do macs really handle files in the gigabyte range a lot better than pcs?

[drag]

Sure makes sense to me. When I copy a file from ext2 to reiserfs 3 for instance we should make sure the save all the extra EA's and stuff like that in a special .#file.ext2.ea file. Then when I copy that to a XFS partition over NFS I should have a special .%filename.xfs.stf file, oh and then since I am using nautilas for that last transaction it should save it's own .Nautinfo file then when I copy that file to a Mac share then back to a ext3 filing system I should add the normal ._filename stuff, so I end up with a grand total of:
filename
._filename
.#filename.ext2.ea
.%filename.xfs.stf
.Nautinfo
.DS_Store
._.Nautinfo
._#filename.ext.ea
._%filename.xfs.stf
.%#filename.ext2.ea.xfs.stf

Seems like it makes sense to me. After all we don't want to give up any features or anything..

My only question is what happens after you copy the file to a Fat16 partition? 😛

 

[rjain]

just because you don't use them doesn't mean they're rare

 

[Nothinman]

Just because you use them doesn't mean they're popular.

 

[rjain]

ok, let's remove every feature that isn't used by a majority of users. e.g., that feature called unix. that'll solve all these problems.

 

[Nothinman]

All I mean is that since Apple implemented a HFS+ only feature in their OS they should patch the userland tools they ship to properly deal with those features and one of those things IMO is patching tar to safe the EAs that contain the icon and association information. The fact that I can backup things on one box and restore them on another and have them not work properly is a bad thing.

 

[rjain]

Fair enough, but people who tend to use tar will probably not care about the apple-specific features of their FS. It's probably just not a high priority.

 

[drag]

features suck.. mostly. I know it usually seems worth it, but it usually isn't.

For instance, what if GNU had redesigned all of it's tools to take advantage of ext2 specific features? Sure it would look cool and give the OS extra functionality that people would probably find very usefull eventually, but what about now?

Nobody realy uses ext2 for anything realy important anymore and all that work that went into modifing the thousands of tools gets wasted. And what now? People have gotten used to writing programs and designing scemes to take advantage of ext2 features and now that's obsolete. People could stick with ext3 or maybe require that all new file systems incorporate backward compatability with ext2's special features. This makes it much harder to create new ways of doing things and stiffles innovation, each generation of file systems would have to incorporate the features of the previous generation of filing systems.

Don't forget the wonderfull results of Microsoft's desicion to limit file names to a 8.3 sceme, they still have issues with it ten years later.

It's very hard to create new software technology that is realy worth it. If it's not going to be around 5 years from now, you are potentially wasting a whole lot of work over temporary gains. I can pretty much garrentee you that the 5 years from now Apple would of fixed it's little issues with it's apple/unix personality conflicts and hfs+ will be a thing of the past. That's is of course if apple has act together. Then all the effort to modifing userland tools would be wasted and the development of new features that may actually be usefull would be delayed.

Some features are good, like for instance journalling, but some features nobody exept in relatively few cases are going to give a crap about, like ext2's EA.

Fair enough, but people who tend to use tar will probably not care about the apple-specific features of their FS. It's probably just not a high priority.

It just sucks when you use those tools and ignoring apple-specific features will break finder, and cause other strange side effects that you can't anticipate exept thru trial and error.

but it's not realy that big of a deal, there are work arounds and extra steps that it makes you go thru, but it's just a pain. But I figure the issues will eventually be resolved.

 

[chsh1ca]

Originally posted by: rjain
ok, let's remove every feature that isn't used by a majority of users. e.g., that feature called unix. that'll solve all these problems.

Your comparison is utterly flawed.
One could argue that unix is more used than any other OS when it comes to web browsing. Here's how that tracks:
Apache is the world's most popular webserver, according to the servers surveyed at Netcraft. Of those servers, a majority run on unix operating systems. Now, assume that the vast majority of browsers run on Windows. They load pages off of those webservers, so your windows userbase for any given unix webserver gives you a value of n. I'm surfing along to this fictional webserver using a non-genetic unix (Linux Slackware 9.0), therefore at the very minimum, this fictional server has a value of n+1 users using 'unix' (vs Unix), while there are only n users using Windows.
Let's not get into the large unix systems that manage ATMs, are used in renderfarms to create complicated 3d animations for movies, etc., etc..
'unix' isn't a feature, and it most definitely is used everywhere, directly or indirectly by billions of people each day.

 

[drag]

He wasn't talking about unix in general, he ment the underlining operating system of OS X, Darwin.

And as far as that goes, every Mac user using OS X uses Darwin everyday, weither they realise it or not.