iPhone Copy and Paste
19 July, 2008
Many people have complained the iPhone does not yet have cut, copy and paste. It seems so easy/trivial to people, so why hasn't Apple added gestures for these actions? Leo LaPorte complained about this on the MacBreak Weekly recorded on 16 July, 2008.
Mac news sites are reporting that Greg Joswiak of Apple has stated copy and paste is simply not as high on the priority list as other features. Is this the real reason?
I don't think so. To Leo and all of you hoping for these new features, I don't believe it's as easy as you are thinking. Coding a gesture or two into the iPhone OS is not difficult, but there are other factors. Let's look at some of them.
First of all, what are gestures? Apple seems to refer to the entire iPhone interface as gestures, but this is a stretch. After all, touching a button is touching a button. There is nothing "gesture" about that. True gestures are things like the pinch zoom. Use two fingers in a precise manner to zoom or shrink a picture. This harkens back to the original ideas for gestures. Clear back in the late 60s and into the 70s, people working on computer interfaces envisioned a camera hooked up to a computer. The computer would recognize a user nodding her head yes or no to a question posed by the computer, thus reinforcing what the user types in. Think Hal 9000.
Gestures such as nodding your head are intuitive, natural and in most cultures obvious. The gestures on the iPhone/iPod Touch are do not fit this definition. Yes, nodding your head is a learned action, but it applies to life away from computing devices. Babies learn to nod. Give a Windows user a Mac, and ask them to do some tasks and they can figure out a large percentage of work needed to be done. Give an iPhone to someone who has never use one, even if they have used other smart phones, and they will have problems getting around. If you ask them to find a specific photo and enlarge the photo, they will likely be stuck.
This is not a criticism of the iPhone interface. I make this point, so you will realize developing a touch interface is not easy. Apple has to come up with gestures that once learned, seem logical and are therefore relatively easy to remember.
So, are the gestures for cut, copy and paste easy to define? Nope. What will be the gesture for any one of these three interface concepts? Remember that before cutting or copying, you have to select an object. When the object is a block or range of text, the gesture shouldn't be too difficult. A single finger can indicate a range of text from point A to pint B. A swipe can do the same thing. But, will this be confused with other gestures. So, do we specify that two fingers be used? How do we make these gestures consistent with the gestures for cut, copy and paste?
And how do we select things other than text? How do we select one or more photos or elements of a drawing? Touching a photo will probably have significance other than selecting it in most applications. Will each photo have to have a little box indicating you are selecting it? And selecting an object will be important, if Apple is going to add the ability for objects to have contextual menus. Which gestures will be used for selection, cut, copy and paste will being logical enough that once learned, they will be easy to remember? Do, we specify two fingers for selecting objects, and which two fingers?
Now, I will throw in another level of complexity. Remember Apple is trying to come up with a touch interface that will do so much more in a few years. They may come out with a tablet in a couple of years, a device expected to do a lot more than the iPhone or iPod Touch. No, it won't be a computer, but it will be expected to be a good input device for text and much more.
And, Apple will have to have a complete set of gestures for such a device, not just cut, copy and paste. So, before Apple settles on gestures for object selection and cut, copy and paste, they must think several years ahead and make sure those gestures are consistent with anything else they may need to come up with. Apple has been working on this interface for some time now.
No, adding cut copy and paste is not a trivial exercise. So, it's difficult to guess when these gestures will be added to the iPhone interface. Should we even discuss whether the current touch hardware can support the touch resolution needed for whatever set of gestures Apple settles on?
And as was pointed out in Leo's podcast, what about the clipboard needed to transfer the copied objects from application to application? Apple has sandboxed iPhone applications, isolating them from each other. If they were to create a clipboard bridge, would security be compromised? And the methods for performing cut, copy and paste could be automated for things like text boxes, but what about other objects? Would the methods require special implementation by developers, many of whom are new to Apple development, and wouldn't understand how all of this works in the Apple world. Maybe Apple is trying to avoid a situation in which some applications support these methods and others do not.
And, a few things really needed to the iPhone OS first. Am I the only person greatly disappointed Apple did not add file synchronization to iPhone OS 2.0? Do we still have to email notes to ourselves? What good is cut, copy and paste if non-trivial text can not be directly synchronized to your computer? As a developer, I have lots of ideas for iPhone apps, but many of them will rely on some easy way to get data to and from the iPhone without having to rely on an email interface. I suspect that this type of synchronization is not trivial either, and for many of the same reasons.
Oh, and for those of you who think Apple will come out with a tablet device in the next few months, reread this article and think about the factors I have written about in past articles.
MacWorld 2008 Wrap-up
16 January, 2008
It was nice to have a year in which MacWorld announcements followed my predictions closely. Here is a summary of the products announced by Steve Jobs at the keynote.
One week prior to MacWorld, Apple released a press statement detailing the new Mac Pro towers. The towers now come standard with 2 quad core processors. Options include a single quad core chip or you can move up to dual quad cores running at 3.2 GHz. They even have an optional graphics card that is top of the line in th eindustry, but costs $2850!
At MacWorld:
1. Time Capsule -- is a WiFi base station. Essentially an Airport Base Station with "N" speeds combined with a hard drive for back ups; this is cost effective for those using the Time Machine capabilities of the Leopard operating system.
2. iPhone -- Jobs mentioned the new developer tools for iPhone coming in February. He discussed a free update for the iPhone that includes better Google maps integration and fairly accurate position determination using cell towers and WiFi signals. The update also includes the ability to SMS with multiple people, WebClips, which are live clips of web pages, the ability to rearrange icons on the home page of the iPhone, and to create up to 9 home pages.
The iPod Touch was updated with applications for Mail, Maps, Stocks, Notes and Weather, but this collection will cost current Touch owners $20.
I was a little surprised the iPhone was not updated to 16 GB of storage. But, I have heard via a source the circuits inside the iPhone are so dense, Apple cannot add memory at this time. I was very disappointed the software updates did not include better integration of Notes with the Notes found in the Leopard operating system. Users still have to do some manual manipulation to get notes they started on the iPhone back to their Mac or PC.
3. Movie Rentals -- lots of discussion of iTunes movie rentals. Apple has a deal with most of the movie studios, including Touchstone, Miramax, MGM, New Line, Lionsgate, Fox, WB, Disney Paramount, Universal, and Sony. Users can rent a movie, start playing within 30 days of downloading it and watch the whole movie within 24 hours of starting to watch. $2.99 for older movies, $3.99 for new releases, older HD releases for $3.99 and $4.99 for new releases.
The cost is reasonable, but I was disappointed by the 24 hour limit. And more DRM free music was not announced.
4. AppleTV 2 -- As I mentioned to a few people, I expected Apple to and they did update the AppleTV via software, a free update for current AppleTV owners. The update allows better sharing of movies, pictures, podcasts and more from your computer. More importantly, it allows you to get items directly from the iTunes Store directly from the AppleTV and your TV screen without a computer acting as intermediary. You can download podcasts from iTunes or purchase movies, TV shows and music directly from the Apple iTunes Store (online). You can also go to .Mac and Flickr and display photographs.
A very nice update that will help AppleTV sales, as will the price drop from $299 to $229.
5. MacBook Air -- Apple introduced the new MacBook Air, the star of the show. This new MacBook weighs just 3 pounds, yet has a 13.3 inch screen. It has no optical drive (DVD), a single USB port, a mini-DVI port and an audio out port. It relies on wireless for networking; to get an Ethernet port, you must purchase a separate USB to Ethernet dongle. Due to the need to limit heat, processor speeds are limited to 1.6 or 1.8 GHz. expect the battery to last 5 hours!
The amazing part is its thickness. With the lid closed, the maximum thickness is 3/4 inch, tapering to barely over 1/8 inch at the front edge. People at the show have reported its elegance. It is designed for people who do a lot of email, documents, presentation and web searches. It is not for people with the need for many peripherals or heavy computations.
As I predicted, it uses a tiny 1.8 inch hard drive that stores 80 GB. I also said they would not use Flash memory for storage. Well, they do offer a 64 GB SSD (solid sate drive) as an option, but it costs $999! The RAM is fixed at 2 GB with no slot to increase the total to 4 GB. The battery is fixed and cannot be swapped by the user. If you heavily use this Mac and need a battery replacement in 2 or 3 years, a service call will be required.
The MacBook Air also comes with an expanded trackpad that responds to gestures. For example, you can use the "pinch-zoom" as used on the iPhone to expand and shrink photos. I don't think this has fully been absorbed by the general Mac public; this type of input will expand in the coming years as the hardware gets a bit cheaper.
The MacBook Air costs $1800, a couple hundred more than I expected. Perhaps the cost is due to the tiny components and custom machining of parts. Or, it may be due to the new screens and circuit boards used; Jobs said this MacBook does not contain or use in its manufacture various harmful compounds an elements, such as arsenic, mercury, bromides, PCVs and so on. The aluminum case is fully recyclable. That's a good start!
So, a good MacWorld for Apple and a disappointment for all of those people who really thought Apple would pull another iPhone out of its hat.
Pre-MacWorld 2008 Notes
13 January, 2008
Just a few notes before MacWorld 2008. One of my predictions has already come true. I said that Apple would not introduce new Mac Pro towers at the MacWorld presentations. So, my record for predictions for 2008 is now better than last year!
Apple introduced the new Mac Pros last week in a press release. Very low key, clearing the way for announcements in 2 days. I will stick to the rest of my predictions, including the idea MacWorld will be heavy with announcements in the area of music, movies and the iTunes Store.
In my previous article on predictions for MacWorld 2008, I did not go much beyond the MacWorld time frame. The distant future (say 6 months from now) is difficult to predict, because Apple will be guided by market forces, not by the prototypes found in their labs.
I *do* believe Apple will announce an iPhone with 3G data speeds during 2008. But, the announcement will not come before the WWDC in June. There is some possibility Apple will skip a 3G phone and go directly to a 4G phone. But, this possibility is low as it will take longer for cell phone companies to decide on 4G standards and deploy these capabilities than they are now admitting.
I do not believe we will see a larger form factor touch device during 2008. The hardware technology cannot yet be delivered at a low enough cost, the interface software for more complete touch interfaces is not ready, and Apple cannot afford to release another touch interface device, when the general public has not completely absorbed the idea of the simple touch interface found in the iPod Touch and the iPhone.
But, we will see more touch devices in the coming months. The sales growth of MP3 players has slowed apparently due to the beginning of market saturation. Apple will transition iPods towards PDA type devices with more and more features for personal data management. Better management of calendars, personal data bases, and the ability to input text and tweak presentations will be among these. The abilities of the iPod will remain, but will be considered an essential add-on.
As Apple develops these new-age "Newtons," the devices will split into 2 lines. Those with cell phone abilities and those without. Not that the public's need for cell phone abilities will go away soon, but many people's communications needs will be satisfied by advanced Wi-Fi networks coming in the next few years.
So, we will see 2 lines of small touch devices differentiated by cell phone abilities, not unlike the differences between the MacBook laptops and the MacBook Pro laptops. Not everyone needs the ExpressCard slot, the faster processors and graphics, and the video out capabilities of the Pro laptops.
Will we see the touch devices with 50% to 200% larger screens than the iPhone and iPod Touch? Hmm... Maybe not. Larger screens may not make touch input including text any easier, and people do not want to carry a device any larger than necessary. So, a larger touch device is not certain, even though such experimental devices have undoubtedly been in Apple labs for several years.
And in closing, what about the future of the Mac? The Mac will also transition towards touch interfaces with the multi-touch track pad a first step. As hardware comes available, the Mac could easily transition towards pad devices with full screen touch interfaces and optional keyboards. But, how will the Mac OS X human interface transition beyond the touch interface? Well that's the real question!
Prognostications for MacWorld 2008
30 December, 2007
Last year, my predictions for new products at MacWorld could not have been more wrong. The Steve Job keynote was all about the iPhone, which I did not predict for a few more months. The "Law of Averages" indicates this year's predictions should be spot on.
New Product Development
I believe this year's MacWorld keynote will be used to announce a number of upgrades and some new products, but no products defining new categories in spite of rumors floating around the 'Net. I don't believe people starting those rumors take into consideration many factors determining the products Apple will release.
The biggest rumor is for a large iPhone or iPod Touch type of device with a 5 to 9 inch screen (depending on rumor source), full touch capability, and functionality that brings it closer to that of a PDA rather than a phone or a music/video player. Perhaps it would be the 'Newton Touch'.
I will not wait to tell you I believe the likelihood of such a device to be rather remote. Many factors will keep it from seeing the light of day for at least another year. Let's look at these factors, which affect Apple's product development and deployment.
One set of factors is Apple's ability to integrate a brand new product into their sales support, marketing, warranty support, and so on. The sales support and warranty support staff need to fully understand products, so they can help the customers. Marketing people need to fully understand products, so they can differentiate these products in their own heads before trying to market these products to potential customers. In all cases, these Apple groups are still trying settle in with the iPhone and iPod Touch following introductions outside of the US. And, let's not forget the effort required to set up product channels for whole new products.
The next factor is the ability of Apple to make this product. Are the hardware components available? Well, a large touch device would need a large touch screen. Apple would need a large touch screen that is cost effective, reliable, of high resolution and in very large quantities. I don't think such a touch screen yet exists. The large touch device would also require copious quantities of Flash storage; I would not expect the consumer to be satisfied with a hard drive based device. Of course, Flash storage exists, but the supplies are constrained. We are constantly hearing the sales of the iPod and iPhone is straining the Flash now being manufactured. At this time, Apple doesn't need to release a new device that strains Flash supplies even more.
And as far as Apple being able to make this device, what about the software, the system support software? A large touch device with more PDA type capabilities than the iPhone or iPod Touch would require more touch capabilities integrated into the system software. The touch software would have to support more advanced gestures (true gestures) for cut, copy, paste, and other text editing needs, and so on. I do not think these capabilities are yet in their final form in the Apple labs. I expect some of these capabilities to be more slowly released in the iPhone and iPod Touch before a leap to a more capable Newton Touch.
Cost -- the large touch device would be expensive with its large touch surface and copious amounts of Flash storage, more expensive than customers will accept.
While throwing in these various factors I will mention product strategy. To be brief, does Apple need a Newton Touch at this time? No! Apple still needs to develop the iPhone first, so that it becomes a clear market technology leader among smart phones. And further convince people technology such as the touch keyboard works fine. The iPod remains the technology leader among MP3 players. So, no new device is needed at this time; introduce the Newton Touch now, and Apple's resources will be strained, while telling the world "Hey, we (Apple) have introduced a whole new device that proves we can make neat stuff, even though you do not understand how it's high cost will make your life better." Apple has done this more than once: remember the Newton?
Music
So, let's start 2008 predictions with existing product lines. MacWorld has always had an emphasis on consumer products, so we will see a number of tweaks to the iTunes Store. Jobs will likely announce more music labels selling DRM free music at higher sample rates. This trend helps reduce the hassle getting music moved to iPods and other devices, while delivering music with greater quality. Wally-Mart and HP have just announced they are closing their DVD online store. Most observers believe they have had poor sales due to high prices and the use of Windows DRM the iPod cannot use.
Even though Universal has defected from the iTunes Store, I hope Apple will introduce movie rentals at the $3 price point. Various factors have pointed in this direction, including changes in the iTunes software. I have always felt people prefer to purchase their music but rent movies, so this is a needed addition. If Apple introduces movie rentals, then we should expect the AppleTV system will be upgraded to handle rentals and downloads directly from the box and not need a computer as an intermediary. AppleTV needs an enhancement like this to spur sales.
Many Internet pundits agree the news in Apple's world of music and movies will focus on the iTunes Store and associated software. I am not expecting any iPod upgrades; they could increase the amount of Flash storage in some of the models, but don't count on it. Flash RAM supplies remain a bit constrained. Expect a couple of tweaks to the iPod Touch operating system.
iPhones
On the other hand, I am expecting a new iPhone model or the current iPhone to be upgraded to 16 GB of Flash storage. The iPod Touch has a 16 GB model, and the iPhone needs to match it. I am also expecting a nice upgrade to the iPhone system software. People are clamoring for new features, and the iPhone upgrades thus far have been for bug fixes and tweaks. What new features will be added? That's really hard to say. I don't have an iPhone, and so am not invested in the outcome. But, there are some features that would help iPhone sales, such as note synchronization via the iTunes application, better input for calendars, and other PDA like features. I would like to see an iChat application added, but Apple's contract with AT&T may prevent this. I am hoping for more than just tweaks to the Google maps capabilities.
Many people are expecting the announcement of an iPhone with G3 data network speeds. I am not. The people expecting this assume Apple will need a 5 or 6 month lead time for FCC approval. Only the first iPhone required this much time, and Apple can ask the FCC to keep new approvals secret, not caring if a little information leaks out. Apple will not officially announce a new iPhone so far ahead that it will cut into sales of the current model. The 3G iPhone may not come until mid-year.
One sure thing will be Jobs' discussion of the new developer's kit for iPhone. At this time, I expect the application development for iPhone will be based on HTML, CSS and so on. Full Cocoa support may be a few months off.
Macs
So, we come to the Mac, the original focus of MacWorld. Here, the predictions become difficult. Intel has some new processors that could be used in the Macs, but they are primarily for the higher end machine, like the Mac Pro. And in recent years, Apple has not used the occasion of MacWorld to update the "pro" line of Macs. Some people out there believe Apple will drop the Mac mini, but I hope not. Many people have found it to be a very useful machine at a good price point. Many Windows users come into Apple stores to see the iPod, and Apple needs the low Mac mini price to help create new Mac owners and draw them towards the higher priced Macs.
There is another possibility. I have heard via the grapevine, but not on the web Apple has developed a desktop machine somewhere in capability between the Mac mini and the Mac Pro. Apple certainly needs a machine to fill this gap for those who do not want the all-in-one solution of the iMac. The new Mac might use 3.5 inch hard drives instead of the 2.5 inch hard drive used in the iMac and mini. I hope it will also have a full size slot with enough room for the largest and most capable graphics cards. Such a Mac would be purchased by gamers, engineering companies and many in between who need to run Windows along side OS X. The biggest factor against its release is Apple's desire to keep the Mac line of computers lean and mean.
In a recent podcast, Merlin Mann pointed out that Apple will be guided by market forces, and the continuing desire to pull people from Windows to the Mac. This will also be true in laptops, and Apple sells more laptops now than desktop machines. Apple's MacBook and MacBook Pro could be updated at MacWorld; don't expect major changes. The biggest rumor concerns a new ultra-portable that could join Apple's laptop line. An ultra-portable would weigh 3 to 4 pounds and would have a 10 to 13 inch screen, but no DVD drive.
Many people think this machine would have no hard drive, but instead a 64 GB Flash drive. I am a bit dubious due to the high cost of the Flash storage versus a hard drive, and the constrained supplies of Flash modules now being manufactured. If a laptop appears with only Flash memory, expect to pay a premium over a MacBook, while enjoying significantly longer battery life. But, another possibility exists. Toshiba recently announced a ruggedized 1.8 inch hard drive series that could be used in an extra small laptop. Or, imagine a 1.8 inch hard drive combined with 16 GB of Flash storage. This *might* be possible if Apple has managed to integrate the zpool capabilities of ZFS volume management into Leopard as I believe they have been doing. But, this is the subject of a different article.
So, an ultra-portable MacBook may be the most realistic prediction for a whole new product at MacWorld, yet I am dubious due to the constrained supplies of Flash storage. Expect a 12 inch MacBook Pro with a micro hard drive. The least realistic prediction is the rumor of a new touch device like the iPod Touch, but designed to be a tablet PDA. The whole first section of this article describes why I think it is highly unlikely Apple will come out with a tablet touch device during the next 12 months. The most exciting product we can realistically expect is a new MacBook or MacBook Pro with a large multi-touch pad for inputting simple gestures (for now) like the pinch used on the iPhone and iPod Touch.
Summary
Well, this covers my basic predictions for MacWorld 2008. They will concentrate on music, movie rentals, iPhone updates, maybe a new Mac or two, and throw in a new display or other peripheral such as an external BlueRay player. Of course, there will be plenty of announcements outside of Apple's products; the big one will be the new version of MS Office.
I really need to score some runs this year after last year's complete surprise. But who knows? Maybe Apple will come out with the Beatles iPod *this* year. I am going to keep predicting that one until it happens!
Leopard Under the Hood
7 November, 2008
Now that Leopard is installed on my PowerBook, it is time to write the article I meant to write this past summer. It concerns the biggest changes found in the new version of the Mac operating system.
No, the biggest changes are not Time Machine or QuickLook or the integration of Notes throughout many of the applications. The biggest changes are under the hood.
Now to a developer, "under the hood" does not refer to new utilities, such as Core Animation. Such utilities are very useful to developers creating programs, because they contain code the developer doesn't have to write. And, developers like the fact that these utilities are available to all programs, and updates are supplied by Apple, so the developer can concentrate on the special code that makes the application what it is. Imagine the developer lives in a world like the Matrix. These utilities are like the buildings surrounding you as you walk down the street. The developer doesn't have to go into the building to take advantage of the services created by the business within that building. No, to a developer, the real OS is all of the plumbing that lies below the street.
OK, that analogy is going way way too far, but the point is that developers are excited by the changes in Leopard found beneath the street. And, there are a lot of changes.
Leopard is now UNIX compliant, and like Sun Solaris it can be called "UNIX" and not just "UNIX-like". This is important in enterprise arenas.
The OpenGL engine in Leopard has been improved and now meets the Open GL 2.1 standard with a new virtual engine that runs as much graphics processing as possible on the graphics card, depending on the Mac and the graphics card it is using.
Leopard includes a new "sandboxing" system that will help reduce the hackability of the system. Not that Mac OS X was bad to begin with, but security is a moving target.
All of these changes are important, but 3 changes to the system really stand out, and the third most people have no idea is there.
The first of these changes is memory management. A primary job is to manage the memory in a system, swapping processes in and out of memory, using virtual memory, and so on. Apple has stated that "automatic garbage collection" is now included in the development system included with leopard. That sounds simple, but there is much more to it.
Automatic garbage collection requires changes to memory handling through the operating system. In Tiger and previous OS X versions, the user can be distracted by the "spinning beach ball of death", which most often twirls in the user's eyes when an application needs more memory and the system must move stuff to swap files to make room in RAM. Sometimes, this means moving pieces of the current application out to virtual memory, and then soon moving them back to RAM.
Apple's engineers working on the core of OS X have wanted to work on memory management for a long time. So, they took the time to do it in Leopard, and this is one of the reasons leopard took more than 2 years to get out the door. But, memory management in Leopard is smarter, knowing better which chunks of memory to swap out and which to keep. It is also more "fine-grained". When moving stuff in memory, it is no longer necessary to large pieces that may contain application code that might be needed 2 seconds later. Stuff can be moved around in smaller chunks.
So, I expect the user will notice that the SPOD will twirl less often under Leopard. In my own brief tests, I like to torture Safari, opening numerous tabs to web pages containing all kinds of stuff such as flash animations. Under Tiger, I could get Safari to slow to a crawl after a few dozen trips to the web. Leopard's Safari is much more difficult to clobber, and based on other subjective tests, I believe this has to do with these improvements under the hood more than any improvements to Safari specifically.
The second big under the hood change in Leopard is process and thread management. Processes and threads have to do with the work an operating system is asking the CPU to do. OS X is a fully threaded operating system. This simply means that one process can get started while other processes are continuing to execute. Processes on Apple's OS 9 and previous systems were not completely independent of an application, and depended on the co-operation of each application. In OS X, processes such as printing can be handled independently, while you go on with some other work.
In Leopard, the management of processes has been improved. Apple admits that the scheduling of these processes is improved. This means that the OS is more intelligent about which processes get time on the CPU. "OK process A, you get 2 more milliseconds, and process B gets 50 microseconds." And, the scheduling had to be improved more than simple measures. The Macs are now completely dominated by CPUs with multiple cores, and it is the process schedular that decides that certain process can use core 1 and 2, and another process can use core 3.
Of course, there is a lot more to it than this. Threads are little bits of processes that run on a given core, and a developer can explicitly define processes and threads or the operating system can infer these pieces for itself. Leopard is much better at this inference, and so everything will run better, more smoothly and so on. Less twirling of the SPOD. Apple *had* to make this stuff better now that Macs contain all of these multiple cores. The core or kernel of the operating system itself was tweaked throughout to run better and faster with all of this new memory and process management.
But, one more improvement was put into place. This is an improvement no one has talked about. Without having the time to tear Mac OS X apart, I cannot tell how much of this new improvement has been put into place, but I feel sure it is there.
This new improvement is ZFS or the Zettabyte File System. A file system is the thing that tells the operating system how to manage storage, such as a hard disk; how to construct the indexes and pointers, how to divide a file up into available space in such as way that the user never knows the file is stored in pieces on the hard drive. Files systems you may be familiar with are HFS+ on the Mac and NTFS or FAT32 on Windows.
And ZFS is a file system in these ways, a file system with greater capabilities. Where these other file systems can manage hard drives of gigabytes or even terabytes of data, ZFS can manage hard drives so large the numbers are meaningless.
And, so the pundits out there are asking in their own blogs, why is Apple messing with ZFS. Do we really need another file system, and the changes that implies? But , they miss the point entirely.
ZFS, in spite of its name, is more than a file system. It is also a volume manager. A volume manager is the part of the operating system that says, "Let's divide that hard drive into 2 partitions, each an independent volume" or "Let's RAID those 4 hard drives together for better performance."
ZFS as a volume manager can do so much more. It changes the concept of volume manager to include all of these traditional ideas and much more. It contains the concept of a "zpool" which is a collection of storage (almost) unbounded in size. More than one hardware device can make up a zpool, and another can be added when ever the user wishes.
For that matter, a zpool can be any type of storage. A zpool of 2 hard drives can be supplemented by some Flash RAM storage and a removable disk. And a zpool can be used to create a storage volume or more than one volume.
And I believe Leopard contains the basic volume manager from ZFS. It is not a mature volume manager, and cannot do the many more things ZFS can do but not described here. But, it can handle basic volume manager chores.
My simple evidence includes the zpool utility that appears to be within Leopard. Go into the Terminal and type "man zpool" and have fun reading the details. The zpool command is in Leopard.
Why is this important? Well, it lays the ground work for some really nifty future improvements. But, it also could give us some intriguing capabilities right now.
Remember I said a zpool can combine different kinds of storage, such as a hard drive and Flash RAM? Flash RAM is too expensive right now to equip a MAC with 128 GB of Flash RAM, thus replacing a hard drive. But, what if they combined a micro hard drive with a smaller amount of Flash RAM? Using ZFS, this two forms of storage could form a single volume that is fast, and uses less power while costing less than pure Flash RAM. Windows Vista supports special hard drives equipped with Flash RAM, but Leopard would not need special Flash RAM or special hard drives. Think about this in a very slim Mac laptop; a very special idea!
Leopard Delay
30 April, 2007
Well, I will have to stop making predictions of Apple product releases (OK, that won't happen). In mid-April, Apple officially announced the Leopard operating system or OS X 10.5 will be delayed. The likely timeframe of the release will now be October.
Apple stated the delay is due to the shifting of resources from Leopard development to the iPhone. Key OS X engineers were needed on the iPhone due to some problems encountered getting the iPhone version of OS X to the level of quality demanded by Apple.
People had been suspicious of the Leopard time table in recent weeks, because the number of bugs in recent beta releases of Leopard had not decreased. Apple will probably take the opportunity to complete a number of features that would not have made the initial release of Leopard. Rumors indicate Apple is working on a substantial update to Quicktime that would not have been ready until October in any case.
Leopard should be a very substantial upgrade to OS X, including features in the core of the operating system not normally visible to the average user. These core improvements may include complete 64-bit support for the newest CPU chips, automatic garbage collection, better multi-thread support for multi-core chips, an improved version of OpenGL for graphics, a new Core Animation layer and improved core layers, including Core Image. All of these improvements will help performance and should reduce occurrences of the dreaded spinning beach ball of death. But, it is these types of improvements that can be difficult to implement, because they affect the entire operating system. Apple is also adding a feature called Resolution Independence, difficult to implement.
I am most excited about the automatic garbage collection and improved multi-thread support. The addition of automatic garbage collection APIs implies work is being done to improve the memory system in the OS. OS X tends to do a lot of memory swapping, when you would think little is needed. The heavy memory swapping is also caused by not having enough multi-thread support. This problem is sometimes caused by old Carbon code or simply by modern code needing more work. There are APIs Apple documents as being "not thread safe." This may be caused by the need in those APIs to make calls to old Carbon routines. Or it may be that the whole system simply needs better multi-thread support. I do know that developers need help in the OS and the developer tools to easily implement fine-grained threads.
And, while we are discussing Leopard features, vague rumors on the Internet would indicate Leopard might, might, might contain the first cut at ZFS. ZFS, formerly called the Zeta file system, could allow Apple to create a very lightweight laptop with a hard drive smaller than 2.5 inches and flash memory combined into a single storage pool. ZFS would point the way to the future with greater support for metadata, among its many abilities.
Some rumor web sites have assumed the Leopard delay means big problems for Apple. These sites imply Apple is not paying attention to the Macintosh product, because they haven't added a bunch of engineers to their staff. This cannot be assumed! Apple operates a lightweight development staff for more efficient effort. They cannot simply add programmers and engineers to an effort and expect work to progress, as has been suggested by these web sites. The key people needed on both Leopard and the iPhone are experts on hundreds of thousands of lines of code in both products.
Adding programmers and engineers takes time. The best programmers need plenty of time to get up to speed and familiarize themselves with complex code, such as OS X. And, the best are needed. I have led development efforts for government contracts for which average programmers were hired. I would have been much happy with half the number of more-experienced programmers. And, although those experienced are paid more, the total cost would have been lower.
The delay brings advantages to future development for both the iPhone and Mac OS X. We will assume for the moment the iPhone will be using the kernel and many objects from Leopard, rather than falling back to Tiger. By solving all of the little problems in Leopard-iPhone, Apple can fold the fixes into the general release of Leopard for the Mac. This will help keep Mac OS X a single development effort, rather than splitting into two different products. Why am I pretty sure the iPhone is using Leopard technology? In previous paragraphs I mentioned the better memory handling and thread support going into Leopard. These are very necessary for a device like the iPhone using small CPUs that need to benefit from every bit of efficiency these improvements can bring.
The reasonable question is why Apple decided to announce the iPhone prior to the release of Leopard, causing the release of the two products to be within a month or so of each other. Perhaps a delay of the iPhone would have caused a big problem for Apple trying to get into the cell phone market.
The bottom line is that the iPhone should still be delivered in June with perhaps a one or two week delay over Apple's original intentions. And, we should learn about all of Leopard's features at the World Wide Developer's Conference (a conference for developers of Apple compatible products) during the week of 11 June.
You can also expect a nice version of OS X in October. With the addition of automatic garbage collection, and better multi-thread support, Leopard will have a strengthened memory management system. This will result in better performance, fewer instances of the spinning beach ball, and much better use of the multi-core CPUs Intel is introducing.
12 September Apple Announcements
14 September, 2006
Steve Jobs took the stage to announce upgrades to the current lines of iPods, make available full length movies for download, and in the most surprising move, pre-announce a product! Apple hasn't pre-announced a product in a couple of decades.
The pre-announced product is the "iTV", a device that will allow multimedia content to be streamed from your computer to your TV and home entertainment system. The name will likely be changed as will the final specs. I am sure Apple is hoping to include the new 802.11n technology, so there is enough wireless bandwidth to stream high definition TV.
I am surprised about this product, but it certainly seems to tie into my comments of August 30. This product could easily evolved into the home based server I mentioned with the addition of a hard drive and operating system. Of course, the legal problems of sharing multimedia content in this manner will first have to be overcome.
Fear of Technology
30 August, 2006
A couple of years ago, I wrote that Apple could create a small home server. This could be either a pure server, one that is not used as a workstation and has no screen or Apple could add the server software to a desktop Mac, one that is always up and running.
This server would be used to collect email, store photos, download TV shows and movies similar in function to TIVO, and would act as a communications center for members of a family or small business. It would store account information and user specific files for multiple users.
This device would be useful for a household with a single computer user or a group of users. Imagine being able to pick up a small portable Mac, and logging in as any user found on that server. The portable Mac is automatically set up with a mobile account and populated with the files belonging to the logged-in user using a high speed connection. When the user returns to a location close to the user, the user's files are updated on the server, so any portable Mac can be set up with a mobile account for that user in just minutes. Another person can then log in and set up on the same portable Mac in minutes.
A portable Mac that operates in conjunction with such a server could be very portable. Using flash memory instead of a hard drive, the portable Mac would need smaller batteries. If the portable Mac were being used within wireless range of this server, the portable would not need an optical drive or I/O ports making it even smaller.
Some of the technology needed is coming along. The high speed wireless technology known as WiMax is now being implemented. Flash memory is increasing in size to the point that it could replace laptop hard drives in the next year or two.
I am sure Apple is working on some of this technology as an experiment, but now I am wondering if market forces will prevent this type of technology from coming to fruition.
www.Publish.com has reported the DVD Copy Control Association is suing Kaleidescape. Kaleidescape makes servers for the distribution of multimedia content within the home. If I understand their product, Kaleidescape stores the content from DVDs and other media on hard drives, and uses a network to distribute this content to various rooms. They do everything possible to remain legal, including making buyers of their systems sign agreements to only use content they have purchased legally. But, they are being sued for breach of contract and licensing. According to Publish.com, "... the DMCA prevents individuals or companies from bypassing content protection schemes; even, theoretically, for fair use purposes.."
The various movie publishers clearly do not know how to make a new business model in the face of changing technology. They simply have a fear of technology.
Could this keep Apple from creating a wonder system people would buy in droves? For now, Apple may be limited to creating an Airport Base Station that will transmit video wirelessly, like the Airport Express already does.
WWDC Observations
18 August, 2006
Earlier this week, Apple made announcements at the World Wide Developers Conference in California. Many columnists and bloggers complained that not enough announcements were made. To make things clear, this conference is for developers, who pay an average of $1400 to attend conference sessions designed to teach them the latest programming and administrative techniques for Apple's OS X. WWDC is not designed to be a platform for making big announcements.
Steve Jobs did debut new features that will be found in the next version of OS X version 10.5, also called Leopard. Leopard will come out in the "spring" of 2007. This is about three months later than previously assumed.
Jobs and friends described ten new features in Leopard. The entire list can be found at www.apple.com. I found the most intriguing to be Time Machine. It is an automated back-up application allowing old versions of files or trashed files to be recovered. It can also be used to back up the entire hard drive, including operating system. Third party utilities are available for this, but included as part of the OS, I feel more confident that Apple will create a reliable utility. But, the interface is the most important aspect of Time Machine. It appears to be so easy to use that anyone with a spare storage volume will take advantage of this feature. I think this feature alone will sell copies of Leopard.
But, Jobs mentioned that not all new OS features were to be revealed at the WWDC. It has been assumed this is to keep Microsoft from taking ideas from Apple, and implementing quick and dirty look-alikes for Vista. This is certainly one reason, but the other is to save some great features for future, splashy announcements. It is likely that some new features will be shown at the MacWorld in January, 2007, and yet others when Leopard is released in the spring.
At WWDC, Apple also released the new Mac Pro, a tower Mac powered by Intel chips as a replacement for the PowerMacs using the IBM built G5 chips. The new towers have two Intel chips, each chip containing two CPU cores. Let's clear up some misconceptions.
Apple has used the Dual Core Xenon in the new towers. Some people have wondered why Apple didn't use the Core 2 chips rather than the Xenon chips. This question seems to assume that Core 2 chips are faster than the Xenons; they are probably fairly equal, but good benchmarks are not yet out. The important reason Apple has used the Core Xenon chips is that Core 2 chips cannot be used in multiple chip configurations, so Apple's direction was predetermined.
The important thing to remember is the new Pro Mac is a very powerful machine built to take advantage of the 64 bit capabilities of the Core Xenon (and Core 2) chips. Is the Mac Pro faster than the PowerMac G5 it replaces? Good benchmarks have yet to be performed, but preliminary marks indicate the new machines are more powerful, but not overwhelmingly so. Some types of tasks are faster than others, but it seems the new units are 15% to 20% faster.
People considering the purchase of a Mac Pro should be aware of some limitations of the new Mac Pros. These computers use a new type of memory called fully buffered RAM. Apple is using this type of RAM, because they are using the Core Xenon architecture (many Intel chips), and this pushed them into using fully buffered RAM. This type of RAM is a limitation, because it is expensive compared to other types, especially when the RAM comes with its own heat-sink. Apple says you need those heat sinks for maximum performance.
Another thing to consider when purchasing the Mac Pro is the graphics card. You have a choice of three graphics cards: the NVIDIA GeForce 7300GT, the ATI Radeon X1900XT, and the NVIDIA Quadro FX 4500. I know of no other graphics cards that support the Mac OS for the simple fact that Windows is still based on the BIOS bootstrap system, rather than Intel's EFI used in the Intel powered Macs, which is similar to the OpenFirmware boot protocol Apple used in the PowerPC Macintoshes. Windows Vista does not use EFI, and Microsoft has stated that it will support EFI (UEFI 2.0) in some future version of 64 bit Vista , but not the 32 bit version. So it will be a while before graphics cards commonly support EFI, giving Mac Pro owners more choice.I also found some interesting differences between the lower cost 7300GT and X1900XT and the very expensive FX 4500. The lower cost cards are intended to be for gamers, and limit their OpenGL throughput, presumably optimizing DirectX throughput for gamers on windows computers. The FX 4500 is considered a "workstation" graphics card, not limiting OpenGL throughput. The high cost comes from the support NVIDIA claims to provide card owners, support gamers will not receive. As I learn more information about graphics cards and these system, the information will be added to this forum.
One other thing Mac Pro owners should know. The Mac Pro comes with four PCI-Express slots, but only one has support for sixteen channels of bandwidth. You see, like ATA drive interfaces going serial (known as SATA), PCI-Express is the serialized version of PCI. To provide greater throughput to the graphics cards, an increased number of serial channels are used for each card. The top end graphics cards can make use of sixteen channels, but if you add a second graphics card, the maximum number of channels it will be provided is eight, thereby decreasing performance compared to the first graphics card.
Well, that's the top of the announcements from Apple's WWDC. The next hardware announcements should be in September, when Apple introduces Macs using the new Intel Core 2 chips.