mnottertail
Posts: 60698
Joined: 11/3/2004 Status: offline
|
quote:
Well, if you and Ken want to look into it, you will find that they have a different mechanism for error correction for large drive arrays. Don't let me stop you guys from going and educating yourself. As for indexing - again commercial indexes were doing 43 billion files in an hour. I don't know how that scales. Ken already explained RAID, but yeah, I am aware of the different error mechanisms having been a computer programmer that geeked in places the depth and breath of this country, for over twenty years and creating (not alone) the E911 system (among other things) for verizon, and oh, and voice compression and recognition for automated operators for verizon. And have done most of my work in telephony and computing gizmos. Yanno, like telemarketing stuff. Hooking telephones and computers together to transform that stuff. Yeah, cute shit on indexing, you ain't in the universe, let alone the galaxy. Let us consider the simple word how....have you any idea of the state of voice recognition in the field of free form voice translation? How, midwestern Haow, Texan 'ow, cockney HHHHHHHHHHHHHHow, posh Huuuuuuw, indian (not native) well, I suppose if you had a very very good detector you would probably want to transform it into soundex so that you could sort of normalize it........and then you would be able to boyer-more the search patterns......but you have no context word for word. And if looking for something..........(which is what you would use an index for)........ I will bomb your ass into the ground........I am gonna bomb his fuckin ass, we are bombing the eiffel tower wednesday at 10pm, do you have a bombadier snowmobile? Heres a bombshell, got free tickets and went to the Megadeth concert at the Bomarcy center. It was the bomb, and boy it fuckin' near killed me, was I bombed. But I gotta hit the situation room, I'm turtling, it is touching cloth, so I gotta go bomb china. The cascade of indexes are bigger than the data, even at a simple hash table using y = mx + b, which is going to work out about 1.4 times the data for every index, and then the joins (inner and outer), coalesces, clauses, and whatnot from the query would take hours and weeks to come up with something across the spectrum, so that you are looking for a quark in the universe. That's one reason why cellphones have been captured for so long.........they got text on em, and there is far better error correction for text than is on fuckin cell phones. So, we gotta go at least 4 or 5 times the data to even make any correlation by indices whatsoever. At best, they figure out a perps foneno and then get all records from and to and then correlate those and then the ancillaries and find out what is important between here and there. (And that assumes they have access to the SS7 records from the switches to get the to and from) I will be long dead before they can index my conversations with my family, my friends and verizon wireless bill collectors. That's why I ain't worried, hey, I don't fuckin like it on GP, but I ain't fuckin hysterical.
< Message edited by mnottertail -- 7/20/2013 7:50:36 AM >
_____________________________
Have they not divided the prey; to every man a damsel or two? Judges 5:30
|