The cult of Apple secrecy
Spencer Uresk
being a little lighter in news and Java content. One thing that
intrigued me a bit in this episode was Joe's discussion of the
measures Apple takes to maintain secrecy around new products, and
their reasons for doing so.
Certainly, I understand the benefits of maintaining secrecy around
consumer products - I think Apple does a really good job at launching
products and that is one of their competitive advantages. However,
I've often wondered why they are often so secretive on the developer
side of things as well, and was wondering if Joe or anyone else had
any insight as to why that is the case.
Two examples:
1) The (in)famous Java 6 on OS X issues, which the Posse talked about
in this episode as well.
From my perspective, this wasn't so much of an issue of Apple being
slow to release Java 6 for OS X (other implementers are often behind
Sun as well) as it was an issue of them not communicating what they
were doing at all. There was no acknowledgment that they were working
on Java 6 support, when they were planning on releasing it, or if they
even wanted to support Java on OS X at all anymore. This situation was
somewhat exacerbated by some anti-Java statements made by Steve Jobs
at around the same time.
One unfortunate effect of this - At work, we'd been trying to convince
management to let us buy MacBook Pros instead of the usual Dell
laptops to use as our primary development machines. They didn't really
want to allow us to purchase different hardware, and the lack of
information from Apple surrounding Java's future on OS X gave them
quite a bit of ammunition.
2) iPhone Developer Program NDAs
Back when I was first taking a look at doing iPhone development, I had
joined the developer program and was looking for some books on iPhone
development. A handful of authors had books ready (or at least in beta
form) but were waiting for Apple to lift the NDA before publishing
them. Several of the publishers I talked to were pretty frustrated by
the lack of communication from Apple - they weren't being told when
they planned on lifting the NDA. At this point, the reason for even
having NDAs in place was unclear - tens of thousands of developers
were already working with the APIs and that isn't where the real
competitive advantage for the iPhone was anyway.
This delayed potential developers from getting their hands on iPhone
development books and frustrated authors and publishers who wanted to
get their books out.
So, while I fully understand the need for secrecy when it comes to the
launch of consumer devices, I'm left scratching my head at cases like
these 2 examples where secrecy seems to only confound developers and
offers no real benefit to Apple. What does Apple gain by not telling
anyone what their plans are for Java on OS X? As a consumer, surprises
are often pretty cool, but as a developer or someone looking to invest
a lot of time and money into something, surprises are something I'm
looking to avoid.
Is there some value in keeping stuff like this secret that I'm
missing, or is this simply a case of a culture of secrecy that has
maybe gone a little too far?
Casper Bang
here:
http://gizmodo.com/5427058/apple-gestapo-how-apple-hunts-down-leaks
For what it's worth, I don't think the Gestapo methods scale with
their new found success. You order 45 mio. OLED screens and it's bound
to leak.
/Casper
Fabrizio Giudici
> Is there some value in keeping stuff like this secret that I'm
> missing, or is this simply a case of a culture of secrecy that has
> maybe gone a little too far?
>
>
distinguish between two cases:
1. When we're talking about a new product, as the rumored
iTablet-or-what-else, I find that the secrecy is legitimate. They're
probably introducing new ideas and technologies and are just protecting
their work from being copied by competitors; additionally, marketing
strategies are carefully picked, they are expensive, and can be
jeopardized by a leak (*). I'd sum up that when they're going to release
a new product, secrecy is legitimate.
2. Things are completely different when we're talking about supporting
others' technologies, such as Java 6. There are no news here, as
everybody knows that would be inside the product, and there's just very
scarce innovation in Apple's Java editions (I mean, for instance there's
no specific integration with new technologies such as Grand Central, and
even the look and feel integration is superior in third parties'
products, such as Quaqua). For a while I tried to find a bit of logic
reasoning behind this, and I thought that it was bound to the idea of
keeping their hands free; in other words, they could be working in this
moment to Java 7, but if they later decide that they don't want to
support it they don't have to give to people any explanation. I had to
change my mind after the drop of ZFS from Snow Leopard: they introduced
a read only version of ZFS in Leopard and even put a full open source
customized version on their forge; while it lacks many integrations with
Mac OS X (e.g. the Finder, trashes etc...) at a certain point a release
started to work pretty well. So it was clear that they were working on
it - contrast this with the complete secrecy of their Java 6 support.
Still, when Snow Leopard came out, ZFS was dropped and the project on
the forge removed all of a sudden. Of course, no public communication
about the rationale. This means that they think they can change their
mind at any time even when facts seemed to confirm their commitment. So
I'd only say that their attitude is plain arrogance. They can afford it
since most of the business comes from fashion industry (iPhone) where
sells are guaranteed by masses of fanboys. Every corporate get arrogant
if it can afford to be.
(*) I suspect that some leaks are driven by Apple itself as they have a
great teasing effect on fanboys.
--
Fabrizio Giudici - Java Architect, Project Manager
Tidalwave s.a.s. - "We make Java work. Everywhere."
weblogs.java.net/blog/fabriziogiudici - www.tidalwave.it/people
Fabrizio...@tidalwave.it
Ryan Waterer
licensing issues that they were not able to resolve. I'll find links
later if others do not beat me to it.
Sent remotely
On Jan 6, 2010, at 7:54 AM, Fabrizio Giudici <fabrizio...@tidalwave.it
> wrote:
> --
> You received this message because you are subscribed to the Google
> Groups "The Java Posse" group.
> To post to this group, send email to java...@googlegroups.com.
> To unsubscribe from this group, send email to javaposse+...@googlegroups.com
> .
> For more options, visit this group at http://groups.google.com/group/javaposse?hl=en
> .
>
>
Casper Bang
as Oracle would (rightfully IMO) deprecate it in favor of BTRFS -
which is GPL and of course much more friendly than CDDL.
/Casper
On Jan 6, 4:07 pm, Ryan Waterer <aguitadel...@gmail.com> wrote:
> Correct me if I'm wrong; the reason they pulled zfs was due to patent/
> licensing issues that they were not able to resolve. I'll find links
> later if others do not beat me to it.
>
> Sent remotely
>
> On Jan 6, 2010, at 7:54 AM, Fabrizio Giudici <fabrizio.giud...@tidalwave.it
> > Fabrizio.Giud...@tidalwave.it
Fabrizio Giudici
> Yeah ZFS is CDDL. I suppose Apple could also have dropped ZFS seeing
> as Oracle would (rightfully IMO) deprecate it in favor of BTRFS -
> which is GPL and of course much more friendly than CDDL
>
illogical. BTRFS is still unstable and unfit for production (there's a
flashing warning at the top of the BTRFS official wiki), while ZFS has
been for years. When Oracle completes the Sun buy, they can relicense
ZFS as they wish, so for the short and middle term ZFS stays here for
sure. For the future, I don't see the point in spending money for
cloning an asset that has been bought, so personally I'd not bet on
BTRFS future (with the exception of the open source community - but if
ZFS got GPLed, I see many Linux users adopting it).
Back to the Apple point, ZFS has been CDDL since from the beginning -
nothing changed recently (there has been a legal action against Sun from
another corporate for an alleged patent breach, but as for as I know Sun
got out of it). Half-said things over mailing lists and blogs referred
to Apple not willing to pay the money for a double licensing for
liability issues (a thing that would be exactly the same with GPL, so
CDDL it's not an issue with Apple). I think that everybody understands
that a such a deal doesn't fail all of a sudden from a day to another,
yet Apple's customers only saw the surprise when Snow Leopard was
released (of course ADC signers got it with a few weeks of advance, when
they got the first beta with missing ZFS). As far as we know, Apple is
developing Yet Another Advanced File System, but who know when it will
see the light. Of course, no word about it.
PS As far as I understand, Apple can't know anything about what Oracle
will do with ZFS as it would be illegal for a third party to know about
plans related to a buy in progress - I mean, even Oracle and Sun
personnel can't talk about it.
--
Fabrizio Giudici - Java Architect, Project Manager
Tidalwave s.a.s. - "We make Java work. Everywhere."
weblogs.java.net/blog/fabriziogiudici - www.tidalwave.it/people
Casper Bang
> illogical. BTRFS is still unstable and unfit for production (there's a
> flashing warning at the top of the BTRFS official wiki), while ZFS has
> been for years. When Oracle completes the Sun buy, they can relicense
> ZFS as they wish, so for the short and middle term ZFS stays here for
> sure. For the future, I don't see the point in spending money for
> cloning an asset that has been bought, so personally I'd not bet on
> BTRFS future (with the exception of the open source community - but if
> ZFS got GPLed, I see many Linux users adopting it).
I disagree with that analysis. BTRFS improves upon ZFS in many ways
[http://lwn.net/Articles/342892/] and is widely believed to be the
dominating, if not yet default, Linux file system within a few years.
ZFS is more stable since it has a few more years of development under
it's belt, but that really only goes for Solaris and FreeBSD doesn't
it?
> As far as we know, Apple is
> developing Yet Another Advanced File System, but who know when it will
> see the light. Of course, no word about it.
Possibly. I've always been longing for a file system that allows true
meta-data and labeling, which breaks the age old hierarchical
structure - why MUST I choose one location for my item? (Answer:
Because that's how file-dialogs have always looked) And as much as I
despise the practices of Apple, they are the one company (Microsoft
had to give up, WinFS anyone?) who could and would do this.
/Casper
Karsten Silz
wrote:
[...]
> (*) I suspect that some leaks are driven by Apple itself as they have a
> great teasing effect on fanboys.
Correct, apparently:
http://www.appleinsider.com/articles/10/01/06/former_apple_marketing_manager_describes_companys_controlled_leaks.html
> --
> Fabrizio Giudici - Java Architect, Project Manager
> Tidalwave s.a.s. - "We make Java work. Everywhere."
> weblogs.java.net/blog/fabriziogiudici -www.tidalwave.it/people
Fabrizio Giudici
>
> I disagree with that analysis. BTRFS improves upon ZFS in many ways
> [http://lwn.net/Articles/342892/] and is widely believed to be the
> dominating, if not yet default, Linux file system within a few years.
> ZFS is more stable since it has a few more years of development under
> it's belt, but that really only goes for Solaris and FreeBSD doesn't
> it?
>
relevant they are for end users. From a business perspective, the famous
sentence #2 to Sun customers in the document "Oracle Plans to" is "spend
more money developing Solaris than Sun does now". Given that Solaris
stays here, ZFS stays here at least for the two years that BTRFS needs
to be production ready. At that point, I think that you need some huge
differences in features to have the customers migrate to a different
file system. For Linux this is different, and indeed since there aren't
two distros that agree on the same bits there will be probably both. I'm
still skeptical on Oracle spending big money on R&D for both.
I'd like to make it clear that I'm not trying any prediction. Most of
predictions are wrong even when made by people that are supposed to be
able to make them. It's' just my 2 cents. Of course if the Oracle
documents about the buy are just a pile of horseshit, any reasoning
doesn't make sense.
Leopard users can still use the open source version of ZFS (I do) and
some people have started working on the sources, so I think it will be
usable with Snow Leopard sooner or later. Given that the refusal of
integrating ZFS into Linux made a lot of people to scream, I believe
that *if* Oracle make ZFS GPLed it will take very short to have it
available in a distro.
> Possibly. I've always been longing for a file system that allows true
> meta-data and labeling, which breaks the age old hierarchical
> structure - why MUST I choose one location for my item? (Answer:
> Because that's how file-dialogs have always looked) And as much as I
> despise the practices of Apple, they are the one company (Microsoft
> had to give up, WinFS anyone?) who could and would do this.
>
We'll see. At the moment they are at the bottom of the list, since HFS+
is the worse in the market, even worse than NTFS in my opinion.
Michael Neale
product engineer would go through: keeping your mouth shut over
something really exciting and cool for a couple of weeks is nothing
compared to certain death (dismissal) for a product you are beavering
away on for YEARS.
Casper Bang
gestapo methods, controlling leaks while carefully preparing gigantic
hype-fests with Steve Jobs as high priest. Google just invites 100
journalists, hands them a new superior phone, and says "Oh btw, people
can buy it now from google.com/phone and we'll ship so you have it
tomorrow".
/Casper
Chris Adamson
new filesystem suited to the nature of its core products: flash memory
in the 10-100 GB range for iPods and iPhones, and SSDs for laptops as
they get more affordable. Once we take away the assumption of ferrous
oxide on a spinning disc, lots of things are presumably in play, and
ZFS' advantages may not apply. More at:
http://www.devwhy.com/blog/2009/10/24/the-loss-of-zfs.html
On Jan 6, 8:06 pm, Fabrizio Giudici <fabrizio.giud...@tidalwave.it>
wrote:
Fabrizio Giudici
> There's some speculation that Apple would be better off developing a
> new filesystem suited to the nature of its core products: flash memory
> in the 10-100 GB range for iPods and iPhones, and SSDs for laptops as
> they get more affordable. Once we take away the assumption of ferrous
> oxide on a spinning disc, lots of things are presumably in play, and
> ZFS' advantages may not apply. More at:
> http://www.devwhy.com/blog/2009/10/24/the-loss-of-zfs.html
>
HFS+. I have at least one customer that runs a large repository of media
that are used by a web application on Mac OS X (very bad idea, in my
opinion, anyway it's up to him) and I don't see any SSD in that place
for many years. And the blog, correctly, says that one FS doesn't fit
both roles, the "data center" and the "exchange media". So, what would
Apple be working on? Back to the original topic, we don't know at all.
PS Reading that blog I've just learned that I'm "psyched".
Michael Neale
their mouths shut in order to have the best splash marketing wise, in
many ways that is good business.
I think the difficult bit is that for apple to do this as well as it
does (and it does it VERY well), it seems to have to penetrate the
business in every single area - where by default things are
supersecret and closed - which I think it where it tends to get
annoying. But its Apple, they are free to do that if it works for
them, and if you want to work there, you have to live with that.
Steven Herod
on my old MacBook. Stuff I've never had with NTFS (i.e. failing to
boot for no reason, and having to recover using DiskWarrior).
On Jan 9, 11:17 pm, Fabrizio Giudici <fabrizio.giud...@tidalwave.it>
wrote:
Joe Nuxoll (Java Posse)
Michael Neale
wouldn't be too hard to not let it out. But it would be hard to get
the grin off my face knowing at what was coming.
On Jan 12, 1:15 pm, "Joe Nuxoll (Java Posse)" <jnux...@gmail.com>
wrote:
Fabrizio Giudici
> On the subject of HFS+, I've had numerous issues with the filesystem
> on my old MacBook. Stuff I've never had with NTFS (i.e. failing to
> boot for no reason, and having to recover using DiskWarrior).
>
you get a few bugs with the finder, after I repeatedly discovered that
some files were getting corrupted and the operating system didn't gave
me any warning (of course, also the check with DiskUtility reported no
errors).
--
Fabrizio Giudici - Java Architect, Project Manager
Tidalwave s.a.s. - "We make Java work. Everywhere."
weblogs.java.net/blog/fabriziogiudici - www.tidalwave.it/people