Main Area and Open Discussion > General Software Discussion
What is the currently best Desktop Search software?
Curt:
I was reading about the features, when I saw this:
Indexing speed:
1Gb per Hour (this value can be influenced by Computer performance)
--- End quote ---
- and my interest disappeared...
-Curt (December 18, 2008, 04:14 PM)
--- End quote ---
On second thought, I realize that it is the content of documents, it is scanning at 1 GB per hour, so maybe the indexing speed really is okay? What do "you" think? :tellme:
f0dder:
"Indexing" means "scanning documents to build the index used for fast searches" - so once the indexing is done, the actual searches (or "index table lookup", kinda) should be blazingly fast.
Curt:
Thanks, f0dder, but... Of course I must first pardon me for having used the wrong words, indexing and scanning are not the same. Sorry! And then I ask: How fast is this "indexing speed: 1 GB per hour", compared to other documents'_content indexers (if there are any)?
f0dder:
Don't worry, programmers don't always use the clearest language to express themselves.
1GB/hr sounds pretty slow to me, and like a pretty arbitrary figure. Surely it should depend on harddisk and CPU speed, as well as the content being indexed? Unless they're artificially limiting the indexing speed?
Harddrives are fast today, and even a disk that's some years old should be able to read 50MB/s, and really a lot more. If we reduce that to 40MB/s to factor in disk fragmentation, and generally being pessimistic, that would still be 140GB/hr. RAM speed is measured in several GB/s (even for old DDR-1 RAM), but of course there's some CPU processing done and index file being written, but even if we pessimize by 10x, that would still be 14GB/hr :)
Of course these are just figures pulled more or less out of the blue air, but 1GB/hr seems weird to me.
Curt:
hmm... are not this two different things - reading the drive, and reading the content of documents? Doesn't the indexer sort of has to open each and every document, before being able to read it? This thought was why I was thinking that it actually COULD be fast, because I have no clue about what it takes to index the content of documents.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version