Parents came, and went...the weekend was filled with much laughter and silliness...and I actually got my Mom to eat sushi (well, tempura). Hehe...that was an experience... I'll save the chopstick stories, though. Need more sushi...too bad it's expensive enough to break my wallet in two.
Also went exploring...and I discovered how to get from Atascadero to Arroyo Grande using only one CA highway (as opposed to 3). Of course, it takes 2 hours and involves a dirt road. :-D
I think the little black cat is "ours" now...she technically lives across the street, but spends most of her time over here. (Still don't know what her name is, though.) She's been here every day this week when I leave in the morning, and when I come home in the evening. I always have to stop and say hi, and pet her, and scratch her under the chin before I'm allowed to leave or go inside. Sooooo cute... I want one that I can keep! :(
I've been having more programming ideas lately...mainly because I'm pissed at the inadequacy of a hierarchical filing system that doesn't even incorporate all my information (contacts? email? blogs/news?). Had a great idea in the shower tonight...which I'm fairly sure I'll remember. Now if I could just code it up in any sane length of time... I think I could make it user-friendly enough for normal people to use, too.
Haystack is a good start in the direction I want...but it only works on one machine, and it's incredibly fscking slow, even on my P4 2.0 GHz laptop with 512MB RAM. I suspect this is due to the combination of the Java VM and the storage mechanism (RDF)...but who knows. I know I can build a much more efficient storage system than XML...using XML for data storage is just silly. (Yes, you could keep everything in RAM, but that doesn't scale, and it kills your startup time.) It's much nicer when it's being used to communicate data, not to store it.
Believe it or not, the most efficient way of storing data that I know of is with a GDBM (or similar) file. Key/value pairs in a binary format where keys are very easily and quickly retrievable...it's being used in my Bayesian filter right now, because it's the only thing fast enough to do the job, and it doesn't require loading a dozen 5 MB corpuses into Python dictionaries in RAM.
GDBM would be great for storing indexes (because of the FAST search times), and with a filesystem like ReiserFS or XFS, you can create lots of small flat files that contain individual objects. (By lots, I mean thousands or more, all in the same directory.) Yummy...database. I suppose we'll have to see how well it performs...I know that opening and reading individual files can be a bottleneck (particularly the open() call), but that should be reduced by (a) doing some sort of in-memory caching, and (b) using a filesystem which organizes directories in trees (a la XFS/Reiser).
Oh, and yes, I've thought about using a real RDBMS (a la PostgreSQL) for this project as well. I just haven't decided, based on usage patterns, if an RDBMS would be better, or if my way would be better. I'm leaning towards my way, because I *think* it would be faster to access individual objects. (You wouldn't need to do a full query to find/retrieve them...just open the file you want.) But, I probably ought to save all this for when I get around to actually writing up something explaining what this grandiose program is supposed to do. :)
I have ants crawling up my legs. *shiver*
Discovered a new artist on eMusic today...this one isn't *quite* as good as Enigma, though it's right up there. It could be because I have certain...erm, associations with specific Enigma songs. Enigma tweaks my memory in rather pleasant ways.
So, eMusic has already paid for itself...the two CDs worth of music I downloaded tonight are, by themselves, more than enough to pay for this month's subscription.
Let me back up a second, though, and explain this partial change of heart. I don't have any problem with artists making money for producing their works. In fact, I think it's great. I also don't have too much of a problem paying the distributors a certain amount of money for doing the distributing (after all, pressing those CDs and maintaining eMusic servers does take a certain amount of money).
What I do have a problem with, is the way the distributors (RIAA) are gouging customers. Anything over about $10-15 for a CD is absurd. I also have a problem with DRM, and with copy protection. At least for now, eMusic is providing songs that are completely free of copy protection, and they are doing so at a very reasonable price. (Not to mention, they have some fairly decent music out there.) I think it is important to show the RIAA that if they play nice, then we'll play nice (or at least, I'll play nice).
Should eMusic raise their prices extraordinarily high, or stop delivering music in "free" formats, you can bet my subscription will be cancelled in the next month, and I'll make damn sure to tell them why.
That only leaves the artists...and the best way for them to take care of themselves is, if they don't like the way the RIAA is handling things, they should stop doing business with them, and go out on their own. If the RIAA is treating artists sufficiently poorly, then the artists should take their business elsewhere. If enough artists do, and they have no music, then customers will take their business elsewere also.
Unfortunately, the music industry isn't exactly a free-market economy... so I guess you do the best you can.
22:37:03 up 4 days, 15:48, 0 users, load average: 0.00, 0.00, 0.00