A few years ago my sister was looking for very simple, object-oriented, effective family history software. Not a superexpensive geneology package that would help you track primary-source documents in overseas libraries, but a way to jot down who was married to whom, link in photos and audio recordings and stories, and maybe even keep track of conflicting or questionable information. I'd love to do that someday but I know I don't have time, and it would be a widely useful project.
2006-11-06 11:18 pm (UTC)
This software may require a little knowledge to setup, but has everything you described, and is open source (hence free). It uses a web interface, and can be set either for individual usage, or as an internet server. With minimal understanding of html, it is possible to link whatever you want (after you set up apache web server). Basically, it can be slightly challenging to get started, but once setup, trivial to use and maintain, fully featured, and free. http://cristal.inria.fr/~ddr/GeneWeb/en/index.html
2006-11-06 11:55 pm (UTC)
Re: genealogy software
Cool -- thanks!!
So, here's one. It's more of a difficult design problem than a difficult programming problem. There are a lot of situations where people get together: lunch, game nights, parties, quartetting, etc. What would be nice is to have a place where you can say, "I can game on Tuesday and Thursday evening", and other people can do the same, and an Event pops up.
Now, Events need to have their own dependancies before firing. For instance, every Game Night needs a Host. A quartet needs tenor, lead, baritone, bass. Also, you need to put in how often (maximum) that you want that game night to fire. That way you're not doing game night every night. It also means that you can say, "I'll host maximum every 6 weeks".
And inviduals need to be able to prioritize their Events. I want the Northern Game Night before the Southern Game night.
This also works for restaurants. I can say, "Chili's every two weeks max". Every day at 10:00, an email is sent out with the "best" choice of restaurant(s)/people.
Hmmmm... I wonder if you should be able to "I will go to any restaurant unless xxx goes along". Naaaa.
Oh, and you'll want to do this in django
I love this idea too. I envision a spreading database where people get to meet friends of friends and get more organized about forming wonderful groups for various purposes and everyone's happier and more social.
Then again, that's rather idealistic. Surely the higher the number of people, the higher the potential for drama (your "I will go to any restaurant unless xxx goes along" comment made me laugh...). But such is life.
Yeah, such a thing woudl be pretty useful and interesting. I guess when I picture "the future", I already sorta had in mind that we'll have computer calendars that can do stuff like this and make socializing much easier.
THE FUTURE IS NOW!!
well, if greg makes the app anyway :P
This is indeed a neat idea. If I can figure out how to do it and such, I'll make it my next project. Also, I know you've prodded me about django
in the past, and it does look like a neat system, and I have been wanting a more database-y project. And Python is good, although I'm more in a Ruby state of mind ATM...
OMG do the Livejournal thing. Better yet, figure out a way to make all your past entries friends only with one mouse click.
Can you have it export to a calendar with entry titles on the days you made entries? And then you click on the title (or beginning of entry) to read it?
That would be excellent. Sounds kinda involved, but VERY useful.
OK, I'm gonna work on the LJ backup thing.
For making everything friends only, you might want to check out the various clients
available. If you can't find anything that works I can try to add that functionality...
So for the calendar thing, do you mean export to a "real" calendar? (Outlook, Google Calendar or something) Or do you mean just a page like my archive page
Either one. Google would be cool. Hey that event idea up there is ingenious.
2006-11-12 10:31 am (UTC)
I will suggest you to not give up on netflix yet. Even, I have been disappointed at my inability to get into the leaderboard. I am always a week too late for the scores on the leaderboard. I am now able to get an RMSE of 0.958-0.962, that would have put me on the leaderboard a week back.
But, I am still trying. I would suggest you to look at alternatives to your current strategy. A plain correlation based analysis(using clustering to reduce finding correlations only for related movies), didn't take me far either, and so I am now trying to use SVD to do dimensionality reduction, and then trying various alternatives.
Suprisingly, some people have claimed that correlation based analysis took them to an RMSE of 0.94ish. I wonder what special trick did they use.