Implications of Adobe's Intel-Only Soundbooth
When I read Adobe's new Soundbooth app would only run on Intel-based Macs, I wondered the same thing as everyone else: why? John Nack, Photoshop's product manager, goes through the reasoning on his blog, but it presents more questions to me than answers.First, kudos to John for having the guts to answer something like this in public. No matter what I say below, I have more respect for someone who is willing to explain their stance.
(I also want to be clear that I'm discussing the merit of philsophy and technique, not the merit of the people themselves, and I don't take any of this too seriously. I'd buy any of them a beer and talk it over, and I'm not going to lose any sleep over it anyway.)
Where is Adobe's Focus?
John Nack speaks to the core reason for the decision about Soundbooth:
Apple's migration to Intel chips means that it's easier to develop for both Mac and Windows, because instead of splitting development resources optimizing for two different chip architectures, you can focus on just one.
There are two sides to this. On one hand, yes, focusing on one chip makes things easier, but the point of Apple's developer push was not switching to Intel, it was cleaning out assumptions about the underlying chip. That's why the key word they use is Universal.
Apple's focus is on Intel right now, but this stuff is unpredictable. Well, actually, it is predictable. It always changes. There was 16-bit to 32-bit, 32-bit to 64-bit, 68k to PowerPC, PowerPC to Intel. What if Apple released a new type of portable which was not x86-based at all?
At each step along the way, the speedbumps were related to cleaning up code that was too closely married to the processor. This begs the question: why is Adobe intent on writing so much processor-specific code that a PowerPC version of a new app becomes impractical? Is this a framework issue? A tools issue?
I'm struck by John's comments that the end of PowerPC-based Macs validates the idea that new apps can assume Intel. This isn't what "newness" should buy you. It should be an opportunity to not repeat the mistakes of earlier software designs.
In simpler terms, the goal shouldn't be to write an app that runs on both PowerPC and Intel, it should be to write an app that is not built around a single processor design.
Adobe may have thoughts on why they think this is impossible, but I haven't seen it in written form. My question is how much they're betting on Intel being the only game in town.
The Subtext
The other half of my thoughts on this relates to the subtext of Adobe's decision. Two Adobe program managers have now gone on record as saying it would be too hard to support two processors on the Mac side.
This would make more sense to me if it was a small, independent developer, but this is Adobe. I think they see themselves as an industry heavyweight. If that's the case, why can they not afford the resources for something like this?
It also sends the wrong message to users and developers, as David Chartier of The Unofficial Mac Weblog illustrates:
I am certainly no software engineer, and I know equally little about the intricacies of audio software, but if a company with Adobe's girth says that now is a bad time to start building PowerPC support into a brand new product - I'll listen.
The assumption on David's part (and it's not his fault), is that the problem must be too difficult if even Adobe can't do it. That assumption could be wrong. The challenge may be more due to Adobe's technique than the issue itself.
Optimizing low-latency software is not easy, but it is possible, and Apple provides at least some of the groundwork in Core Audio. Soundbooth actually requires Core Audio support, so they're using it somewhere. Core Audio, of course, supports both PowerPC and Intel-based Macs.
The Goodwill Issue
The one sentiment that I find particularly strange is the one that John brings up here:
If you're a Mac user, I think it's important to ask yourself, "Would I rather encourage software developers to bring their titles to the Mac, or would I rather jump down their throats given any opportunity?
If Adobe were to bring other Windows-only apps to the Mac, would I be happy about that, or would I rather give them hell for focusing on features & functionality rather than a discontinued chip architecture?"
[...] You're making it harder (and less appealing) for people of goodwill to make the effort to support the Mac.
Yikes. So now John has suggested that the Mac is a charity case. Do Adobe's numbers back this up? I doubt it. In any case, what kind of relationship would Adobe have with its customers if critical feedback resulted in withdrawing products?
Now this is John's personal blog. What he says may reflect company culture, but it may not. If it does, it's more telling about Adobe than the Mac. In the past, Adobe was the final word on Mac applications, but that's been changing. Apple's pro apps make some of Adobe's look positively dated.
In any case, Adobe went from being a champion of Mac design and technology to being among the last to adopt any Mac standards. There is also the issue of keeping pace with evolving Mac user experience. Mapping between the Leopard and Vista user experience will be likely be more than just drawing widgets differently, and this likely means the Mac version of Adobe apps will draw the short straw.
For Adobe, watering down their Mac apps to remain even with Windows is a high-stakes gamble. They must either believe their customers are Adobe users who happen to use a Mac, or they believe that they can outclass Apple in application design on an ongoing basis, negating the loyalty issue. They could also simply be betting there are enough professional media users on Windows to compensate for lost sales to Apple.
In my opinion, none are sure bets by a long shot, and the fence-sitting is chipping away at their position. Meanwhile, Apple continues to move forward with their OS, API and applications. It's completely within Adobe's power to do something about this — Apple wants them to use these APIs — but it would mean changing their philosophy about cross-platform apps.
In any case, I think Apple should feel emboldened now to press ahead with their own pro apps. They don't want the Mac to be held captive to Adobe's whim, and neither do I.
Implications of Adobe's Intel-Only Soundbooth
Posted Oct 30, 2006 — 29 comments below
Posted Oct 30, 2006 — 29 comments below
Dale — Oct 30, 06 2223
1) They're porting a Windows app to the Mac, and the framework for the app is Intel specific. So they're just doing it the easy way; and
2) They're a slow slumbering company that are threatened by their competitor in this category, Apple. They know they're behind the eight ball releasing the product (just like PhotoLightInABoxOrShop, or whatever they're calling it) and know their product development lifecycle has more in common with Vista's.
ssp — Oct 30, 06 2224
I assume that such a media intensive project may require some low-level work, possibly with that processor specific SIMD stuff. I can very well imagine that there aren't all that many people who are good at doing that kind of programming, that they're quite expensive and that Adobe may prefer to use them for their other applications.
Sure, it's not nice to see the PowerPC platform being left by companies. But I fully assume that it makes 'business sense' for the people who do it. Let's say it takes another year until their Soundbooth is ready, how many of the people who do sound stuff at a level where GarageBand doesn't cut it won't have Intel Macs by then?
Just from a superficial look... why aren't people more concerned about pretty much every new application coming with its own (ugly) UI these days?
Dan Price — Oct 30, 06 2226
Adobe have had a monoploy with their sofware for far too long. They're showing all the signs of a monolpolistic company that has forgotten how to compete and now sees it's brand name, not it's product, as it's most important asset.
They're also not the first dual-platform company to make compromises on the Mac for the benefit of the larger Windows platform.
Apple is right to challenge them, and any other company that thinks it can hold the platform hostage. I don't see Adobe dieing anytime soon - Photoshop is still the killer app - but strong competition is in everyone's best interest.
As for only developing for Intel, have they learned nothing?!
Frank 'viperteq' Young — Oct 30, 06 2227
I think that is a dangerous assumtion to make. Many Mac users that I know (myself included) can't go out and buy a brand new Intel-based Mac simply because the money isn't there. For many of us, buying used has always been the way to go and until some of those Intel-based machines start hitting the used Mac channels, PowerPC is what we use as it still works and functions great.
I think I read somewhere that only about 30% of total Mac users have actually made the jump to the new machines. If that number is correct, in a years time that number could've increased to 60% or 70% barring no more problems hardware wise. But that STILL leaves a good 40% or so that are still using PowerPC and to assume that these users out there should not have the opportunity to use the software is a bad thing. I agree with Scott: This does send a really bad message to users and other developers for the Mac.
I hate to say it, but as large companies go, Microsoft provides the best example of how to develop software for two different platforms as a large company. Yes, their Mac offerings always ship 6 to 8 months after their Windows-centric counterparts, but at least Microsoft can say that they have a dedicated team of developers that insures that their Mac-centric products fit into what users feel a Mac-centric application should be.
Adobe however doesn't see things this way. Their developers seek to write code one time that works everywhere and that's not always the best course of action to take. As of late, while their applications work, they often seem bloated and out of place in context to the other applications that are finding space in my Applications folder as of late. I say that if Adobe no longer cares about properly making software that fits the Mac like it's supposed to, get out of the way and let some other promising company step in.......
Roger Dragon — Oct 30, 06 2228
ssp — Oct 30, 06 2229
At the end of the day it's Adobe's decision anyway. If they don't see a market on PPC Macs, well then they won't make the effort to program for it. That's just how things work...
Dan Price — Oct 30, 06 2230
So why don't we as developers save ourselves a lot of heartache and switch to Windows? :P
The most successful pioneers saw the potential of their product and created the market. Apple and DTP for example.
Apple was so successful with it's ppc-x86 transition because they engineered their APIs to be abstract from the hardware from the very beginning with OSX. Adobe are struggling to release UB-versions of their products because their codebases are tied too closely to the old Carbon APIs and the PPC. By tying their future products to a particular CPU, even Intel, they risk the same hardship in future should Apple migrate again (it could happen).
I find this attitude incredible coming from a mature company like Adobe. Intel-only products are inevitable, just like the first PPC-only products. But it's far too soon. The last range of G5 Macs were not snails by any means. Why limit their product's appeal to what is a small 1st-gen Intel userbase, when the bulk of their customers are still on PPC precisely because they haven't ported their main product suite yet?
Preston — Oct 30, 06 2231
Preston — Oct 30, 06 2232
Why limit their product's appeal to what is a small 1st-gen Intel userbase, when the bulk of their customers are still on PPC precisely because they haven't ported their main product suite yet?
The suite was actually successfully ported to Xcode earlier this year, and what they're doing now is finishing up CS3.
Ben — Oct 30, 06 2233
Machine-level optimizations are hard, and most certainly do not come for free. Developing PPC -specific code from scratch buys Adobe no value in the (much) larger Windows market, but x86 code is a win all-round. Within 2 years, any Mac that matters will have an Intel chip inside it, and by next summer, most serious creative professionals will already have one. The worst case for a user is that if you have a real business critical need for an Intel-only app, it will just force your hardware upgrade cycle to happen a little sooner. This is not about abandoning the Mac platform; this is about Adobe resource tradeoffs. They're not abandoning Macs, they're abandoning *old* Macs, because it would be significantly more costly to support them. If Soundbooth were written entirely in generic C++, and they could get a PPC build out of it for free (or very cheaply), they surely would do so. As he says, the CS apps already have PPC stuff that just needs tweaking/updating-- the PPC wizards at Adobe are surely working on the CS apps, where the mileage is much better.
To Scott's point about making apps architecture independent, rather than moving to a new, specific architecture: this might be what we all learned in school about design elegance, but it's really only an academic argument here. Apple won't incur the cost of another complete architecture change for at least another decade, if ever again. In the long run, we're all dead (users, computers, products, and software companies). There's no use doing any more planning than is reasonably necessary for future events that are unforseeable and unknowable.
Samo — Oct 30, 06 2235
They have competition for their various technologies coming from Apple _and_ from Microsoft. It doesn't look like they want to use innovation to battle these "threats", though. They bought Macromedia and are trying to (ab)use Flash to strengthen their position.
With the current crop of APIs I don't think that developing a Photoshop competitor, for example, that does 80% of what PS is used for and does it better is _that_ hard anymore. Neither on Windows (Vista) nor on OS X. Would be fun if someone did to Photoshop what Adobe did to Quark XPress.
Ben — Oct 30, 06 2236
Large software companies (Adobe, Microsoft, Apple) do their thinking about product roadmaps in terms of years, not months, and brand-new products don't look at the way the user base is *today*, they look at it over the next decade.
Here's a followup example to illustrate. When Apple shipped iTunes for Windows in late 2003, a huge chunk of the consumer Windows users were still running Win98 (or even Win95). But iTunes required XP. Why? Because Win9x compatibility was just legacy compatibility at that point, and it was understood that within a few years, any Windows box that mattered would be running XP (or some NT-based successor). iTunes and QuickTime do stuff that requires driver-level behavior, and all that media and I/O stuff is totally different on XP than it was in Win9x. It wasn't worth it to Apple to write a bunch of code up front for a platform that was already on its way out. It was smarter to invest in the future.
Frankly, this was a riskier decision for Apple than what Adobe is doing. Apple was looking to push iPods into Windows households, and they alienated a bunch of the immediate market (e.g. here) when they made this call (obviously it hasn't been a big problem...). Whereas Adobe's customers in this case are professionals, who will upgrade hardware more frequently and readily anyways.
Samo — Oct 30, 06 2237
Let's hope, then, that CS3 will actually be a good improvement over CS2, worthy of a full product version bump. For a change.
Keith D — Oct 30, 06 2238
I think this is an incredibly daft decision on Adobe's part, the PPC products will still be around for a while and people will want their software to run reguardless of the intricacies of the platform they're using.
Scott Stevenson — Oct 30, 06 2240
Of course it is. We're discussing how they reached that decision. Nothing I wrote tries to convince Adobe how many PowerPC users there are.
@Preston: Apple has publicly committed to the Intel platform by releasing Boot Camp and pledging to make the software a part of 10.5
I'm not sure if publicly committed really means much. Today, Apple is on Intel because it makes sense to be. That could change two years from now. Or Apple could add another sort of processor to the lineup without actually phasing out Intel.
Another option is that Apple may add another processor to the existing Intel-based design. Perhaps the GPUs are going to start doing a lot more generalized processing.
so I think that goes beyond simply telling developers to remove their processor assumptions
I disagree.
Adobe supporting only Intel in a new Mac application is news, but I'm not sure if it's as big as it's being made out to be, given when and where
I think you're misreading the point of my post. It's not about PowerPC versus non-PowerPC. It's about shipping apps with a significant amount of processor-specific code. Adobe may have no choice in doing this, but it is risky.
@Ben: Apple won't incur the cost of another complete architecture change for at least another decade
Can I quote you on that? :)
There's no use doing any more planning than is reasonably necessary for future events that are unforseeable and unknowable.
We don't know that processors changes will come, but history suggests they will, and we better hope they do or the industry will have some big problems.
Apple to write a bunch of code up front for a platform that was already on its way out. It was smarter to invest in the future.
I agree with you, but betting on one processor is not (in my opinion) investing in the future. Also, be clear I'm saying they should do this for their benefit.
xyz — Oct 30, 06 2244
Select Hardware, then CPU Type.
Of course these are only statistics for Omni..
Ben Zotto — Oct 30, 06 2247
Can I quote you on that? :)
Absolutely! :) The Apple universe is unusual in that it has survived a couple really severe architecture upsets in its history. This is largely only possible because Apple owns the hardware roadmap for all Mac users. They've done a great job at the current transition, but they don't do it for fun; it's hard and expensive for them and their users. The x86 platform has a lot of gravity because of its use in the PC world, where it will remain in use, sadly, for the forseeable future. That gravity makes x86 attractive to Apple due to economies of scale and better supply chain, etc. I think it's highly unlikely we'll see Apple arbitrarily leave x86 for some other proprietary platform-- the whole reason for being on x86 is to leverage the scale (ie cost, innovation) advantages that it provides from the PC world. (God knows it's a crappy architecture.)
We don't know that processors changes will come, but history suggests they will, and we better hope they do or the industry will have some big problems.
I'm speculating of course, but we're far more likely to see evolutions in the x86 design (64 bit variants, what have you) than wholesale changes in architecture, at least until someone invents something so head-smackingly clever but incompatible that the entire PC industry has no choice but to migrate.
I agree with you, but betting on one processor is not (in my opinion) investing in the future. Also, be clear I'm saying they should do this for their benefit.
Sure. But how do we know they haven't already got an architecture-specific shim layer in there that will make it easy for Soundbooth to work fast on, say, 64-bit Intel as well as 32-bit Intel? The vast majority of an application's code is in the user interface (Cocoa, MFC, etc, etc) and in the data processing underbelly (for example, basically portable C/C++). If there's really architecture specific code, it comprises a tiny % of the code base, and if the day comes that it makes good sense for Adobe to adapt that piece, they will.
I agree with your general thinking Scott about maintainability and future-proofing. But since the issue at hand is PPC compatibility, we're really talking about past-proofing, which isn't the same thing. Not supporting PPC is a business resource decision, not a technical one. (No one but the engineers at Adobe could tell us whether the design of the software would actually allow multiple architectures easily or not.)
Samo — Oct 31, 06 2252
The way Adobe people make it sound, that's not the case. But that might be more of a problem with Photoshop, for example, than newer applications like InDesign.
cjwl — Oct 31, 06 2263
Ok, yea, it would be fun for the PowerPC guys to dink around with the public beta but Adobe knows when the price goes up these guys won't be opening their pocketbooks quickly.
Jesse — Oct 31, 06 2264
Now, do I think that is a good idea? NO! Adobe software is getting worse by my measure, not as bad as MS but bad.
The revolution was from NeXT and now at Apple and it takes real thinking and forward looking attitude to really embrace it. Adobe sees little advantage in changing and I can't imagine they will anytime soon.
They make too much mony to worry about changing the way they create software and have a ton of legacy code and legacy programmers around.
Sad, but true.
Juanxer — Oct 31, 06 2265
Actually, Sound Booth and Lightroom look like a very interesting experiment in app development for a company like Adobe: do some fast "cheap" prototyping and test user reaction, instead of the traditional route these giant software businesses usually follow.
I completely believe Adobe when it says it has to balance resources when determining things like PPC compatibility. It has an ENORMOUS lot of apps to port to Mac Intel (nd nobody would be complaining if Apple had happened to stick to the original Mac Intel launch timetable).
I don't expect Apple to transition to any new processor technology that isn't endorsed by the whole personal computer industry: PPC, ultimately, has been a long and tortuous detour, and going Intel surely was the plan since the Copland crisis onwards. By going mainstream, Apple eludes lots of hardware development and production misery: now the whole PC world is its hardware design and fabrication unit!
John C. Randolph — Nov 01, 06 2272
Adobe has long had the problem of believing that they can do no wrong, and this is largely because the first customer they ever had (Apple Computer), made them extremely rich. Let me tell you, if you think it's bad dealing with them as a Photoshop user, try dealing with them if you want to be a Postscript OEM.
-jcr
scott lewis — Nov 02, 06 2286
Te Apple II line went from 8-bit to 16-bit with the introduction of the IIgs, but that cannot be considered part of the Mac evolution since DOS, ProDOS and GS/OS did not have binary compatibility with the Mac.
David Ramos — Nov 02, 06 2288
Nor do those figures pass the common sense test. I know dozens of Mac owners, most of them design professionals, and few of them upgraded their computers in the last year. Many of the studios I'm familiar with are plodding along on candy-colored hardware and Quicksilver G4s. Few people, even those who depend on computers for a living, have the budget to upgrade just because of a processor switch.
Hardware makers and software companies aren't the only organizations with large budgets and the need for long-term planning. When I worked in higher ed, our institution upgraded lab machines on a two-year cycle, and faculty computers every three years. Administrative users generally received trickle-down machines from labs. Mind you, this was a school with an enormous endowment, and our upgrade schedules seemed luxurious compared to what most other colleges used. In a large organization, machines get upgraded when they're scheduled to be upgraded. Cycles don't shift to accommodate architecture changes; a school can't just scare up a million dollars whenever a manufacturer changes their roadmap.
The consequence of this are that many, many machines in large institutions won't get upgraded just to match industry trends. They match the curve, but only eventually, through slow and stable adjustments that are moderated by budgets and upgrade cycles.
For the most intensive (and expensive) production environments, the computer is of less importance than other necessary equipment. It represents a miniscule capital investment, compared to other items - and it also provides a correspondingly small portion of the facility's capabilities. In audio and video, things like loudspeakers and online storage matter a great deal more than the machine itself. When a microphone costs $2000, upgrading the computer becomes relatively easy, but it also sinks to a low priority.
I suspect that all this matters relatively little, though. Soundbooth isn't a high-end DAW. It's more of a casual audio editor, useful perhaps for preparing sound for use in websites and podcasts or as an adjunct for Adobe's faltering video apps. Adobe's potential audience already comes from the margins of the audio world, and further restricting that audience might not represent a great loss of sales.
Harvard Irving — Nov 04, 06 2296
I disagree with a number of things in this article. First, Apple has publicly committed to the Intel platform by releasing Boot Camp and pledging to make the software a part of 10.5, so I think that goes beyond simply telling developers to remove their processor assumptions.
What does Bootcamp have to do with the way that developers should write their programs? I've tried really hard to think from various other viewpoints, but there is no way I can comprehend what Bootcamp has to do with development or Apple's "processor assumptions."
Second, by the time Soundbooth comes out in the middle of 2007 at the earliest, Intel Macs will be commonplace, and the abandonment of PowerPC will be that much more justified.
Second, by the time Soundbooth comes out in the middle of 2007 at the earliest, Intel Macs will be commonplace, and the abandonment of PowerPC will be that much more justified.
Is there any evidence for this whatsoever? PowerPC Macs are commonplace, but there is no way that Intel Macs will be commonplace as early as 2007.
The suite was actually successfully ported to Xcode earlier this year, and what they're doing now is finishing up CS3.
This makes Preston sound like an Adobe astroturfer. Why should we, the loyal Adobe users, care about CS3? Where the hell is the Intel-native version of Photoshop?
I don't think any serious Photoshop user gives two craps about CS3. They'd probably rather have a version of Photoshop 2.5 that runs on all platforms. So if, as Preston claims, the suite has already been converted to Xcode, then why is Adobe holding back on releasing it?
cjwl writes:
I think this makes complete sense for Adobe, new products are bought by early adopters, early adopters, are, well, early adopters, they buy new stuff, they have the latest Mac, the latest iPod, the latest this and that, these are the people that companies rely on to actually spend money on the product, review it, tell their friends about it, etc. It's just how it works, Sales 101.
Again, I'm not sure what planet this is coming from.
Most of the professional users that I know, are "early adopters" of software, and very evangelistic about good software - but they only update their hardware very slowly. Most professional Photoshop studios and graphic designers are still running old G4 towers. They are spending way too much money on software already too be buying the latest hardware. Even if they have the money for new hardware, the risk is not worth it, because new hardware is not proven to be reliable and solid.
In contrast, the people who seem to have the newest Intel Macs and iPods, hardly buy any software. They mostly use the free software that comes with the Mac, like iTunes and Garageband. They rarely talk about software, and certainly don't evangelize.
Those people who are talking about 30% of Mac users being Intel-based are most certainly exaggerating. It is nowhere near this number. I might believe 5%, but even that is really pushing it.
Among professionals, and particularly Adobe software users, I think the number is closer to 1%. In the consumer market, it may be as high as 2%. And I find 2% to be a staggeringly high number that indicates how quickly Apple is growing. If 30% of Macs in use today were Intel-based, you could elect Steve Jobs as ruler of planet Earth, and he would have solved world hunger by now.
Scott Stevenson — Nov 04, 06 2297
Why do you think that?
Joseph R. Simkins — Nov 04, 06 2298
Business Plan Writers — Jan 05, 10 7073
Excellent post. I want to thank you for this informative read; I really appreciate sharing this great post. Keep up your work.
<a href="http://www.bizplancorner.com">Business Plan Writers</a>
Dior Sunglasses — Jan 20, 10 7132