My need is not urgent because the idea (no, it's not a game) I want to explore has been gathering dust in my head for more than a decade but I want to start collecting necessary parts, one of which is zero-g chair with fine-grained force feedback control. Biggest concern I have with my idea is motion sickness and fatigue (imagine playing Grand Turismo for 8 hours a day, 5 days a week). Only workaround I have is selective process like the way fighter pilots are chosen.
Before Web exploded, I used to haunt Stanford libraries, looking for referenced papers and proceedings. Even now, Web offers breadth but rarely the depth of information I want to research without hitting some fee-barriers. Most often, I would hit the wall at ACM when I am find an interesting paper then try to follow the trail of referenced papers. I used to find interesting USENIX proceedings to buy when I used to visit bookstores weekly but I haven't had the need nor urge to visit a bookstore in years.
The good news is that USENIX papers are now free. The bad news is that ACM and IEEE papers are still not.
Being able to write code for any platform in any languages is a good thing but having to do it in parallel is not. While overseeing java-based server-side development, I am tasked with developing clients in Firefox (javascript, XUL), Internet Explorer (C++/Win32), AIR (actionscript/javascript, Flex/HTML), and now iPhone (Objective-C/OS X). That's five language and six platforms I have to juggle. It's times like this that recent abundance of syntax-aware editors with auto-completion shine. And, of course, Google code search provides instance access to all the example codes (good, bad, and ugly) in the world.
It's been coming for a while in plain sight but Dave has now put a name to it: FlickrFan. Is it something new? Seemingly not but, knowing Dave, I know he is seeing something I currently cannot see. And the chance of mis-focusing is pretty high. Software is magic, folks. Like the way one can put meanings between lines of words, one can capture something wonderful between lines of code, even crappy code, code that anyone can write but all too often fails to capture anything but vain eye candy or crippled features.
I still remember hearing about Dave's Living Video Text two decades ago. It seemed mundane at first glance and it took many years before I saw it on store shelves. But his outliner idea dawned on me eventually. Since then, there were scriptable apps, blogging, and podcasting. I don't know why ideas he gets attracted to always end up being late bloomers but, thankfully, his persistence gives us time to come around eventually. When a mountaineer points at something, one should not focus on his finger.
Is FlickrFan a start of something wonderful or will it be much ado about nothing? Frankly, I don't know but I also don't see much return in being a premature critic or a flip-flop.
This is what I am seeing in FlickrFan at the moment:
FlickrFan captures and projects images from a user's social network without all the clutter of web user interface. Photos of friends floating by on the big ass screen you paid thousands of dollars for, every day and every glance, in every gorgeous detail, each throbbing HD pixel filled with bits and hearts of the world you choose to care.
Now, if you had to choose between the FlickrFan show and "Deal or No Deal", which would you choose?
I spent some time Saturday with AIR/Flex3 B3. There were pockets of improvements but overall quality of engineering left me unimpressed. Enough said.
Whoa. Adobe released AIR B3 and Flex 3 B3 last night. I won't have time to look at the goodies until the weekend but this is great news. Adobe also released AMF3 spec (Flash binary wire format) and reference implementation called BlazeDS.
I wrote my own AMF3 library in Java while back and there are at least two public implementation of AMF3 library so BlazeDS is not a big news but the spec will be useful to see if I missed anything in my implementation. But revealing the details of AMF won't make a major impact because 1) it was mostly known already and 2) it's not the wire format but the format of proprietary contents flowing over the wire that is keeping third party Flash servers back.
While I am well past 24, by two decades egad, I'll throw in my two cents to help Facebook fix Beacon. I figured this is more constructive than joining the negative flash flood going on now.
Facebook Groups as Limited Opt-in
The way out is to use groups to encourage as well as letting users control Beacon opt-in. If I am fanatical about Coke, I would click yes when joining "I Love Coke" group required opting into beacon news about my recent Coke-related purchases. If I like Sci-Fi, I would join Sci-Fi groups and it would feel natural to opt-in when asked if I want my friends informed of my Sci-Fi book purchases at Amazon and not for other types of books.
Group Controlled Privacy
Members of a group should have a direct say in which beacons should be required or offered as optional for members. Minimal number of votes will hamper use of polling for this purpose but, even if not enough people voted to enable or disable a beacon, each vote can be seen as opting-in as an individual decision instead of a group decision.
Group Rewards
Corporate sponsored groups can use rewards to members to entice more users to enable beacons. Opting-in to join random drawing for limited collection of Star Wars action figures is, I think, a lot of Star Wars fans will chooses to do.
To set the world on fire, one must know which way the winds are blowing. As someone wise once said that one can learn even from a 3-year old, Facebook could learn a thing or two from pyros.
One thought I've had for a long time regarding the database technology is that, despite SQL, it's mostly designed from database service provider's point of view. Programmers appreciate fine grained control but shoving most of the decisions to programmers also creates unnecessary burdens, particularly for mainstream use cases.
Why should I have to dictate which indexes to create? Databases should be able to figure my needs over time and make necessary modifications as needed when access pattern changes.
Why do I even have to define and alter table columns explicitly? Why can't databases figure things out based on queries I make and optimize/adjust as needed?
Why can't I just insert some pojos encoded as JSON then query them which ever way I want and leave the rest to the database? Except for extreme use cases, there aren't that many varieties of queries and modifications an application requests or makes so it shouldn't take much work to automatically optimize and morph database schema behind the scene, including automatic creation of indexes and incremental strip show of data fields.
And finally, why the hell do I have to tune anything? If a car's engine can tune itself, a database should be able to tune itself. I mean the level of complexity involved is not rocket science after all. Forcing 'users' to dictate everything leaves less room for automation.
While new/reborn technologies like CouchDB is interesting, I think there are quite a bit of improvements in the database middleware technology that remains to be explored.
