[G4] defraging a Mac

[Alex]

On Monday, Apr 5, 2004, at 08:52 Canada/Eastern, sr ferenczy wrote:

> [...] actually, i WAS wrong... no specific number, but since os 9.1, 
> apple has allowed "more than 2 forks"
> http://www.speedtools.com/Users%20Guides/Disk%20Defrag%20Guide.pdf 
> [...]

TN1150 (to which I referred you) explains:

    "Future Support for Named Forks

     Files on an HFS volume have two forks: a data fork and a resource
     fork, either of which may be empty (zero length). [...]

     HFS Plus has an attribute file, another B-tree, that can be used to
     store additional information for a file or directory. [...] The
     contents of the attribute file's records have not been fully defined
     yet, but the goal is to provide an arbitrary number of forks,
     identified by Unicode names, for any file or directory."

The operative terms from Apple (i.e., the horse's mouth) are "future 
support", "have not been fully defined yet", and "goal". In other 
words, it's something that Apple intends to implement (and has put the 
APIs in place for), but hasn't yet -- and may never do. So let's make 
things clear: HFS+ files can have two -- and only two -- forks now. In 
the future they may, or may not, have an arbitrary number of forks, but 
we ain't there yet.

>>> [...] i do know that many os x defragging tools do NOT worry about 
>>> keeping individual forks contiguous with the entire file, rather 
>>> they only keep forks contiguous with themselves.
>>
>> That's a new one for me -- but I'm always keen to learn new stuff. 
>> Could you provide more details?
>
> http://www.speedtools.com/Users%20Guides/Disk%20Defrag%20Guide.pdf

AFAICT, the only relevant passages are

    "[...] Intech believes it is much more important to keep each fork 
within a
     file contiguous (in one piece) than it is to keep an entire file
     with multiple forks contiguous. [...]"

    "[...] while Disk Defrag will attempt to make the forks of a file
     sequential to each other, the main priority is to remove the
     fragments of each fork of a file. [...]"

I'm not quite sure how you read there that "many os x defragging tools" 
do it. My reading is that only one is supposed to do what you describe: 
Disk Defrag.

>> Is HFS+, as implemented under Panther, the first desktop OS to do 
>> this type of clustering? Is its algorithm superior to other FSs which 
>> do adaptive clustering? And on what criteria did you decide HFS+ is 
>> better than, say, ext3 or ReiserFS?
>
> i was speaking desktop - basically windows and macintosh - consumer 
> machines [...]

It appears we do not have the same definition of desktop machine OSs. 
My definition also includes Linux, OS/2, BeOS.

> [...] i know very few (in fact none) consumer types who run linux.

Check out the OS X for Users list -- you'll find quite a few. And it's 
fairly natural -- if you know your way around Linux well, you are 
already half-way to being a Mac OS X expert.

>> But this whole discussion is about a bogus issue, driven chiefly by 
>> the "mine is bigger than yours" syndrome. (Or perhaps not bogus, but 
>> certainly of interest chiefly to geeks who eat OS specs for 
>> breakfast.) It's not the file system that really counts, it's a 
>> combination of features, FS included.
>
> nah, its about trying to get the highest performance out of the 
> machines we have without having to splurge for a brand spanking new 
> machine.

How's that? If you decide that NTFS is better than HFS+, you'll buy a 
new PC, but if you know that HFS+ is better than NTFS, you'll get the 
highest performance out of your Mac?! Gimme a break!

> its also about trying to maintain a very useful knowledge about the 
> working of software i use daily [...]

Arguing FS esoterica and which FS is "better" helps you maintain "a 
very useful knowledge about the working of software i use daily"? 
Wouldn't reading Apple Tech Notes or the Inside Macintosh series or the 
Darwin man pages be of slightly more value in achieving such a laudable 
goal?

Let's get real. You've made the choice to use a Mac. If you use a Mac, 
you've got one FS choice: HFS+. (For obvious reasons, UFS is only 
suitable for specific tasks.) If you use Win, you have basically two 
choices: FAT32 or NTFS. If you use Linux, you have multiple choices: 
ext2, ext3, Reiser, XFS, JFS, etc. (For the sake of brevity, I omitted 
older FSs, like MFS, FAT16, VFAT, etc.) So, for Linux users it makes 
sense to argue about FSs -- they have a choice. For Win users, it makes 
less sense -- they're pretty much argued out. For us, it makes no sense 
whatever.

> [...] simply pointing out what appear to be erroneous/misleading 
> statements.

And those are?...

> [...] i got most of my information from one source, but trust software 
> developers as big as intech to know what they are doing.

You trust Intech more than Apple when it comes to the specifications of 
the file system designed by Apple? It's your privilege, of course.

f