OS Obsolescence: a Rant
by Ctein
Consider yourself warned—I'm on a tear. I'm sick of software and hardware manufacturers forcing us into system upgrades and breaking our existing ones. I'm not talking about needing to upgrade single applications to newer versions. I'm talking out about the them breaking your whole machine. There are some very, very good practical and technical reasons for not upgrading an OS if you don't need to. Lots of professionals are a generation back in their OS's. I'm one of them.
My old PowerBook is running both OS9 and OS X 10.3.9 (the newest OS it can). Why did I bother even putting OSX on this pre-Y2K platform? Because I'm probably never going to be able to get rid of this computer, and there are some situations where I need OS X (like drivers for new peripherals). I've got "mission critical" applications that only run under OS 9, not diddly stuff but professional business applications. Up until a year ago, Apple built OS 9 backward compatibility into OS X. They killed it with the MacIntels. Which means even when I get my new Mac I'm going to be stuck supporting this one. Doesn't matter to Apple that 20% or so of Mac users are in my boat. They just said, "The heck with you all; your new computers are never, ever going to be able to run those apps."
Now Adobe announces Photoshop CS3 (hardly a major upgrade) and guess what? It'll only run under the latest Mac or Windows OS (that's 10.4 for you Mac addicts and XP for PCers). So my laptop won't support it and my PC desktop is stable at Win2K, thank you very much. I'm screwed again.
None of this was technically necessary. MacIntel OSX could support OS 9 apps. And the core code and functionality of OS 10.4 vs 10.3 and Windows 2K vs XP is similar enough that Photoshop compatibility for one generation back would be no big deal.
I understand how corporate profits drive this. Intentionally killing off acknowledgment of an older system, even if many people still use it, means that much less money spent to keep those customers up and running. Well, frankly, my dear, I don't give a damn. When I define my business and my pleasure by the enhanced profit margins of these companies, then I will care about them making an extra buck by inconveniencing me. Until that unlikely day, it's nothing but a posterior pain.
Recommended in the Comments: Dynamics of Software Development by Jim McCarthy.
Consider yourself warned—I'm on a tear. I'm sick of software and hardware manufacturers forcing us into system upgrades and breaking our existing ones. I'm not talking about needing to upgrade single applications to newer versions. I'm talking out about the them breaking your whole machine. There are some very, very good practical and technical reasons for not upgrading an OS if you don't need to. Lots of professionals are a generation back in their OS's. I'm one of them.
My old PowerBook is running both OS9 and OS X 10.3.9 (the newest OS it can). Why did I bother even putting OSX on this pre-Y2K platform? Because I'm probably never going to be able to get rid of this computer, and there are some situations where I need OS X (like drivers for new peripherals). I've got "mission critical" applications that only run under OS 9, not diddly stuff but professional business applications. Up until a year ago, Apple built OS 9 backward compatibility into OS X. They killed it with the MacIntels. Which means even when I get my new Mac I'm going to be stuck supporting this one. Doesn't matter to Apple that 20% or so of Mac users are in my boat. They just said, "The heck with you all; your new computers are never, ever going to be able to run those apps."
Now Adobe announces Photoshop CS3 (hardly a major upgrade) and guess what? It'll only run under the latest Mac or Windows OS (that's 10.4 for you Mac addicts and XP for PCers). So my laptop won't support it and my PC desktop is stable at Win2K, thank you very much. I'm screwed again.
None of this was technically necessary. MacIntel OSX could support OS 9 apps. And the core code and functionality of OS 10.4 vs 10.3 and Windows 2K vs XP is similar enough that Photoshop compatibility for one generation back would be no big deal.
I understand how corporate profits drive this. Intentionally killing off acknowledgment of an older system, even if many people still use it, means that much less money spent to keep those customers up and running. Well, frankly, my dear, I don't give a damn. When I define my business and my pleasure by the enhanced profit margins of these companies, then I will care about them making an extra buck by inconveniencing me. Until that unlikely day, it's nothing but a posterior pain.
Posted by: CTEIN
Recommended in the Comments: Dynamics of Software Development by Jim McCarthy.
29 Comments:
the digital domain - computers, operating systems, software, RAW formats, et al - are an ever increasing-in-complexity Tower of Babel. Consequently, I have adopted the word "proprietary" as my most favored epithet.
The phrase "indentured servitude" also comes to mind.
As Mike knows I'm a Windows man (don't like the Mac look - either the user interface or the hardware) and because my day job involves sorting out the mess that Joe Public makes of his PC I'm just dreading Vista.
It would seem that software makers are often solving problems we didn't know we had or offering us new functionality we didn't ask for.
This is not a new problem but surprisingly things may be better now than they were some years ago! I have a wonderful heavy book about the design of audio amplifiers which was written in the 90's before the PC became a ubiquitous household item - the author bemoans the fact that during production of the book he had to move between 5 different computers! At least these days most people can do all they want on one machine and those with "legacy" applications can get away with just 2 machines (or even a dual-boot single box).
Computers or cameras, we live in a world of built in obsolescence but at least you can still make photographs with a 60 year old camera.
Cheers, Robin
p.s. Ctein's new book arrived a couple of days ago - what a refreshing change to see all the screenshots done on a Windows machine!
My Mac iBook G4 will work both ways;
however can agree with Ctein.
I still miss the simplicity of OS
8.6 and yes too OS 9.2.
All this mad rush to electronic oblivion will doom us all. You WILL
be buried under electronic garbage!
Stupid is as stupid does.
As for me, I'm very happy with Mac OS 10.4. I think it's a wonder. I'm not sure it would have been worth the upgrade from Panther, but I bought this dual G5 just as Tiger came out. By 10.2, OSX was clearly leaving the old system in the dust, as far as being able to get serious work done. 10.3 was better, and 10.4 better still.
When I end up booting in OS 9 I wonder how I used to work in it all the time. I'm very happy to have a stable, usable OS; I'm glad to pay for the continued evolution of it. Better and better, I'm all for it. I spend all day with it, and I'm glad it's working so well.
Ctein, if you don't think CS3 is a worthwhile upgrade, you can keep working in CS2 just fine. CS2 is a wonder itself. If you get a newer Mac, you'll have a newer OS. Nobody's really making you upgrade. From what I hear of CS3, there are some nice usability enhancements to the RAW interface, and I'm all for that. I'll pay my annual Adobe tax for that. And if it runs on my next intel laptop, to be purchased soon, blazingly fast and stable, I'll be a happy user.
I understand Ctein's frustration, but I don't agree with him that Apple ought to be supporting OS 9 forever.
Supporting very old OS environments is what Microsoft does. They are only now with Vista dropping support for some very old DOS and Windows 3.x apps. They are widely criticized for late releases and crashes, while Apple is praised for up-to-date releases, full of fancy features, and for stability of their OSes.
If Apple were as committed to backwards compatibility as Microsoft is, they would have a similar set of difficult engineering challenges that would take resources away from other work.
There's another side to this, also: Until a few months ago, I was unable to buy any Mac games for my kids that ran under OS X. Only now, years after OS X first appeared, are they starting to become available. Clearly, they saw no need to upgrade until their hand was forced.
In the interest of full disclosure, I am the developer of ImageIngester, which does not run on 10.3.x. I could have made it run, but that would have complicated development and testing, and would have required me to keep a 10.3 system around and then to support users still running 10.3. My choice was to put all my efforts into development on the current platform, and to forego selling to customers running only 10.3. This was not a technical decision, since I could have supported 10.3, but a decision based on how I choose to spend my time.
Supporting a long comet's tail of old environments is a valid way to do an OS, and so is a sliding window. Microsoft does it one way, which its customers appreciate, and Apple does it the other way. There's no way Apple could both support old environments and also give us the other advantages we all love.
--Marc
On the other side of the coin, look at what happened to MS trying to stick by its legacy support - 6 years developing Vista. And it is still compromised because of it.
Apple avoids such issues by just resetting. (well, somewhat)
Why doesn't CS3 run on W2K?
Most people who want old legacy software to run on the latest OS or hardware also decry the increasing complexity of new operating systems, and complain bitterly about code bloat and increasing costs.
You can't have it both ways.
For instance, it's easy to support OS9 on G4 and G5 chips as these chips support a PPC emulation mode for the older 601, 603, and 604 PPC chipsets.
The Intel chipset doesn't have this emulation mode; in fact they would need to support a Motorola/IBM PPC instruction set, which is simply not going to happen. It's true that Apple could probably do it in software emulation, but that would require engineering resources that would be better spent on latest OS. It's just doesn't make any economic sense to deploy engineers to write a complicated emulation system that only a few people would ever use.
At some point it becomes economically unfeasable to test the latest OS with 6 or 7 year old legacy apps. If a problem shows up in testing, who is going to fix it? The software developer? They've long ago retired the code base and redeployed their engineers elsewhere. The OS developer? Ditto. Who will pay for the considerable engineering and QC resources to fix problems like this?
Computers are not like camera systems where you can jam a 25 year old lens onto a new body and get some functionality; they are orders of magnitude more complicated.
If you want a better idea of what goes into software developement, check out the book "The Dynamics of Software Development" by James McCarthy.
Actually, there is a ray of hope here. Virtual machines work very well (e.g., VMWare appliances). VMWare is about to come out with a Mac virutal machine (but I don't know about it supporting OS9/powerPC), and Parallels is apparently getting there.
So you can, for example, run Win2000 on XP in a virtual machine, or the other way around, or run Linux on a Windows box, or either on a Mac (or you will soon be able to) And there is almost no performance hit with these newer virtual machines.
So there can be at least some hope of flexibility. Of course, you may still have to pay for a software license (e.g., for Windows) to run in the VM
I personally love some of the new options and abilities the new OSs give me. Having said that, I don't upgrade my computer OS anymore. Too many problems, too many things to go wrong, too risky. I figure I'll replace my computer roughly every 3 years. That's how I replace my OS. Don't have to worry about software compatibility and things just going plain old whacky because I still got the old one while I sort it all out. If you upgrade and it goes bad, you're just up a creek, and I don't like that level of panic, worry and stress.
I'm far from an expert about computers, but I think I agree with John Lehet. In fact if there's anything that strikes me as odd about the current computer Universe, it's the development of using Windows on Mac machines. OSX is Da Bomb as far as I'm concerned--it really MAKES the Mac. The obvious improvement going begging is to make OSX run on PCs, not the other way around.
But playing Devil's advocate to Ctein, what's the alternative? Do we really want them to STOP development of new OS's, or do we really want to be using OS9 still? Or a further evolution of it? Personally if OSX were only usable on PCs tomorrow, I'd switch to PCs. Anything they need or want to do to keep the development of OSX healthy and current is fine with me. I'll take any of the downsides just because the upside is too big to miss out on.
--Mike
It's hardly a situation of doom and gloom. Just like any other young industry in the past there are growing pains. The computer industry is no different and is just now starting to mature. Remember, the GUI has only been available to consumers for 20 years and it has been an even shorter period (much shorter) that computers were not just for geeks and business.
I've read various analysts here and there since the announcement of Vista's release saying we are seeing the last of the major updates to operating systems due to the maturing nature of the industry. I have to agree. Computers have recently moved into areas where people are less concerned about having the latest tech and are more interested in just using their computer (this is part of what defines the maturity of the industry).
Microsoft is in a position where the stability of the platform has improved to the point it is hard for them to convince people to upgrade and due to their position and size in the market, they lack the ability to keep upgrading the OS without loosing more market share to Linux in the more profitable markets (business). However well Vista works and how adaptive it is to new things will largely determine the long-term success of the Windows platform.
Apple has slowed down the rate at which they release major updates too. What was every year is now two years. With the system very well optimized as it is and some new core functions in the upcoming Leopard (namely resolution independent UI) there will be little need to upgrade as the OS will be able to handle all technology that is currently on the horizon. A release from time to time that adds new functionality or refinement for current tech to the core that already exists is all that may be needed in the near future. Heck, like Unix/Linux systems, it can be adapted to run on any kind of processor out there. PPC, x86, and whatever the Panasonic chip the iPhone has... and apple has proven it is more than capable of making such a switch almost seamlessly as needed. Apple's software exists simply to support their profitable hardware and offer a compelling user experience. They don't profit from OS upgrades in any significant way. If they did, then they would have anti-copying provisions like Windows.
I don't think it will be long before OSs never get major upgrades again. We may see it in the next 5 years starting with MS. What would be the point and how could they sell it? It does not matter how hard a company may try to make an OS obsolete, if enough consumers say "to hell with that," then there's not a thing the company can do about it.
Until then, if you still want to use software from previous OS iterations, there are always plentiful choices in OS emulators freely available a Google search away.
Ctein highlights a problem small businesses never had to face before microcomputers. When creating mission-critical or production systems, probably they biggest factor that should be considered is 'how long will the software/hardware be supported and viable'? I manage desktop support for a large asset management company, and was aghast to learn a few years ago one of our critical systems depended on a software development platform that had been abandonded by the manufacturer (Sybase). We didn't avail ourselves of the end of life notification and ended up needing to support an orphaned system. We should have replaced the Sybase dependencies on our schedule instead of an emergency case.
Frankly, neither Windows nor Mac OS X are viable platforms in my opinion for production/critical systems as both manufacturer's interests are in obsolescence. A platform such as Unix is what should be used if you need indefinite ability to run the same software. And in the future, you can more likely run an old version of a Unix on new hardware.
Patrick
I'll try to helpful first, before ranting back. :)
If you just have a few mission-critical apps, why not try running them under SheepShaver? It will let you run OS 9 anywhere you want. See http://theappleblog.com/2006/07/01/classic-on-intel-macs-courtesy-of-sheepshaver/ and http://gwenole.beauchesne.info/projects/sheepshaver/ .
Alternatively, you can try installing OS X 10.4 on the Powerbook using XPostFacto, from http://eshop.macsales.com/OSXCenter/XPostFacto/Framework.cfm?page=XPostFacto.html This program allows you to install OS X on unsupported hardware. I have heard that after disabling things like Spotlight and Dashboard in 10.4, it'll run even faster than 10.3.
A final option is to get a refurbished or second-hand PowerPC Mac so you can run your Classic apps. Those Quad G5s are still killer machines, and the Powerbooks aren't too bad either.
And now for the rant.
As a software developer, with regards to running OS 9 "Classic" apps on OS X Intel, I can see how it just wasn't viable. It would be a major undertaking to still run all those old apps, and it would take time away from new development. While this loss isn't "technically necessary", it's bordering on technically unviable.
It's a bit disingenuous to claim that manufacturers are "breaking your whole machine". No one is breaking anything. Disappointment with not being able to run CS3 is okay, but to blame the computer industry because an old computer can't run the so-new-it's-not-released version of one of the most demanding applications out there is a bit silly. Your old software (including CS2) works just fine, doesn't it?
While CS3 is a minor upgrade in terms of features, it is a major upgrade in terms of underlying core code. The "core code and functionality" between OS X 10.3 and 10.4 is very different in terms of supporting Intel machines.
The loss of support for 10.3 is the price paid for having an Intel native Photoshop, which is a major upgrade for anyone owning a newer machine. If it is only a minor upgrade to you, then why bother upgrading?
But not only is it just corporate profits driving this, but the relentless advance of technology and (gasp) user requests. Sometimes we suffer when we can't keep up, but on the whole the industry moves forward. Otherwise we wouldn't have the amazing technology we do. Have you had the opportunity to try apps like Aperture or Lightroom? On a quad processor machine? Wow!
Now, tell me why I can't use Velvia in my new DSLR? Those annoying camera manufacturers, breaking my hardware...
All I can suggest is asking yourself the question I ask whenever I'm tempted to make a big purchase: "what do I need the new one to do that my old one can't?"
If nothing, don't buy it.
Also, I've done the Photoshop upgrade every time since I started owning it at version 2. There may have been one time around the 4,5,or 6 transition -- I can't remember which -- when I wasn't particularly thrilled at the new version. But basically, every time I consider it money well spent. It's amazing, and we're lucky to have it. We're lucky they keep making it better. Someone's got to pay them to do that work.
Oh, and I'm a fine art photographer and part time web designer. I'm do *not* have money to burn.
patricperez says, "A platform such as Unix is what should be used if you need indefinite ability to run the same software."
Mac OS X is Unix. It runs a Mach kernel that is linked to BSD.
As a software developer in a previous life, there comes a point where supporting old operating systems and hardware becomes untenable. The code gets unmanageable with workarounds for OS bugs that never got fixed. It's also an added burden testing new versions on old platforms. It all boils down to where the main market for your product is at. For the Photoshop CS3 beta, that's 10.4 on a Intel Mac. The reason the beta is public is to allow third-party developers to update their plug-ins for Intel. That it also runs on my PPC Mac is a bonus.
My upgrade strategy nowadays is to wait until the last minute and update everything. Not that anybody has a choice but there's little reason to buy an Intel Mac today (most of the major apps aren't Intel native) but when they are I'll be up for a new machine with the latest OS and quite a few software upgrades. The upgrade can take a week out of your life though. Incremental upgrades are a bit of a trap as you always find that you need to upgrade something else. Case in point: Photoshop CS3 will run on older machines but you'll find that its Smart Objects/Filters will be a tad slow.
I'm still after the elusive computer that is as fast as the first one I owned.
John Lehet and Josh both suggested continuing to use CS2 if you don't need anything that's in CS3.
One problem with this is that Adobe stops supporting the older versions. If you buy a new model camera, you won't be able to open its Raw file in PS unless you upgrade to CS3 so that you can use the latest version of Adobe Camera Raw.
Adobe Camera Raw is the hooked barb on the harpoon. To use ACR with a new model of camera, you need a new version of ACR. To run that new version of ACR, you need the latest version of PhotoShop. To run the latest version of PhotoShop, you need the latest version of your operating system.
Unfortunately, while there is no strong technical need for this, there is a very strong financial motivation to not support older platforms. Every single platform that is being targeted requires a fixed amount of additional development, testing, and that means more staff, longer project cycles and so on.
If Adobe sold CS3 for all OS's since Mac OS 9 and Windows 98, but charged double the price (regardless of the OS), would they make as many sales? That's how the software makers see the situation. It's unfortunate that there will be some victims along the way, but that's the price you pay when you use closed proprietary systems.
> To use ACR with a new model of camera, you need a new version of ACR.
Probably the DNG converter has the same hook? Ouch.
Personally, I think Mac does a fine job of keeping everyone compatible. The Rossetta thing they deployed in their transition to MacIntels was ingenious, even better than the great job they did stepping up from 9.2 to 10.1.
Oh, and I'll never throw away my Powerbook 140 which is happily running MacOS 7....
And I'll keep my TiBook. I too am bound to The Old Ways, for now. I have to find a way to move HyperCard forward. Believe it or not, all 100+ individual-view pages on www.lehet.com are built from a HyperCard stack, and so are some of my other sites. I have to move that forward though; it's ridiculous. But I've never found anything as quick and easy.
CTein -- your rant about the "core code and functionality" being the same between OSX 10.3 and 10.4 is ill informed.
There were numerous "under the hood" optimizations for client software in 10.4, including CoreImage, a framework which harnesses the otherwise unusable power of your graphics card's GPU to process video.
This means filtering and other operations can be completed much faster and in the background, so you can work on other stuff. This is a nice benefit of Aperture (and has the side effect of providing RAW read support for any OSX application built on CoreImage). I'm sure if Lightroom offers this capability, it will need to have CoreImage available, too.
As for 2k/XP, they changed the DirectX video pipeline between 2k and XP and I'm guessing it offers similar capabilities to CoreImage.
Consider: why are you willing to upgrade client software and not the complex frameworks it runs on top of? This would be like putting shims under all the furniture in a sinking house rather than replacing the foundation. Secondly, if there's nothing wrong with your current image processing pipeline, what do you gain by investing in the latest Photoshop?
One or two folks mentioned Unix/Linux and their ilk.
As a long-time Linux user (since kernel 0.99pl14), I found that I finally had to give up my beloved development platform, when I found I wanted to run Photoshop to support my newly discovered digital camera habit. Unfortunately, in spite of claims made by the WINE folks, nothing more recent than Photoshop 7.0 will run reliably on Linux under emulation.
Also, X, the base graphical system underlying the various desktops on Unix/Linux systems, has no real support for monitor calibration or color spaces. That turns out to be a terrible crippling factor for anyone that wants to run a color-managed workflow on a Unix/Linux platform.
So, what's so great about Unix/Linux? Consider that programs that are over 30 years old still run just peachy, alongside the latest and greatest, like Firefox/Thunderbird, and OpenOffice. The open source aspects of these systems are their salvations, when it comes to legacy support. As long as a distribution maker has access to the source code for all programs to be included, they can be tweaked and re-compiled to match whatever changes/fixes might be in the latest OS kernel and interfaces.
In a proprietary world, like Windows, with all the various third-party for-pay applications, this just isn't simply doable.
Most folks have never experienced installing a Windows system, for instance, from scratch. Most everyone these days gets a PC, bundled with Windows, and a bunch of 'shovel-ware'. They may install a couple of killer apps, like Photoshop, etc., but beyond that, have never had to deal with the intracacies of installing a modern OS on a bare-bones system.
As a software developer in a previous life, I have had more experience than I care, loading and re-loading various OSes on PCs, and I can tell you that bringing up a PC with Windows, and a full suite of development applications, and the obnoxious 'install-reboot-repeat' process of service packs and patches, is enough to make you scream. It used to take me several days of this routine to get a Windows PC up and ready to do development on.
Constrast that with installing a Linux distribution from CD/DVD. You essentially select the 'Everything' selection and just let it rip. A couple of hours later, and usually a single reboot, and you have a completely configured system with over 4,000 utility, office, web, and development programs installed.
Now, if companies like Adobe and Intuit (the makes of Quicken and Turbotax, my other two 'killer' apps) would take the Linux platform seriously, I'd dump Windows and its ilk like a hot potato, and go back to my first love.
I will echo what Dave New says. I'm a long-time Linux user who's gone to the dark side in recent years because of PhotoShop (and the occasional game). I still use Linux at work. (I've been intrigued by Mac OS X, and if I have to buy my own laptop I may buy a Mac. We'll see, though.)
My personal hope is that virtualization will help alleviate some of these problems. There is no reason why you can't run a newer operating system in an emulator. Performance may well be an issue, but you can't have everything. I've gotten decent performance with VMWare virtual machines for some things, albeit not all.
PhotoShop will always be something of a problem, since it's so resource-intensive. Nonetheless, I get by on a laptop with an apparently broken video card (it locks up when I load the proper drivers) and less RAM than many people seem to use.
By the way, my absolute worst OS installations were all Windows (95 and NT 4, I do believe). My worst Linux installation (Slackware with about 50 floppies when I didn't know what I was doing) was not nearly as bad.
When I see such rants, I always like to give a big, Nelsonian, "Ha-ha!" to everyone who has been convinced to migrate in droves to digital.
Let's face it, none of this transition has ever been made with sustainability in mind, only short-term profits, recurring upgrades, and a messianic vision of the future.
Those who suffer the most from this balderdash are actualy those to whom digital is the most useful, i.e. professionals with a high-volume output, newspapers, magazines, etc. Because these people actually have to rely on their material instead of just using it for fun, they get bitten harder. So when will we finally see digital equipment that will last as long as traditional equipment? At some point professional should lobby for open standards and sustainable equipment otherwise we'll never leave this swamp.
A look at the issue from the developer's perspective.
As dasmb pointed out, there are very valid technical reasons to require an OS upgrade with the application upgrade.
While there is validity to some of what Ctein says, I think he's clearly wrong when he says that Adobe and Apple are "breaking his whole machine".
His old system is CS2 on 10.3. Nobody's breaking that. CS2 doesn't have a drop-dead date in it after which it won't work any more. Ctein can keep working the way he's accustomed to.
If he wants CS3, he'll need to upgrade. And it's his right to be unhappy about that. But no one is forcing him to move to CS3.
>So when will we finally see digital equipment that will >last as long as traditional equipment?
When it stops improving?