“TUAW talks about the growing swell of upcoming apps which will use Leopard APIs, and therefore, will require Leopard to run. I’m sure some users are wondering why developers are doing this, and some developers are wondering if they should do the same. Many early adopter end-users are planning to upgrade to Leopard regardless, but some might want to wait for various reasons. From the casual observer’s perspective, the obvious thing to do is support older versions of Mac OS X in order to have access to a larger market. In a sense, that’s true. But there’s more to the story.”
The developer has already decided there are things he wants to take advantage of in Leopard so Textmate 2 will be Leopard only while he will maintain the 1.5 series. I see no reason not to do that.
This is one of major reasons OSX and Linux can move along faster than Windows. MS simply has to maintain large amount of backward compatibility, due to uncomparably larger user base with gazillions of HW combinations. I am not advocating either way, it’s just the way it is.
Microsoft sets their own policies. And it isn’t that OS X is not offering backward compatibility, their old APIs are still there… I can generally run older apps on linux (even older gtk apps) because often the older libraries are included just for that reason – to support the older API.
Mind you, I am not exactly arguing against what you say, just that there still is backward compatibility of a sort.
There’s backward compatibility and then there’s backward compatibility. There’s a difference between being compatible all the way back to win16 and being compatible only to last week.
Ideally, software should be compatible to the previous platform release. That means new Mac apps should work on Tiger.
Backwards compatibility for Leopard would mean that apps written for Tiger or previous versions work on Leopard. It does not mean that the apps written for Leopard work on Tiger.
The linux Kernel is just as backwards compatable with hardware as Windows. Maybe even more.
The funny thing about Windows and this software compatability stuff is that it’s a myth. MS keeps all this software in Windows to be backwards compatable but out in the real world it doesn’t work like that.
There are companies that still run Windows NT because their applications don’t work in Windows 2000 (1 release ahead)
I am working on a project now where we have to move .NET web developers from Windows 2000 to Windows XP and it’s a nightmare. The funny part about it is that the developers are all using MS products.
First stupid thing we found is that you can not run IIS on Windows 2000 pro or Windows XP pro without being an admin. The next thing that we found is that MS made a full version of IIS 5 for developers on Windows 2000 pro but for Windows XP pro instead of putting IIS 6 in with Visual Studio they include IIS 5.1 With Windows XP pro. (Which is IIS 6 sort of) So they have problems doing tasks that require IIS 6. LOL!
So all this backwards compatability talk when it comes to Windows is just talk.
I assume it just means that if you make a Windows XP applications it won’t be hard to get it to maybe work in Windows 2000 (Maybe)
The linux Kernel is just as backwards compatable with hardware as Windows. Maybe even more.
That may be true but you don’t run applications on the Linux kernel per se. You run them under the window managers. So if KDE decides that KDE4 is a major refit then you have to decide to stay with KDE3 or move on to KDE4.
Right which is why I said “The linux Kernel is just as backwards compatible with HARDWARE as Windows. Maybe even more.”
Didn’t mention anything about software or applications.
But anyway, that is no different then if MS decides to upgrade Direct X or one of the other core technologies. Then you have to decide to install the new one to protect your self with security and get new features or you have to keep the old one to play your old games etc.
That even applies to Service packs. People are still running SP1 on XP for compatibility issues etc.
This happens all the time – look at all the apps that are incompatible with Panther for example. Sometimes there are two different versions. Home users are upgrading their OS anyway and business users usually have a predefined set of tested apps… So nothing new here.
It’s quite another issue that there’s such a huge Tiger userbase by now that makes matters more complicated.
I just bought a MacBook running 10.4. I’ll have to buy a full blown copy of 10.5 when it comes out just to keep my new machine new.
This same thing happened with my iMac running 10.2. To this day, I haven’t upgraded. I might upgrade it to 10.5. We’ll see.
Oh well. I’ve got Kubuntu under Parallels, so I’m happy. And my wife is a remedial user, so she’ll be happy staying at 10.4 for years to come. Bottom line, I’m just left with a slightly sour taste in relation to Apple, a working Linux OS, a happy wife, and a couple nice pieces of hardware.
I’m confused. Nobody is actually forcing you to upgrade, they’re just offering incentives. And without any of those incentives what would be the point of Apple (or anybody else) ever updating their software?
The point that he’s making is that once Apple moves on to Leopard, and indy developers move on to develop applications which take advantage of Leopard-specific features as well, there will be LESS incentive to update applications which are already made for Tiger or any prior version of OS X.
The only point in which you may not want to go from a previous version to Leopard is when you’re that completely satisfied with your computer and what it does that you simply don’t want whatever’s in Leopard.
That applies to XP and Vista as well: at this point, there is very little obvious incentive for the vast majority of people to upgrade to Vista from XP or 2000, save for DirectX10 and……I guess that’s it.
Plus, there’s a hefty number of people who received their first home computing experience with XP (like myself), more so than previous Windows versions due to MS’s abnormally-long delay on Longhorn/Vista.
IMO, it would help for software developers to release upgrades (updates are different) for their applications or operating systems only if they want to do something radical with them, or take it in an entirely different direction or to reach for a brand new audience than previously.
Maybe Apple is reaching for a wider audience with Leopard, one that wants to take advantage of its new “Intel-ligence”?
Yeah, that seemed a tad vague for a pun…..so does any target audience that Microsoft is reaching for with Vista.
That’s one of the major parts, yes. For me, it also feels like they’re pinching pennies out of users. Why not backport new APIs a version or two? Features like fast user switching, expose, dashboard, spotlight, better UI design in built-in apps, performance upgrades on otherwise compatible libraries… I can understand making that exclusively available in new versions of an OS, but app compatibility?
That just tears apart the “just works” factor. This is something I think Windows and OSS get right and OS X gets wrong.
If you upgrade OS X because you want new features, you’re upgrading out of a positive cause. If you’re upgrading because Apple made it so your new app is jiggered not to work unless you have the freshest, hottest stuff, you’re upgrading out of a negative cause. Which one makes for a happier userbase, and how does this relate to the popular negative views people have about Apple?
How much would it hurt Apple’s bottom line to buy some good PR?
Missing the point you say they as if they referring to Apple these are developers choices to make leopard only apps has nothing to do with Apple!
Apple releases cool APIs that are Leopard only, so developers can only support the latest and greatest.
If Apple backported the APIs, devs could keep compatibility.
Is This Called “Buyer’s Remorse”?
Nope. This is called “Whining.”
If Apple backported the APIs, devs could keep compatibility.
And what incentive would people have to upgrade Leopard then? How would Apple recoup the R&D costs to backport these APIs as well as develop a new OS? They certainly wouldn’t be getting money form the folks that already have Tiger. These folks would end up getting Leopard features for free.
Now let’s look at the developers side of thing: with Leopard, it’s a sure thing the API is included in the OS. If the new APIs are backported, now the developer has the responsibility of managing these dependnecies themselves. So rather than the app being a 2MB download, it becomes 20MB.
I’ll have to buy a full blown copy of 10.5 when it comes out just to keep my new machine new.
Please. Did your iMac stop working the day 10.3 came out? Didn’t think so. I just got a new MacBook too but I don’t think it’s going to be taken away from me if I don’t upgrade. I probably will upgrade just so that doesn’t happen. Oh, and to also take advantage of things like Java 6, Cocoa with GC, new multi-threading and things of that sort.
> If Apple backported the APIs, devs could keep
> compatibility.
You keep saying this as if “backporting the APIs” is done within a week of work. Fact is, you cannot backport the APIs without also backporting the underlying mechanisms, and every mechanism they depend on. Then you’ll have to test all that thoroughly. When you actually try this, the “backport” essentially becomes an update that turns Tiger into Leopard, because the main difference between the two is all the stuff that makes the new APIs possible. By your logic, Apple should give Leopard away for free. Why?
Let me put it this way:
Does anyone else think Apple could make good money giving away the base system and selling all the features for a price? I think people would pay to upgrade from the base Tiger system with no features to the base Tiger system plus every visible feature that’s been added since the last time they upgraded.
Do you realize how much work it would be to backport new APIs to older OSs? Microsoft can do it, because they have a veritable army of developers, but Apple doesn’t. Indeed, its streamlined development model is one of the reasons they can get an OS out there that competes with Vista while throwing a fraction of the development resources at the problem.
I don’t realize how much work it would be. Linux makes it look easy, but I’m probably misunderstanding. Why is it possible to compile, say, KDE 2 or 3 on either Linux 2.4 or 2.6 if backporting is hard? I know there’s a difference, I’m just not sure what it is, and I know you, rayiner, often post pretty good explanations.
With linux they are not backporting the API to the older library you are installing both the older and the new library on the same computer.
The app then just looks for and uses the correct library.
Take a look at /usr/lib sometime, especially on something like Gentoo where new libraries are added all the time. You may see many versions of the same library.
Mac is likely doing the exact same thing. Leopard will have the new Leopard libraries and also the old Tiger (and Panther) libraries so that it can run old apps.
The only real difference is with Linux the new libraries are free, so you can install them on your old OS. With Mac you have to pay for the new libraries and rather then selling them as separate libraries you have to download and install they package it all up together and sell it as a pretty OS.
Quicktime is an exception though. They upgrade to Quicktime 7 library and they give it to you as a free download and you have both Quicktime 7 and 6 on your old OS. They could do this for all libraries. They just would not make any money doing that if they gave it away and it is much easier to convince people to pay $150 for a shiny new OS then to pay $10 for each new library.
They just would not make any money doing that if they gave it away and it is much easier to convince people to pay $150 for a shiny new OS then to pay $10 for each new library.
I think they’d make plenty of sales based on new features while allowing older OSes access to new libs, and therefor new programs. People get excited about buying spotlight, expose, dashboard, time machine, spaces, etc. People get disappointed when they buy a piece of software and find out their brand new computer is too old to run it.
Which brings me back to my question, what’s the difficult part of making, for example, core animation available for 10.4, or even 10.0 for that matter?
Buy an ADC Select membership to get the answers on how it is that backporting isn’t pragmatic.
Which brings me back to my question, what’s the difficult part of making, for example, core animation available for 10.4, or even 10.0 for that matter?
I don’t know, but I would think that Core Animation builds on already existing APIs in Tiger and Quartz Extreme. Many of the things in Leopard are only possible due to the technologies available in Tiger. Backporting would probably also be difficult because of strong integration between the technologies. If you want to backport A, you need to backport B, C, D, etc.
It makes a lot more sense to use all resources to make a kickass system out of Leopard and sell Leopard with new Macs from spring 2007. If the sales of Macbooks so far is an indicator, then marketshare for Leopard by the end of 2007 would be very large due to many new buyers of Macs that will include Leopard. A large percentage of the rest will follow suit over the course of the year.
Apple, seems to me, want to compete or compare Tiger and Leopard constantly to Vista. If that is to hold true, they must do whatever they can to keep the distance from Vista with Leopard, rather than wasting limited resources, catering to older versions of OSX.
I would be much more upset, if only the very newest Macs would be able to run Leopard. Fortunately, it seems that even newworld G3 Macs will support Leopard.
Edited 2007-01-08 05:21
I think Apple would do well to make OS X appealing, not just whatever the latest version happens to be. They make more on gaining market share than they do on selling software upgrades.
This is no different than in the XP world. Not all windows API’s are backbported to other versions. If the application uses new features, then it will need to use the newer OS. It’s that simple. There is no reason to backport the API’s. That would require alot of work, or just including the features of the new operating in the old operating system, which would essentially be giving the OS away…
Bear in mind that in the case of the 10.3 to 10.4 upgrade, Apple did backport several new APIs:
Quicktime 7 (well some of it) and QTKit work fine of Panther
Safari 1.3 added a lot of improvements under the hood to bring it up to par with 2.0
There’s a reason why a lot of software asks for 10.3.9 rather than 10.3
I understand his sentiment as I’ve done largely the same thing (but I knew it’d work out like that and I’m fine with it). It’s pretty common for Mac applications to stop supporting the last version within a year (this isn’t so common on other platforms).
I don’t really care though, 10.5 is worth it just for virtual desktops!
Simply put, I don’t buy every version of an OS. Not only does the OS cost money, but it frequently involves paying out more for hardware upgrades or replacement, it may mean dumping cash into replacing software that does not operate under the new release. On top of all of that, it takes time and effort to upgrade the operating system.
So if OS release Z + 1 comes out and there are no few applications coming out for OS release Z, I have a problem. I run the risk of being less interoperable with the outside world, I may be missing out on features that I want, and I am probably missing out on toys to play with.
Let’s keep in mind that he is advocating dropping support for an older version of an OS soon after the new one is released. It isn’t like supporting Windows NT a decade after its release. Though I admire the Firefox and Opera development teams for developing innovative new releases that run under NT until this day. In many ways, I think that Firefox and Opera even proves his claims wrong because you can develop fast and stable software for old operating systems.
And it’s not as though he is a lone voice in the wilderness. I found that there was virtually no software being released for 10.1 a couple of months after 10.2 was released. A few months after the release of 10.4, I found that there was much less software for 10.3 (though it was nowhere near as bad as the case was for 10.1). In the end I got tired of playing these games and ended dumping Mac OS X for Windows and Linux. Windows NT is nice because I have a stable work system, Windows XP is nice for the games, and Linux is great because I can be bleeding edge and not be bled dry financially.
Thank-you Mr. Mac developer, but no thank-you. Not only did you drive me away from your products, but attitudes like yours drove a long-time Macintosh user away from the Macintosh.
In a way, Windows is only stable because progressive, meaningful releases only occur every several years. Apple seeks to make underlying improvements that make a measurable difference in the performance of the operating system as quickly as they can.
I am no Windows expert of course, but won’t a lot of apps not be able to take advantage of some of the best gains that Vista is making without a new release?
Cry me a river. The fact that OS X has been a compromise since 1997 has frustrated many who have wanted the direction Leopard is taking from the start.
Why would I cry you a river? I just gave up my evangelism for the product and did what any sensible consumer would do: walked away Apple and started using more suitable products.
Oh, and Mac OS X was released in 2001. Most Mac users didn’t make the transition until 10.1 or 10.2 (which was many moons later).
My take on this is that it’s a good thing to keep updating with the latest OS release. Look at some Windows applications like Lotus Notes and ClearCase. These apps are never updated and contain tons of bugs. Hell, both of these apps still sport the Windows 95 Look and Feel and in Lotus Notes the scroll wheel doesn’t even work in all the windows/views. When applications move forward with the OS, they may lose backward compatility but it will force the software companies to not only bring their software up to date but I bet more than a few software companies would take advantage of this opportunity to fix some long standing bugs. ClearCase and Lotus Notes are rearly updated because they know Microsoft will provide the backward compatiblity.
So when microsoft does the same thing with windows people call them out on it, but when apple does it they are praised for moving forward.. sheesh..
Atleast with Windows you only have to pay up every 4-5 years or so.
Now dont get me wrong, i dont like either of them. but still.
Atleast with Windows you only have to pay up every 4-5 years or so.
What ?
Windows 95, Windows 98, Windows Millenium, Windows 2000, Windows XP.
It is just the same as Apple, a new release each 2 years.
You just haven’t paid windows since 4-5 years just because Windows Vista is late.
If it was ready for the initial schedule, you should have paid for it 3 years ago and waiting to pay the “after Vista” release coming this year.
Seems you never used Windows…
Win 95 – Win ME -> nearly 100% compatible in *both* directions… 1995-2000 = 5 years
Win 2k – Win XP -> nearly 100% compatible in *both* directions… 2000-2007 = 7 years
Win 3.1 to Win XP -> fairly upward compatible and that’s an insane long time in tech.
To bad that Vista is such a scary thing… so i guess i’ll either have to pay the apple tax on every tiny Servicepack or completely go linux which might not be an option since it lacks alternatives to some propretery apps i use.
Posting to clear up a myth, not to dispute your point.
With the exception of 10.1 (which was free to 10.0 users), it isn’t accurate to call Mac OS X releases “service packs.” The desireability of an upgrades is, as always, a subjective matter, but XP Service Pack 2 and and the various Mac OS Xs were released for different reasons, and were marketed in different ways. SP2 was primarily a security patch. What UI changes there were mostly existed because of the OS’s enhanced security features. On the other hand, every OS X release has been driven by laundry lists of new features and applications (see http://en.wikipedia.org/wiki/Mac_OS_X#Versions for an example of what I mean). Simply put, for the people to whom those new applications mattered, there was always a definte sense of value in the upgrade.
And really, the same goes to all those people deriding Apple customers for paying $129 for “point releases.” Linux 2.6 was just a point release, but it was a significant one. It’s all semantics, people. Nothing more.
Yeah, i have seen this 100% compatibility. So tell me why 80% of the hundred games I bought working correctly on win 95/98 are not working anymore on windows 2000/XP ?
i’ll either have to pay the apple tax on every tiny Servicepack
Seems you never used Mac OS X …
You might want to read posts that you reply to.
Win 9x -> 2K wasn’t close to 100% compatible when it comes to games and i wasn’t even commenting on 9x -> 2k, i was commenting on 95 – ME and on 2k – XP, which are impressive 5 and 7 years of nearly 100% compatibility. You demand 95 – XP which would be 12 years of 100% compatibility and who delivers that? Can i run Mac OS 9 apps on an intel OS X boxes? No, i can’t.
Thats because Win 9x/ME to 2k was pretty much a completely different operating system and most games that where incompatible where poorly written and made bad assumptions.
It is no problem though, because no one gets in trouble when he cant play an game. People can get in trouble if an needed application doesn’t work anymore (not that anyone would upgrade production mashines without thinking about that) and 9x/Me -> 2k compatibility is much better when it comes to applications.
Until a year ago, i helped maintain an custom written business app that was written in Turbo Pascal for Windows 3.1 and guess what… that TP for Win IDE and the App itself still work on Win XP.
I remember this article ( http://www.unsanity.org/archives/rant/caring_for_developers.php ) also lays out a contributing reason for the Mac developer habit of only supporting the most recent version of the OS — the sheer difficulty of actually testing the apps on older versions. Apple makes it very hard to test on older versions of the OS without having mutltiple computers. You can’t run anything older than say 10.4.4 on Intel hardware, and of course, shortly after each release of the new OS, Apple rigs its new hardware to not run the old OS at all.
http://www.unsanity.org/archives/rant/caring_for_developers.php
When MS does this it’s the end of the world. When they announced DirectX10 would be Vista only, there where threads full of people complaining.
Strange that it is ok if Apple does it.
Yes, and since Apple also has full control of the hardware platform, they also try to force retirement of older computers — only those with USB (post 1999, I think) were allowed to install Panther, unless you use a third party hack. So unless you’re willing to trust a piece of code that can and will be broken whenever a new kernel security update comes out, you’ll need a brand new Mac to keep up with the times.
This is one of the areas where Apple is far worse than Microsoft.
Apple has often added performance updates in the point releases of Mac OS X, while adding substantial technology changes in the various new versions. This doesn’t hide the fact that Mac OS X often feels incomplete.
With Leopard, Time Machine in API form adds substantial help for application developers to build versioning into their applications without building their own software to do it. It’s also extremely convenient for the user to look back and find a previous version of their project. Prior versions of Mac OS X had nothing quite so compelling.
Usually, some operating system bug fix means that an application requires a certain o.s. version. You’ll notice that games require a certain point version of 10.3 because OpenGL was finally working correctly. Audio applications require certain versions because CoreAudio was fixed at some point.
Apple could be forcing developers to move to Leopard, but I believe that it’s quite voluntary because of the Time Machine and Core Animation APIs.
This doesn’t hide the fact that Mac OS X often feels incomplete.
I agree with this. The release cycle of Mac OS X has been interesting, to say the least. People seem to forget that OS X is a relatively new beast, and it has had to cover a lot of ground in the last six years to reach its current level of sophistication. If the OS still feels incomplete, it is at least in part due to the still-maturing API. Over the last six years, Apple have had to improve and expand this API, and have had to make point releases (instead of continuously running updates) both to recoup some R&D costs and to give 3rd party developers specific versions to support. In a very real sense, the cool applications that have come out with every release have just been a carrot-and-stick routine to lead users through an expensive and irritating cycle of upgrades, in order to get to an API for which more 3rd party developers would actually want to develop.
With any luck, that routine will come to a halt after Leopard!
I think some people here aren’t looking at why Apple is doing this. Too many people see a conspiricy or something. Apple need to keep the OS moving forward, and has been pointed out before, give incentives to upgrade. If Apple didn’t, I guarantee more people would be complaining. Vista is catching up, why isn’t Apple doing something. A lot of times, large changes means a lot of work at the Kernel level, I’m guessing that Leopard is no exception.
As for current apps, lets take iWorks and iLife as an example. These will continue to work with Tiger once they become iWorks 07 and iLife 07. This will be the case for a host of other software, such as updated Pro tools etc…
There will be Leopard only apps, but they will be things like Time Machine and other stuff that uses Core Animation and so on.
Do you need to update iLife or other apps reguardless of whether they only running in Leopard or not? Probably not, if you like iLife 06 now you will probably still like it after the new version has been announced, only there will be some new cool features you won’t have. The great thing is, when you do decide to upgrade to iLife 07 or buy Leopard at some point down the track, Apple has improved the experience in many areas.
Vista will be the only OS to support DirectX 10.0. Will that suddenly make my XP installation less desireable? Well, if it does to a point that I must have DirectX 10.0, I’ll go out and buy Vista. Same goes for Leopard.
Apple will continue to support Tiger for a while, but probably in a year or so, a lot of apps that currently run on Tiger/Panther will be Leopard only, which I think is a good thing. That means the software is constantly improving, getting more stable (I’m sure it will be possible to get Objective C 2.0 running on Tiger by the way) and faster.
Be glad that Apple is continuing to make your experience more interesting…
Simple rule of thumb, upgrade when you need to. As for enterprise, I’m guessing 10.4 will suite their needs for some time to come if it currently does.
if you want a mac, you have to pay the price for what mac is about. you need to get a new computer because an ordinary pc isn’t good enough, you have to pay for new software to replace all your windows software and you have to pay to get new OS updates to keep moving forward.
Obviously most people aren’t willing to keep paying all that money but they won’t have the same experience that Apple feels is ideal.
I’ve been using a Mac since the release of System 7.5, and Apple has always done this. Apple’s mantra is:
out with the old, in with the new, if you don’t like it f–k you.
It’s just how Apple is. If you don’t like it you always have Linux or Windows.
Yes sir. And I, for one, am happy to pay for a rapidly improving computing experience.
Pre-existing apps will continue to run in Leopard. If a 3rd party dev’s latest version adds Leopard only API functionality resulting in one needing to go to Leopard then you must determine how valuable this “functionality” is to your work.
If not then ignore upgrading that application.
It has nothing to do with whether or not you upgrade to Leopard.
The upgrade price for tiger is 130$, so after tax and shipping it’s around 150$. This isn’t a service pack, it’s a brand new operating system, with pretty much every program upgraded.
It also has a brand new set of API’s, and it’s 64bit (if that means anything.)
At 150$ for an upgrade, thats about 20cents a day for something you use every day. I’m not seeing the big deal here.