Right now files are stored in a giant hierarchy in pretty much every OS. My question is, with search technology getting better and faster, will the idea of hierarchy become obsolete? i'm thinking instead something being in a folder, it would instead have some sort of tag that identifies what its for. A good example of this is Apple's SpotLight, leading many users to not even use folders, instead putting everything into one big folder and using search to get the file they are looking for. any thoughts?
Hierarchy vs Search
- Thread starter RAMeater
- Start date
tlarkin
VIP Member
Why not have both? Full paths are important, and mainly they are more secure. I will use Unix as an example.
Let's say I hijack your .bash_profile on your Unix rig and I make an alias of /bin/ls to a new binary called ls. Then I change the binary so it does what I want in addition to having it list.
So, every time you run it, it would list whatever it is you told it to on your terminal session and also execute a script I plotted on your machine.
Now, if I am writing a script, I want to use the full file path in case something like this has happened. So, if I invoke ls to say, list files for a loop in a script, I will use the full path of /bin/ls and that way the script knows exactly what to use, where if I just use ls by itself, it does the command and executes something else.
Editing someone's bash profile and $PATH is rather easy if you got physical access to a machine.
So, using full paths is important and more secure, but that doesn't mean you can't use meta data to search for things. From a programming perspective full paths and heck even a networking perspective it is more efficient and more secure.
If search features went over the WAN were allowed to do that it would generate a lot of over head and lag on a network. If you specify full paths it will search with in that path.
You can have both, but you will always need a hierarchy and to use full file paths.
Let's say I hijack your .bash_profile on your Unix rig and I make an alias of /bin/ls to a new binary called ls. Then I change the binary so it does what I want in addition to having it list.
So, every time you run it, it would list whatever it is you told it to on your terminal session and also execute a script I plotted on your machine.
Now, if I am writing a script, I want to use the full file path in case something like this has happened. So, if I invoke ls to say, list files for a loop in a script, I will use the full path of /bin/ls and that way the script knows exactly what to use, where if I just use ls by itself, it does the command and executes something else.
Editing someone's bash profile and $PATH is rather easy if you got physical access to a machine.
So, using full paths is important and more secure, but that doesn't mean you can't use meta data to search for things. From a programming perspective full paths and heck even a networking perspective it is more efficient and more secure.
If search features went over the WAN were allowed to do that it would generate a lot of over head and lag on a network. If you specify full paths it will search with in that path.
You can have both, but you will always need a hierarchy and to use full file paths.
Folders FTW.
search technology will never be perfect and theres no reason anyone can't have both.
Folders are a fundamental part of every modern file system anyway programs need them, image if office needed to search your entire computer every time you wanted to use a different font, or if crysis need to search your entire computer every time it needed a texture, so many background searches would bring a supercomputers to a crawl.
search technology will never be perfect and theres no reason anyone can't have both.
Folders are a fundamental part of every modern file system anyway programs need them, image if office needed to search your entire computer every time you wanted to use a different font, or if crysis need to search your entire computer every time it needed a texture, so many background searches would bring a supercomputers to a crawl.
chibicitiberiu
New Member
I only use search when I don't know where a file is, usually I know the path of the files I work with. And searching would be extremely slow since hard disks are the slowest part of your computer, so until we find a way to make hard disks without any moving parts and really fast search will slow down your work, not help you.
Only an index put in RAM would work in realtime, but would be inefficient and waste of RAM memory (since it needs a lot; NTFS contains such an index, and it's pretty large).
Only an index put in RAM would work in realtime, but would be inefficient and waste of RAM memory (since it needs a lot; NTFS contains such an index, and it's pretty large).
tlarkin
VIP Member
I had windows search fail on me yesterday. I was trying to ssh into my web server and I had forgot where I put a program called putty.exe. I searched for all file types with putty in the name, returned zero results.
I found the executable in my downloads folder of my home directory. So, Windows search sucks.
I found the executable in my downloads folder of my home directory. So, Windows search sucks.
Concordedly
New Member
Hiyo! Finally someone agrees. Anyway, my thoughts on this are that we as humans also have our own sort of 'filing system' in our minds. We associate things with others, leading to a very defined and profound heirarchy system in just our thoughts. Sure, we have the ability to instantaneously search for what we want in our mind, but sometimes that's not always the best method.
Filing things will probably always be around as it organizes the data we will later need to recall, and with the overall speed issue that cannot be avoided, I doubt that file storing will result in an all-search method in my lifetime, though technology will always surprise.