Hacker News Reader

Comment by chimi | Hacker News Reader

This is why Microsoft is the better long-term play: Backwards compatibility -- their commitment to the investment of labor. If you build your software on Apple, there almost certainly will come a point in the future where your code will cease to work.

Microsoft has made a commitment to backward compatibility. It's such a huge benefit. I have code that has continued to run, untouched, for decades. I don't have to upgrade my development environment. I don't have to buy a new laptop to run the latest xcode. I never have to worry that my years of investment into my own software platform will be worthless three years from now or require unknown amounts of recoding to work.

This is not true of Apple, Ruby, or almost any open source project.

I want to continue moving forward all the time. I loathe the requirement to go back and redo what I'd already done and works. It wastes my time, my money, and my confidence in my vendor.

View on HN

And how much has it cost Microsoft? Apple was able to successfully take the base of MacOS and port it’s operating system to a much lower spec phone back in 2007 and came to dominate the smart phone market (in profit they only thing that matters), the tablet market where MS failed repeatedly, not to mention it was able to port the base OS to everything from Watches and TV set top boxes - all things that Microsoft tried and failed at.

MS being so concerned with backwards compatibility has slowed down its ability to move fast into new markets even compared to Google.

If you are “building your software” for desktop PCs (or Macs) , you’re building on top of market that is becoming less important every year. There is no energy or investments going into desktop computers besides games and even that market is small. The only companies making real money in computer software are Microsoft and Adobe and even they are pivoting more toward browser based software and mobile.

Do you think that Apple should still bundle a PPC emulator with modern Macs? A 68K emulator?

Not being willing to drop legacy code comes with its own costs in maintenance and increases the vulnerability surface. Just the fact that there are over a half dozen ways to represent a string in Windows has led to security vulnerabilities, not to mention all of the other legacy cruft.

It's worth considering the different markets here. Raymond Chen [0] posted this in 2006 - "On the importance of backwards compatibility for large corporations" [1]

> Representatives from the IT department of a major worldwide corporation came to Redmond and took time out of their busy schedule to give a talk on how their operations are set up. I was phenomenally impressed. These people know their stuff. Definitely a world-class operation. One of the tidbits of information they shared with us is some numbers about the programs they have to support. Their operations division is responsible for 9,000 different install scripts for their employees around the world.

> That was not a typo.

> Nine thousand.

> This highlighted for me the fact that backwards compatibility is crucial for adoption in the corporate world. Do the math. Suppose they could install, test and debug ten programs each business day, in my opinion, a very optimistic estimate. Even at that rate, it would take them three years to get through all their scripts. This isn’t a company that bought some software ten years ago and don’t have the source code. They have the source code for all of their scripts. They have people who understand how the scripts work. They are not just on the ball; they are all over the ball. And even then, it would take them three years to go through and check (and possibly fix) each one.

> Oh, did I mention that four hundred of those programs are 16-bit?

16-bit. Ten years after Windows 95 came out.

[0] A great source of backwards compatibility stories.

[1] https://devblogs.microsoft.com/oldnewthing/20061106-01/?p=29...

Well seeing how Microsoft’s profits have increased compare to Apple’s, which was the better idea? Selling to consumers or businesses?

And MS’s recent successes have had nothing to do with Windows. MS’s growth has come from o365 being available everywhere and Azure - where most of the VMs are in fact Linux not Windows. Not to mention that .Net for Windows and all of the related technology (WinForms, WPF, etc) are in maintenance mode while MS concentrates on cross platform .Net Core.

Microsoft is doing the same thing with Windows that SJ said he would do with the Mac “milk it for all it’s worth and move on to the next big thing”. MS is running away from Windows as it’s cash cow quickly.

I’ve always thought Windows was compromised for consumers because of the need to support this sort of enterprise. In some ways it opened up the opportunity for Apple to makes its comeback.

Unless you have shares in Apple your whole point is invalid, you're looking to this from Apple point of view which is very good for her business but not mine, i care about the company that give me the best solutions to my problems not the company that's can milk me the last day of my life.

You've actually made grandparent's point.

What the grandparent said doesn't exclude Apple as a company giving "the best solutions to [a user's] problems". If anything, it can be an enabler.

You say "i care about the company that give me the best solutions to my problems".

Well, hundreds of millions had the same concern on mobile, and they settled on Apple (and others on Android), not Microsoft.

And that was because Apple could (partly thanks to its lax approach on backwards compatibility), be able to shrink down OS X as needed for mobile (and now wearables), make the transition to x86 and later ARM for mobile (and in the future ARM for desktop perhaps), and so on.

In general, they can move forward faster, break things if needed, etc.

Then there are other concerns. I don't want my OS to be burdened with ancient code (and vulnerabilities etc). I don't want developers to stick to ancient versions and not integrate new features etc (Apples forces their hand to -- those that can't be bothered are eliminated). And I don't want a mix of new, old, medium-old UIs on my Mac, like I get on Windows (depending on which decade of native UI framework, widgets, etc they use).

Not having to worry about yet another security vulnerability caused by 10 year old code does impact me. It’s also nice to have other Apple products that MS could never produce because they worshipped at the alter of backwards compatibility.

Running outdated code on a bloated operating system doesn’t benefit me.

>Not having to worry about yet another security vulnerability caused by 10 year old code does impact me.

Surely you're not suggesting that OS X is somehow more secure because it doesn't provide as much backwards compatability?

>It’s also nice to have other Apple products that MS could never produce because they worshipped at the alter of backwards compatibility.

It is looking more and more like Apple cannot produce products other people can, not the other way around. Most of their recent product announcements(2017 and after, especially) have begun to look quite disappointing in my eyes.

By definition, the less code you have to maintain the less surface area for vulnerabilities.

If that were true, you would see Apple’s products trending downward when other competitors are trending upwards.....

>By definition, the less code you have to maintain the less surface area for vulnerabilities.

It does not sound like you have any legitimate information about how Windows compatability increases vulnerabilites beyond vague statements. Trust me when I say that no one here thinks Windows is terribly secure, just that Mac OS X isn't secure at all either, and doesn't provide backwards compatability on top.

>If that were true, you would see Apple’s products trending downward when other competitors are trending upwards.....

Is this not the case? Lowered product buys have been offset by higher prices, but the craze of Apple products of the early to mid 2010s is over; I have also only heard negative things about their recent products too. The failed keyboard on Macbooks, how buggy the recent OS release has been, both of my roommates who have iPhone Xs are thinking about switching to Android after being plagued with constant hardware and software issues. Of course, it is possible I live in a bubble, but most people I know who have Apple technology complain a lot more about their tech than I do.

Samsung and every other vendor outside of China have announced lower sales as people keep their phones longer.

While everyone admits that Mac keyboards are horrible, Mac sales are so small compared to everything else Apple sells, it doesn’t really hurt their bottom line

both of my roommates who have iPhone Xs are thinking about switching to Android after being plagued with constant hardware and software issues.

Anecdote is not data.

It does not sound like you have any legitimate information about how Windows compatability increases vulnerabilites beyond vague statements.

Let me tell you a little story about how the half dozen different ways to represent and encode a string in Windows led to a security issue in IIS where you could just encode a DOS command in a browser window and have it execute remotely....

https://www.sans.org/reading-room/whitepapers/threats/paper/...

Or a DCOM vulnerability in 2019. No one uses DCOM anymore.

https://www.cvedetails.com/cve/CVE-2003-0352/

As it says right on the URL, that CVE is from 2003, not 2019, and it doesn't affect any Windows version past XP.

Yes and by then you already had the 7 plus versions of defining a string dating back to Windows 3 (1992). Backwards compatibility was already a problem then. Since then, MS has had 15 more years worth of cruft on top of that s d yet more ways to define a string. If backwards compatibility caused a security vulnerability back then, how much old Windows 95 code do you think is causing issues now?

Oh, crap! DCOM! I'm have trauma just seeing those letters in that order.

What product on the PC does the Apple have that MS could never produce?

How is MS doing with phones? tablets? Watches ($15-$20 billion dollars in revenue and profitable), heck even set top boxes?

I was not aware that phones, tablets, and watches were on the PC...

But on that note: the Surface tablet is considered the best hybrid PC in its class, and that includes Apple devices. The XBox has sold several tens of millions of units in each generation.

Oh, and here's the kicker: Microsoft's hardware is universally superior to Apple hardware in the same class. You can still use Microsoft hardware from the turn-of-the-century. Even Zunes still work (though you need to connect them as MTP devices instead of using the Zune Store). I have an HP Windows Phone that still works, and still handles call better than an iPhone.

I don't know anyone who still has an Apple laptop, PC, or other hardware more than 10 years old. Apple products simply aren't built to last more than a handful of years before they fall apart because you're using them wrong.

>Not having to worry about yet another security vulnerability caused by 10 year old code does impact me.

10 year old code is insecure as opposed to the just-written code that replaces it?

They didn’t “replace it” they got rid of it. There was no replacement for Carbon for instance.

MS's backwards compatibility is definitely great. It's probably also a big opportunity cost and a hindrance to them developing new things. Tradeoffs everywhere.

”Tradeoffs everywhere”

Exactly. Even though breaking backwards-compatibility is annoying for users, Apple doesn’t do this out of sheer spite.

Apple being able to move quickly definitely benefits consumers. The OP's specific example was Apple's ability to execute on the iPhone using current tech but without the restraints of supporting legacy development.

Microsoft did manage to port Windows to mobile platforms, with success to some extent. Notice that it's Windows for Desktop that maintains support for legacy software. Support for legacy software isn't necessarily detrimental to success. Microsoft's lost battle in phone market has nothing to do with it. In addition, Microsoft seems to be doing just fine in the other markets, and it's been growing on cloud, which we may consider to be a new market. The whole 'support legacy and risk being stuck on the past' argument makes no sense.

Windows on mobile was a mess.

You couldn't just take any Windows app and run it. It had to be a specially designed UWP one of which there weren't many. And then there was different OSs depending on whether it was on ARM or not, running UWP exclusively or not. I am a developer and even I couldn't understand the rules.

It's unquestionable that Microsoft's decision not to encourage developers to rewrite their apps (or at least offer a Carbon solution) really hurt them in the mobile space.

You can't run any Mac apps on iOS.

I think Microsoft's big problem on mobile was if anything a first-mover disadvantage; they had Windows Mobile designed to compete with Blackberry and Java ME devices. The iPhone was able to basically produce a second-generation product that bested all of those things. Microsoft wasn't able to catch up the second generation by the time the market was entrenched, and I suspect that it was less an engineering challenge caused by backwards compatibility (backwards compatibility didn't hinder the NT kernel from powering the Xbox One) and more of a cultural challenge.

The Xbox One isn’t a great example of a successful product. The whole XBox division still hasn’t recouped its losses.

https://www.extremetech.com/gaming/251038-theres-no-profit-m...

The Xbox One has had problems from a business perspective, but I think they have more to do with the Kinect inflating the launch price than they do the NT kernel.

not making money on the xbox 1X. but making profit on the One S. which is an excellent piece of hardware for the sub 300 you pay for it.

This is partially right -- before "modern" Windows Phone spun off of 8.1 and 10 there was a Windows 7-esque mobile OS, and one that tracked XP. They still ran on different processors, but with an API very similar to the desktop OS. But, comparing them to Android they did not have any major benefits or detriments.

What I think made them fail was their relative obscurity. For non-Intel platforms, Microsoft is very much in favor of boot locking and a lack of consumer control, perhaps at the behest of hardware creators who insist on NDAs. So these things were never really publicized.

What “success” was that? Sure it had some modicum of success with Win CE/Win32/and .Net compact framework but success was fleeting. The development environment was nice but they ended up completely killing backwards compatibility with the newer versions and abandoned all of the enterprise/ruggedized devices. The last version of Visual Studio that supported then was VS 2007 while there were plenty of $1000+ ruggedized devices and kiosks that still ran Windows CE - I developed for them back in the day.

Apparently people keep forgetting that laptops and 2-1 hybrids are desktops as well.

Microsoft is doing pretty well there, to the point that in many shopping mall shops in Europe, Android tablets have been slowly being replaced by hybrid laptops running Windows 10.

The surface line is still very much behind in sales compared to Apple.

There are plenty of other options to buy Windows 10 devices from, including Android vendors like Samsung and Huawei.

And all together they sell less than a fifth of the number of mobile devices being sold. There is a reason that MS is running away from Windows and that they put .Net Framework in permanent maintenance mode. They don’t even want to spend money supporting their own browser engine.

App store gold rush is gone.

As for the .NET remark, you seem to be really pissed off about something, always bitching about MS stuff.

.NET couldn't be better.

I’ve never said that .Net wasn’t great. I said it would be dumb to tie your career to .Net Framework that is Windows only and is maintenance mode instead of .Net Core. I’ve had enough conversations with you on HN that I know you know the difference.

And I’ve said that most of the energy is going toward mobile, the web and services. No one outside of game makers are making money selling mobile apps. But they are making money selling services that users interact with over the web and mobile - not the desktop.

.NET Framework and .NET Core are .NET.

Windows is doing pretty well across 80% of desktop, laptop and 2-1 hybrid devices.

As mentioned, I spent the last 5 years doing Forms/WPF/UWP, and there is plenty of work there. Just as there is for anyone wanting to bother with either Qt or macOS.

At least in Europe no one starves by focusing on being a desktop developer.

Again, I know you know this stuff - you were involved in the discussion when it was posted here.

As mentioned, I spent the last 5 years doing Forms/WPF/UWP, and there is plenty of work there.

There were also jobs for Windows CE developers in 2015 after MS abandoned it in 2007, would that have been a good career move?

Microsoft said that they are only focusing on compatibility with Windows for .Net framework going forward and they will be adding features to .Net Core that won’t be ported to .Net Framework.

https://www.itwriting.com/blog/11291-state-of-microsoft-net-...

This is now changing. Microsoft has shifted its position so that .NET Framework is in near-maintenance mode and that new features come only to .NET Core. Last month, Microsoft’s Damian Edwards stated that ASP.NET Core will only run on .NET Core starting from 3.0, the next major version.

From the C# program manager.

Async streams, indexers and ranges all rely on new framework types that will be part of .NET Standard 2.1. As Immo describes in his post Announcing .NET Standard 2.1, .NET Core 3.0 as well as Xamarin, Unity and Mono will all implement .NET Standard 2.1, but .NET Framework 4.8 will not. This means that the types required to use these features won’t be available when you target C# 8.0 to .NET Framework 4.8.

.NET Core is yet to be embraced by most enterprises, everyone is waiting for .NET 5 to actually make the switch. Too much software is now making the transition into .NET Framework 7.x, to bother with half baked support from Core to many APIs like EF6, WCF and components from third party vendors.

From where I am standing, we don't care what Microsoft calls it this month, it is .NET.

It’s not about what you care about the industry is definitely moving further away from Windows desktop apps as it’s main focus and is definitely moving toward services and mobile. How much clearer can you get than the CEO of MS saying they are distancing themselves from Windows as a priority? If enterprises are embracing Linux even on Azure, their definitely not using .Net Framework.

Today more Linux VMs are being hosted by Azure than Windows. Microsoft is making serious investments in cross platform development and making Linux a first class development environment on Windows with improvements in WSL.

EF6 and WCF is legacy. Really? Are we arguing about SOAP in 2019?

It’s not about what Microsoft is “calling it”. It’s about where all of the future emphasis of .Net improvements are.

The industries I move on, we don't care about coffee shop programming trends.

Lets talk again in five years time.

So REST instead of SOAP is a “coffee shop trend”? That migration has been going on for at least a decade.

We already know the “trend” that .Net is going in from no less than the program manager of .Net. That “trend” is no major new features for .Net Framework and it will be stuck at 4.8.x forever while all of the innovation is happening with .Net Core.

We also know the “trend” of MS’s focus from no less than the CEO of Microsoft - a de-emphasis on Windows and an emphasis on cloud and moving to where the customers are - Linux and services for both web and mobile users.

Don’t you think it’s quite telling that the biggest news for Windows 10 lately is improved support for Linux via WSL 2.0 and how much emphasis is being put on VS Code?

It is, just right now I am working for a customer where SOAP keeps being relevant as ever.

.NET Core will be relevant for this kind of customers in about 5 years time, again we are still targeting 4.7.2 for bleeding edge customers.

Microsoft wants to win the hearts of FOSS developers that buy Macs as pretty UNIX for developing GNU/Linux software and are unhappy with the latest "Pro" offerings without having ever written a single line of Objective-C or Swift, that's all.

It is, just right now I am working for a customer where SOAP keeps being relevant as ever.

This might surprise you, but the world doesn’t revolve around your niche.

NET Core will be relevant for this kind of customers in about 5 years time, again we are still targeting 4.7.2 for bleeding edge customers.

By definition, customers using .Net Framework and SOAP are not bleeding edge. They wouldn’t have been bleeding edge 5 years ago. The industry was already moving toward REST.

Are you really saying that all of Microsoft from the C# Program Manager up to the CEO is wrong for focusing more on mobile, services, and the web than Windows desktop because your niche is still focused on the desktop?

You do realize that most of the world’s only interaction with a “computer” is via the web and mobile don’t you?

>App store gold rush is gone.

Depends what "gold rush" means. The App Store makes more money than ever.

For the top 1% of app developers that try their luck, out for people living in first tier countries.

And if we remove the FAANG related products even less.

The world is a little bigger than that.

It’s not just the 1% of app developers making money developing apps. There are plenty of mobile developers making money working for companies that sell services that customers use on mobile devices. I was making a living over a decade ago doing “mobile development” working for a company deploying to ruggedized Win CE devices.

The only way to make money from people using “computers” in developing countries is by targeting services at people using mobile devices. They definitely aren’t using Windows PCs and the companies creating services aren’t hosting on Windows.

As far as “FAANG” related products, none of those companies are making most of their money from selling apps on the App Store. They are also making money from services/advertising that people access via the cell phone.

You don’t make money statistically as a mobile developer by trying to sell an app on the store. You make money from working for a company as a mobile developer just as you do from being a web developer.

Yep, and we are happily doing that by deploying PWAs and mobile ASP.NET applications on IIS.

And how does WPF and Win Forms help there? You’re (hopefully) not communicating with mobile apps and websites using SOAP.

I know IIS like the back of my hand - heck I was deploying Classic ASP apps to it almost 20 years ago. But, I see where the industry - and MS are going. I’m not going to stick my head in the sand like VB6 users 10 years after it was abandoned by MS.

Microsoft Surface sells are estimated to be 20% of Mac sales.

https://www.zdnet.com/article/surface-by-the-numbers-how-mic...

Apple sells about 20 million Macs a year and even that pales in comparison to mobile sells.

The total sales of computers is estimated to be around 63 million last quarter according to Gartner. Compared to 310 million cell phones (https://www.androidcentral.com/phone-sales-continue-decline-...)

> "Microsoft Surface sells are estimated to be 20% of Mac sales."

I'd say hardware sales like that are impressive, coming from a software company.

Apple's original core competency was creating hardware, with the software existing to drive hardware sales. (That there is a "hack" in "hackintosh" suggests that this should still be true today. They don't want you to buy their OS and run it on your own hardware, but they're fine with you buying their hardware and running your own OS on it.) Microsoft's core 'competency' has always been software. (I am of course abusing the term 'competency' in this case...)

Apple’s Mac sales are minuscule though in the grand scheme of things and only a small part of their revenue.

Their other products, while locked down to forbid running third party operating systems, are still nevertheless hardware. With the exception of itunes and safari on Windows, virtually everything Apple does involves selling hardware to consumers.

iTunes on Windows was to sell iPods.

Bingo. So that leaves Safari, which apparently no longer supports Windows but once did in the past.

Apple has a number of services available on other platforms, such as Apple Music.

I remember one story about a computer store in old days that carried lots of systems and parts. The new management decided to optimize this and carry only the popular choices and thus increase the profits. But the profits dropped instead. Before that the place was frequented by tech lovers. When their non-tech friends asked what and where to buy, they referred them there. Once the place was not good for tech people, the schema broke.

Mac sales may be small, but they are important part of the system. Who will write all these apps and on what hardware?

Developers will buy Macs and write apps for iOS regardless because that’s where the money is. Just like game developers use to buy expensive, obtuse, proprietary setups to write games for mobile.

So what? I was speaking about all PC OEMs selling Windows 10 devices, not only Microsoft.

Apple hardware is hardly relevant outside US and a couple of first tier countries.

Mobile phones are good for exactly that, phone calls, Internet and a couple of service apps, hardly for producing content.

So what? I was speaking about all PC OEMs selling Windows 10 devices, not only Microsoft.

I also posted estimates where computer sells are a fifth of mobile sells.

Apple hardware is hardly relevant outside US and a couple of first tier countries.

And neither are PC sales. Most people’s only computer is a phone and most of the money coming from those people are on mobile.

Mobile phones are good for exactly that, phone calls, Internet and a couple of service apps, hardly for producing content.

And unless you’re Microsoft or Adobe, you’re not making money selling PC software to produce content. Companies making money on software are either doing it indirectly through services or by producing applications/services running on the web or mobile.

There is plenty of money to be done doing desktop software, in fact that is what I have done the past five years, before switching employer.

So can you name one software vendor outside of Microsoft and Adobe making any real money selling desktop software - besides games?

How fast do you think a startup founder would get laughed out of an investor’s office of he started his pitch with “I have this great idea for a Windows Application”?

On the development side, it would be crazy for any developer to tie their horse to .Net Framework when MS has publicly said that it is in maintenance mode.

There were still a few jobs for Windows CE developers as late as 2014 even though MS abandoned it after VS 2008 (2007?). It would have been foolish for me to take such a job.

Plenty of companies selling medical and industrial control equipment with Windows 10 or Qt based UIs.

And a big bunch of Fortune 500 buying consulting for internal applications.

So if you want names just pick random companies from Fortune 500 list, whose main business is completely unrelated to selling software packages.

And again, .NET Framework and .NET Core, are as we say in Portugal, sand out of the same bag, .NET.

Finally, Azure runs on Hyper-V, not Linux.

No they aren’t and you know this. .Net framework is in maintenance mode and not being actively updated with new features. That’s like saying .Net Compact Framework and .Net Core are the same thing.

As far as Azure running on Hyper V and not Linux, that’s irrelevant since developers of Azure are running in Linux VMs more than Windows VMs.

You are the only one here fighting hard to make the distinction between .NET Framework and .NET Core, as if that would matter to Fortune 500's.

They don't care for one second, we keep getting greenfield projects with deployment requests on top of 4.7.2. And brownfield projects still using 4.5.

For them even having the possibility to migrate to 4.8 would be a dream come true, let alone something like .NET Core.

So no matter what Microsoft wants to push, they have to keep these guys happy.

I’m not the only one “fighting hard to make the distinction”. Technically the distinction isn’t minor. One runs only on Windows, is in maintenance mode according to Microsoft and won’t get any major updates going forward and one is cross platform, actively being improved upon, and is the future of Microsoft’s development platform according to Microsoft

And you’re in Europe, do you really have an on the ground pulse of what most Fortune 500 American companies are doing? I’m frequently looking at the job boards and most greenfield projects are (unfortunately it’s my favorite language) not focused on C#/.Net at all and those that are trying to get from under both Windows licensing and the increased resources requirements and are moving toward .Net Core.

There are still holdouts for VB6 and I still see jobs wanting people to maintain them. But would you say maintaining legacy VB6 code was a good career move?

Most Fortune 500 operate world wide.

As for VB 6, if they pay alright and offer nice remoting package, why not.

Actually I did a couple of VB.NET (.NET 4.5) support activities during last year.

See, most people around this side of the pound aren't career obsessed.

We get to work, enjoy our IT unions (in some countries), 30 day vacations and have a life outside job.

As long as it pays the bills, most people are happy with what they get.

Until the company decides to change directions, you suffer from salary compression and inversions because HR decides to pay new employees market rate and (the hypothetical) you only get cost of living raises.

The last thing I want is to be tied to a company because I am not competitive in the market and then start complaining about “ageism”

And I’m on the opposite side of the country from the west coast FAANG and startup culture. I’m the definition of a standard “enterprise developer”.

What do you consider "real money"? There's a whole bunch of industry-specific software vendors like Autodesk and The Mathworks that have plenty of dependent companies still buying licenses for desktop software.

Autodesk had a net income of a whopping $40 million last quarter.

https://www.macrotrends.net/stocks/charts/ADSK/autodesk/net-...

And MathWorks revenue is about a billion.

Compare that to even your middling SAAS company. I’m only saying that it is silly for someone to invest their future employability on a platform (desktop Windows) and a runtime (.Net Framework) that even MS is moving away from. Everything they announce is about Azure, .Net Core, and heck even Android these days. I’m not even saying that Windows as a server OS is the worse thing in the world - even though it’s a lot cheaper in both licensing costs and resources to use Linux even if you use .Net Core.

And how much has it cost Microsoft?

Are we really worried about how much money it will cost to the company with the largest free-cash pile in the world? They have literally tens of billions of dollars of cash literally just sitting around doing nothing.

They can afford to throw 0.0001% of that at maintaining backward compatibility.

As pointed out elsewhere in this thread, it’s not just about money. It’s also about path dependence – if you are dedicated to backward compatibility there are lots of changes you can’t easily do to your products. It can hinder innovation.

How much has it cost MSFT? They have a trillion dollar market cap, so I would say that’s not a particularly relevant question.

Well Uber also has a nice market cap. Only on HN is market cap more important than silly things like profit.

But MS recent success has nothing to do with maintaining backwards compatibility. They had to purposefully move away from emphasizing Windows to be successful and move to the cloud, Office on other platforms, abandoning IE, give up the idea of ever selling Windows to end users when it isn’t bundled with the computer, building IE on top of Chrome, and by announcing an Android device.

Does that seem like Windows was a winning strategy?

Blackberry had a nice market cap too, a year or so before they went under...

Windows runs on refrigerators and signage panels and ATMs just fine. It's not based on since inherent technical inadequacy, people just don't like MS products. Usually because they're late and don't quite align with what normal people want... But I don't think this has anything to do with backwards compatibility. From the Zune to WP10, what kills Microsoft in an area is "Also-ran", "not good enough", and "kinda clunky"

How else do you define as “running just fine” than - something people actually want? Most of the kiosk/vertical market type devices are running a version of Android now. The vendors that use to sell ruggedized Windows mobile devices like Intermec have since started reflashing them to run Android.

I have a 32-bit application I use regularly to interface with old hardware (which isn't on the list incidentally; I suspect there are many other niche applications like this).

The hardware vendor isn't going to re-write it. As such I have to...

1) Avoid MacOS upgrades forever, or 2) Let the hardware become nearly impossible to use (there are workarounds on the device itself, but it's really painful to use), or 3) Write a new application myself.

If my 6-year old MacBook can last until I find something competitive with Logic Pro on Linux, I won't be buying another Apple machine. There's nothing else I need it for.

> I have a 32-bit application I use regularly to interface with old hardware

And the relevance of this use case is something that "using a laptop in a coffee shop" Apple users seem to be completely and utterly oblivious to.

Yet that hardware often has a useful life far longer than the computer plugged into it, or the software that runs on that computer. On top of that, there may not even be a suitable/comparable "modern and supported" replacement for that hardware.

Is it possible to virtualization an older release of OS X and run your app? It might be hard to find an image but it can be done. I have a 2011 MBP that bit the dust and I’ve seen been able to virtualize OS X (not sure about MacOS) and load a back up from Time Machine.

Certainly. Proof, ca. 2017: https://jasomill.at/X×X.jpg

Side note: the Tiger, Leopard, and Snow Leopard VMs in this screen shot were running OS X Server to comply with Apple's pre-Lion license restriction prohibiting virtualization of non-Server versions of the OS. I have no idea which hypervisors, if any, enforce this restriction (nor if old versions of OS X Server are still available for purchase directly from Apple and/or available for non-production use through the Apple Developer Program as they once were).

It is possible.

That hardware has software that works on modern Linux kernels?

No, but if it doesn't have software that works on my Mac either, that's one less reason to stay on the Mac. At this point I'm down to one remaining reason to stay on the Mac (Logic Pro) and I'd like to find a feasible alternative.

Have you taken a look at Bitwig Studio? Runs natively on Linux.

Not yet, I'll give it a look.

I feel the same. I started a project just four years ago, and now I feel that I'll need to rewrite a huge portion of the graphics code to work with future Mac versions.

I also have a lesser fear that in order for my application to work on some future version of Windows, I'll need to deal with code signing and directory access rights nonsense in order for customers to think that the software "works properly". But that's just speculative. Apple's already doing it.

Why should I trust your application to run rampant on my computer? Since when was it a bad thing to isolate what an application can do? Even if you aren’t being purposefully malicious and being sloppy you can still cause damage. Look at what happen with the Google Chrome installer on Macs that had system integrity protection turned off.

https://support.google.com/chrome/thread/15235262?hl=en

It's nonsense because it didn't exist before. Imagine you're a long-time user of my software. You upgrade your operating system (because's Apple's marketing tells you to) and find that my software doesn't work anymore. Who made the change? Apple. Who do customers complain to? Me. Despite any argument, Apple's change has cost me $1000+ of development costs for users to continue using my software. Apple is not my customer, and I don't feel that I owe it a penny. If MacOS didn't exist, it'd only make my life easier. If this change only affected new users, I'd have no complaint.

So in that case Apple should never introduced tighter security?

Well, if you want yo stay in business selling your software, you have to keep it updated. Would you also be upset if you couldn’t still sell 68K software in 2019? You either update your software or people will buy something else.

BareBones managed to sell and keep updating BBEdit between Apple transitioning from 68K Macs, PPC Macs, OS X and intel Macs.

Did you really think that you would never have to update your software? How long have you been in business?

It's always true that software must have a maintenance budget due to changing user markets.

My point is that compared to Windows or Linux, Mac is very expensive to maintain software for. If your software has mostly cross-platform code, I would estimate based on my experience that the Mac-specific work is 10x that of the others.

Let's just say thank God that `com.apple.security.cs.disable-library-validation` exists, so the 150 plugin developers for VCV Rack don't have to go through the notarization nonsense that I do. That would truly kill the software, or would force users to disable GateKeeper altogether (with `sudo spctl --master-disable`) to use.

You really think that Microsoft or Adobe spends 10x as much to support the Mac? If the ROI on Macs is so horrible compared to Linux then why wouldn’t Adobe or MS be dropping Mac support and instead support Linux?

Yes, platform-specific expenses are 10x on Mac vs. Linux and Windows. I said nothing about ROI. You're confusing that with platform-specific expenses.

ROI of platform = number of platform users * price - "platform-specific expenses"

Hopefully you can see why even if "platform-specific expenses" is high, the ROI of the platform can be higher than another platform, because the number of platform users is much higher.

If it is so much more expensive to support the Mac than Linux, why won’t the major software vendors support Linux?

Are you going to drop support for the Mac?

Because the "number of platform users" is very low (say 0.2-2% for professional users). I explained that above.

>Are you going to drop support for the Mac?

Probably not, but I might begin charging 20% more for Mac versions of the software.

You're not doing work for Apple, you're doing work that helps your existing customers have more secure systems.

If my customers wanted this, I would gladly budget for it. But nobody has asked for this, even when I bring it up.

The reason (I imagine) is that people use my software in professional/home studios where you typically have a dedicated computer to run the software. People purchase a computer based on the recommended specs of my software, and some don't even bother connecting it to the Internet after getting everything installed. If they do, they don't download random software from the Internet because 2 or 3 packages is all they need. There just aren't really any attack vectors that things like app translocation, notarization, code-signing, runtime entitlements, app stapling, paying $100/yr to be an Apple Developer member, etc would solve.

So these users are unaware of 32 bit apps breaking on update and also going to the trouble up updating a computer that isn’t connected to the internet?

You don’t have to pay anything to get your code signed....

"Beginning in macOS 10.15, notarization is required by default for all software." https://developer.apple.com/documentation/security/notarizin...

Notarization requires that the software is signed with a certificate obtained from your $100/yr Apple Developer membership.

If you don't have a membership, you're not signing with a Developer ID certificate, so you're already requiring users to bypass Gatekeeper (with right-click + Open or by fully disabling it) to use your software. Those processes also bypass the notarization requirement.

You don’t have to have a paid developer account.

Do you have a source of this? I was unable to notarize my software with my free Apple Developer ID account, so I paid $100 this year. This question (https://forums.developer.apple.com/thread/121113) seems to be answered "no". I would love to be mistaken so I can treat my family to a fancy meal tonight.

https://wildfiregames.com/forum/index.php?/topic/26149-macos...

This service requires that the finished application be sent to Apple to be notarized as part of an automatic (and free) process, and that the returned "ticket" be "stapled" to the Mac app before distribution.

https://www.quora.com/How-long-will-it-be-before-Apple-requi...

My Google fu is escaping me because every citation of “free” is in the context of “virus free”

My subscription is about to run out in a few months, so I'll let it expire and test this, as it would save a bit of money, and a continuous Apple Developer membership isn't necessary in order to sign apps infrequently.

> I'll need to deal with code signing and directory access rights nonsense

It might be nonsense (and a nuisance) for developers, but it's far from that for everyone else.

It's about time developers stop assuming they can do whatever they want on everyone else's machines.

If it's my computer, developers don't have the right to screw it. Period.

Absolutely. Microsoft will even preserve behavior that most people would consider bugs, just to keep old software running. I would say backwards compatibility is a guiding principle of consumer software development at MSFT.

Edit: It really says something that the top examples in https://en.wikipedia.org/wiki/Bug_compatibility are from MSFT platforms.

> This is not true of ... almost any open source project.

I don't think that's generally the case, in part because of the very nature of open source. And certainly, for example, I can easily compile over 30-year-old TeX and run 30-year-old Emacs code.

> This is why Microsoft is the better long-term play...

There’s never going to be one single factor that makes one platform ‘the best’ for everybody. Windows platform improvements has been a dumpster fire for over a decade. Windows 10 seems to have staunched the bleeding by being a proper desktop OS again, and Windows subsystem for Linux looks cool. But their dev story is a hot mess, with a trail of the carcasses of half assed and abandoned dev frameworks behind them.

Different users and developers just have different needs and priorities.

Is this really true? I feel like I keep hearing about how $companyX is still on Windows 95/98/XP/7 and old software is often cited as a reason for not upgrading.

There's plenty of areas where Microsoft has dropped the ball on backwards compatibility. They've got operating systems they no longer support, including Windows 10 Mobile in just a couple months. There's hardware platforms they no longer support. Their "PlaysForSure" content can no longer be played. Their web browsers drop support for old technologies all the time, e.g., Edge doesn't support ActiveX. Xbox One is only backwards compatible with some Xbox 360 games, and Xbox 360 was only backwards compatible with some Xbox games.

Really, the only area where Microsoft has bent over backwards for backwards compatibility is Windows-on-x86, because that's been their golden goose for 30 years.

A couple weeks ago, their CEO said "The operating system is no longer the most important layer for us". Will they continue to work their butts off in the name of backwards compatibility, or will they start treating it the same way they do their services, media stores, game consoles, and web browsers? I'm not optimistic Microsoft is going to be a better long-term play in this department.

[1]: https://www.theverge.com/2019/10/3/20896908/microsoft-window...

> Edge doesn't support ActiveX.

They maintain Internet Explorer 11 which has support for ActiveX...

This has not been my experience. Moving from dos/Windows 3 to 95 broke a lot of software, particularly games. The upgrades to Vista and 7 broke so many programs that they included an XP VM to get old stuff to run.

Microsoft's problem is potential customers refuse to do business with them because of Bill Gates failure to act in good faith during the desktop era. See also how everything Intel acquires shrivels and dies.

Same will happen to Google and Facebook.

....for people who choose to sell software it might be. For users this might not matter much at all. I’d rather they remove the bad parts than keep old software working.

And I have New Yorker's Digital Archive which doesn't except in my VM with an old Windows XP.

Nobody is perfect and you can't rely on anyone to solve this problem for you.

> This is why Microsoft is the better long-term play: Backwards compatibility

Small friendly jab: My trusty Photoshop CS2 doesn't agree.

Pretty much why the z and i Series have faithful followings, the investment made in some cases decades ago still has value.

> I have code that has continued to run, untouched, for decades.

This isn’t always a good thing. I usually don’t like when a piece of software feels like it hasn’t been updated since the 90s. But I guess we don’t want some military defense system to break because it’s 32-bit either.

I could hardly care less how modern a piece of software "feels" as long as it's usable. It doesn't stop being usable 5 years later due to being out of fashion. Many recent redesigns (hello Gmail) are quite the opposite to me btw.

I care about fixing security holes, and actually useful improvements (to the UI as well, why not, when they make sense)

Well, when the software doesn’t support all of the new features of the platform, I definitely care. I hate that even today four years after it was introduced, there are still video apps on iOS that don’t support PIP on the iPad and apps that don’t support split view.

That is literally my metric for how good a piece of software is. If it can run for a long period of time with minimal to no maintenance that's the hallmark of good engineering. I couldn't care less about how dated it might look as long as gets the job done.

Reply on news.ycombinator.com