WhatSize | F.A.Q.

WhatSize is a FileSystem utility that lets you view and clean up files, folders based on size. The application is optimized to quickly measure the size in bytes of a given folder or entire FileSystem. Once a measurement has taken place the user can view, filter, sort, find, remove files and folders that are taking too much space.

The application will bring up the System Administrator panel and upon authentication as Administrator will spawn a background process that runs as sudo to measure all files on a given Volume/Folder regardless of permissions. This is useful when measuring an entire Volume. The background process will quit when the measurements are completed or the main application quits.

We measure the Physical Size and Logical Size of the files/folders and display the Physical Size. This is actually a proper way to show how much disk space a file is taking. This is the same value you see when you perform the Get Info on a File/Folder in Finder.

We take into account the Folder Size itself, thus our measurements are a bit bigger than the Finder. The Folder size is what the File System is reserving for the Folder itself. So a Folder with 25000 zero size Files will still take roughly 1MB. The 1MB is used by the Operating System behind the scenes to manage the Folder contents.

This is the size of the file. As an example, a straight text file with one character will be 1B, one byte in Logical Size.

This is the size of that the file takes in the file system. Assuming the file system Block Size is 4KB a straight text file with one character will be 4KB, 4096 bytes in Physical Size.

This is the smallest chunk of information the operating system will use to store data. The default is usually 4KB, but this value can vary between 4KB to 256KB. As an example, if you create a plain text file and have one simple character in it, the operating system will still use one block to store this file. So this file’s Physical Size will be 4KB and the Logical Size will be 1B, one byte.

One Bit is just one tiny piece of information, pretty much useless.
One Byte is 8 Bits.
One KB (Kilo Byte) is 1024 Bytes.
One MB (Mega Byte) is 1024*1024 = 1048576 Bytes. (think 600MB for a CD Audio)
One GB (Giga Byte) is 1024*1024*1024 = 1073741824 Bytes. (think 4.5GB for one DVD, or 25GB for a Blue Ray disk)
One TB (Tera Byte) is 1024*1024*1024*1024 = 1099511627776 Bytes. (1TB is the biggest single hard drive currently)
One PB (Peta Byte) is 1024*1024*1024*1024*1024 = 1125899906842624 Bytes. (think huge, a milion DVDs or Large Hadron Collider)

To make file sizes stand out we use the following color scheme

			file sizes greater than GigaBytes
			file sizes greater than MegaBytes
			file sizes greater than Bytes

With Delocalizer you can remove those unwanted locales for languages you never knew existed. Frankly do you really care about the Elbonian translation of your software ? You can pick and choose which locale to blast out.

Nib files are used to model the User Interface in Mac Applications. Nib’o suction is a feature we added in version 4.6 to allow users to remove all those unnecessary nib files.
The nib files are usually folders with many files inside. Some of the files are not required for the Applications to work normally. Developers simply forget to remove them before shipping the product.

When Apple moved to Intel and Universal binaries this resulted in mac binaries carrying information about each architecture. Same goes for the 64bit architecture. So most of the executable files currently in your mac contain machine symbols for PowerPC, Intel, Intel 64bit or PowerPC 64bit.
When the Operating System runs these binaries, it will choose the proper architecture. For example: On an Intel machine the Operating System will try to load the Intel symbols of the program and ignore the PowerPC portion.
Lipo suction is a new feature we added in version 4.6 to allow users to remove all those unnecessary architecture bits.
If you are running a PowerPC machine, you can safely get rid of all Intel or Intel 64bit symbols, the system will never use them in a PowerPC machine.
If you are running an Intel machine you can get rid of all PowerPC symbols and be fine.
Attention to the details was key in implementing this feature. For example we don’t allow the removal of symbols of the host machine, IE: Removing Intel symbols on an Intel machine. Or removing symbols from a binary that has only one architecture inside, usually older PPC only apps.

We consider two files to be duplicates if they have the same md5 hash value. Figuring out the md5 hashes can be a long and expensive operation, so we will build the hash values in the background and store them into a database for convenience. If the files are modified the md5 hashes will be recalculated. You’d be suprised at how many duplicate files are out there.
We will skip the following file names: .svn, build, cvs, *.app, *.framework, *.bundle, */library/caches/
We plan to allow for more control in a future release using a preference panel.

So you license a copy of the app and go and measure your brand new empty Volume. The numbers don’t seem to match. You just bought and installed a 150GB Western Digital Raptor, Once formated the Volume will show only as 139GB or less available. It’s because drive manufacturers use the decimal system to measure drive capacity (in which 1 GB = 1,000,000,000 bytes, not 1024^3 bytes). Apple switched to using this method in their Snow Leopard release (so in Snow Leopard a 150 GB drive shows up at 150 GB.