[kevs]

It's taking up 2GB what are they?

[Chuckit]

Originally Posted by dictionary.com

in·dex
n. pl. in·dex·es or in·di·ces
Something that serves to guide, point out, or otherwise facilitate reference, especially:
An alphabetized list of names, places, and subjects treated in a printed work, giving the page or pages on which each item is mentioned.
A thumb index.
A table, file, or catalog.
Something that reveals or indicates; a sign: “Her face... was a fair index to her disposition” (Samuel Butler).
A character used in printing to call attention to a particular paragraph or section. Also called fist, hand.
An indicator or pointer, as on a scientific instrument.
Mathematics. A number or symbol, often written as a subscript or superscript to a mathematical expression, that indicates an operation to be performed, an ordering relation, or a use of the associated expression.
A number derived from a formula, used to characterize a set of data.

If you want to offer more information as to specifically what you're talking about, maybe there is a more specific answer.

[Millennium]

My guess is that he's talking about the Spotlight index.

Relational databases like Oracle, MySQL, and SQLite have amazingly powerful search capabilities, but actually searching through entire database tables is quite slow even on today's machines. This is true no matter what machine you are running, because it is not a fault of the hardware: it's inherent to the algorithms and models which databases use. You get immense power, but terrible speed. Faster machines are of course better than slower ones, but they can only do so much. If the only way to search a database were to scan through the whole table every time, it is doubtful that these databases would be nearly as popular as they are, because they would be too slow for many things.

In order to make searches go faster, databases use a technique called indexing. In addition to the actual database tables, they store information about the data they contain in another format which makes searching much easier and faster. This is not as powerful as going through the whole database, and each index you create is only useful for certain specific kinds of searches, but if you design your indices carefully then they may be all that you need. Going through an index is much faster than a full table scan, which is why databases can be used for so many things without driving people nuts.

Spotlight's index files -the large files you are seeing- take this concept and apply it to a hard disk. First, information about the files is put into a SQLite database, so that Spotlight can use SQLite's searching power. Those databases are then indexed (I will call these an internal indices, to tell them apart from the index files) to make the searches fast. The database and its internal index are stored together in one file, which is the Spotlight "index" file. That file lets Spotlight search your hard drive with more speed and flexibility than the Find feature of previous OSX releases could.

[ghporter]

What Millenium said. But Chuckit was right; if you were a little more explicit in your question, Millenium wouldn't have had to guess.

And indexing, while it does take up some (often substantial) drive space, makes searching SO FAST that you'll (probably) think it's worth it.

[Big Mac]

I didn't realize that SQLite was involved in Spotlight. So if the filesystem is actually put in database form, where is meaningful difference between this mechanism and Be's database FS? And why did the BeFS have such a strong reputation of speediness on such ancient hardware? Was it simply because there were so few files on those systems to deal with?

[piracy]

Originally Posted by Big Mac

I didn't realize that SQLite was involved in Spotlight. So if the filesystem is actually put in database form, where is meaningful difference between this mechanism and Be's database FS? And why did the BeFS have such a strong reputation of speediness on such ancient hardware? Was it simply because there were so few files on those systems to deal with?

The filesystem isn't "put in database form", per se. The metadata index is in database form; the rest of the filesystem is still pretty much HFS+, albeit with extended attributes for metadata. But you're right, the lines are blurring between BeFS and modern HFS+ with metadata and indexing. And Dominic Giampaolo was one of the original developers of BeFS, but he's now at Apple in the Mac OS X file system and Spotlight groups...

[kevs]

Ok, I'm learing here. Then clearly, you never want to touch/delete it correct?

[ghporter]

I'd say you could mess with the index file all you want. But the next time you use Spotlight, it will recreate the index-and take so long to do it that you'll hate yourself for messing around in the first place.