On Feb 29, 2012, at 5:45 PM, Charles Dyer wrote:
>> 
>> Yeah I meant just the public domain time frame. Of course the total 
>> development time would be much longer. Even SAMBA on 10.6.x was old, so it's 
>> possible they've been working on their own implementation for a while, maybe 
>> fully 4.5 years since that's about when SAMBA went GPLv3.
> 
> It doesn't show.

Well consider SAMBA 4. It's been in technical preview or alpha (or whatever) 
for 6 years? It was in development for several years before that? I mean...this 
seems like it's one nasty protocol, like aliens that have acid for blood sort 
of thing. It's really complicated. I don't know why it's complicated or why 
it's so popular.

Red Hat bought Gluster last year and they've turned GlusterFS into an open 
source project. It's really, really frigging cool. Mac, Windows, Linux FUSE 
clients (or conventional SMB or NFS if you wish) Replication, distribution, 
striping, self-healing, writable snapshots (like LVM or ZFS), quotas, logical 
volumes, storage aggregation - and all of those features are stackable. With a 
single global name space for the volume. And something ludicrous in the 
brontobyte realm for total storage limit, with a practical limit of ~1000 
storage bricks at 16EB of storage per brick.

If Apple had open sourced AFP ten years ago, what would have happened? It's not 
a cluster file system, so it's nothing like GlusterFS. But it is comparable to 
SMB/NFS. *shrug*

> That's not going to happen. Server 2003 and WinXP are EOL, Redmond is _not_ 
> going to fix any major problems with them. Period. Minor problems, maybe, but 
> nothing which will require serious work. They're too busy with Metrosexual... 
> ah, Win8.

Fair enough.



>> For Apple, this almost doesn't surprise me. Maybe it should, but come on... 
>> I've never taken them seriously when it comes to playing nice with foreign 
>> OS's or making connectivity to anything but the internet, easy or secure.
> 
> They don't care. It's that simple.

Ha, well that's fine. But they got into this sand box voluntarily, and in 
effect claim they want to be taken seriously, in this space. I accept them not 
wanting to be in server space. I accept not wanting to be in enterprise. I 
don't know they can get away with holes of this magnitude and not turn into 
Fisher Price*.

I'm happy to propose 99% of the stuff people think is private/important is 
actually dog crap, not important, does not need to be secure, is already out in 
the world in some other form anyway. But that is not at all how most people 
view their data, and it's not up to Apple (or Microsoft, perhaps in particular 
Microsoft) to not give a crap, in effect second guessing other people's idea of 
what data is to be secured.

> 
> The basic flaw appears to be due to Redmond. Lion just revealed the problem.

Ah ha. Well even that surprises me, that some one else had not stumbled upon 
this. This behavior must pre-date fully encrypted connections. The data is 
clearly in the stream "don't look at this, but here, take it anyway" but the 
client is expected to protect? Bizarre. Totally bizarre from a design 
standpoint. I'm on pins and needles for someone to tell me I'm not 
understanding this correctly.


Chris Murphy


* (And mind you I think Fisher Price makes some cool stuff, but I'm not driving 
to the grocery store on the side walk, let alone in traffic, in a Pretty 
Princess Ride-On).
http://www.fisher-price.com/fp.aspx?st=10&e=outdoorfunproduct&pid=64230
_______________________________________________
MacOSX-admin mailing list
[email protected]
http://www.omnigroup.com/mailman/listinfo/macosx-admin

Reply via email to