Let's keep the discussion above the 'he said, she said' level, shall we?

This isn't a social networking site, after all. :p

Before the thread started to diverge, it started out as a discussion of the quality (or lack thereof) of Apple's s/w releases, and, of course, the opinions did not agree. I believe that differences of opinion in this case seem related to different experience and background, and it seems to me that many of the posters have not considered the bigger picture of what I consider to be "real world" s/w development methodologies, standards, and quality.

For more years than I care to count, I have been involved in design and development in several s/w intensive programs involving multi-billion dollar contracts, starting as a low level worker and progressing to management at various levels. These programs involved hundreds of thousands of lines of code developed in-house as well as utilizing already developed s/w components to assist in the test, simulation, and management functions. In a typical program, the operational s/w segment performed tasks of command and control, navigation and path control, sequencing functions, communication functions, attitude and flight control functions, data acquisition, processing, culling, compression, and formatting functions, target analysis, acquisition and tracking functions, threat detection, assessment and avoidance functions, as well as control of sophisticated sensor packages. All of these functions were integrated within an in-house developed real-time operating system. All of this, when delivered to the customer, worked consistent with the specifications of the contract. As a matter of fact, one of these programs has been operating for well over a decade 24/7/365, essentially unattended except for sequence upload and data download periods. Throughout this time, only one, repeat one, s/w performance fault has occurred that is not attributable either to operator error, hardware failure, or environments outside of what was defined in the specification.

This, then is my background which forms the basis of my contention that the quality of the s/w released by Apple, especially in the recent past, is of inferior quality. Even if one were to believe (and I don't) that more than 99% of Apple s/w users experience no s/w problems, such a "success rate" is considered totally unacceptable for released s/w in most "real world" situations.

There are numerous examples of high-quality software releases in the "real world". Consider the software required to fly a modern airliner - would they even be allowed to fly if the on-board s/w reliability was on the same order as Apple's? Or the s/w involved in the financial industry which processes billions of financial transactions daily - who would patronize a financial institution that guaranteed that 99% of financial transactions would be correct? Or the s/w required to run an automated factory. Or the s/w required to operate a nuclear power plant. Or, even in mundane applications, like the controller involved in microwave ovens. No, when delivered, this s/w has to be virtually error free. If it is not, it is not considered ready for release, period. These are examples of the high standards of s/w that can be achieved by organizations with competent s/w developers competently managed.

I do not state that Apple's s/w developers are incompetent, nor, necessarily their managers. The fact remains that, as a corporate entity, they do not release s/w that even approaches the quality levels that many people are familiar with in their daily lives. Note that the operative term here is "released" - my position is only that Apple releases its s/w a few (quite a few, actually) interim build cycles earlier than the high quality s/w developers do.

I do believe that the fault lies with the particular business model. There is no formal contractual specification between Apple and the people to whom the s/w is delivered. As a matter of fact, in Apple's case, the recipients of the released s/w have no direct control over the specification. When Mr. Jobs makes presentations about the features and capabilities of a coming s/w release, that, at least in my mind, represents an implied de facto contractual specification that he is obliged to fulfill in exchange for my money. Based on what I have seen, heard, and read, 10.5.0 represents a non-fulfillment of this implied contract, to wit, some features he seemed to "promise" were not present or did not perform in accordance with the implied performance. Given past history, he will probably eventually fulfill his promises (more or less, anyway), in a .2 or .3 or later (perhaps a much later) release. In the "real world" with which I am familiar, this type of s/w delivery to a customer would result in at least the loss of millions in contract fees, and very likely the loss of billions in follow-on contracts that would be awarded to another company. In Apple's case, there is no credible alternative company. That is why they have gotten away with these practices. And, unless a viable alternative emerges (and I'm not holding my breath), they are likely to continue to release immature s/w of substandard quality.

I will grant that Apple's business model does not require high quality s/w to be delivered with each release. Neither does Microsoft's. To claim that Apple's is superior to Microsoft's is just the "honor among thieves" argument.

I use Apple s/w. I have used it with whatever version was delivered with the original Mac Classic all the way through 10.4.11, and I will probably continue to use it. After a fashion, it meets the needs I have for personal s/w usage. I also accept that Apple s/w is of much lower quality than other s/w which I encounter in my daily life in the real world. The reason I accept it is because neither my livelihood or my safety depends on it; it is merely a convenience for some of the things I like to do. However, that does not mean that I have to like it or to sing praises of how great it is, as others are wont to do.

I hope you didn't develop carpal tunnel syndrome while writing that.

Where can I go to purchase the retail copy of your product? What is the list price?

If you affiliate with certain gov't organizations and have the necessary clearances, you can get it for free.

Then it's not of much relevance to me, is it?

The implication is yours, my friend, not mine. I don't think Windows is a bad operating system; I think Mac OS is a better one. that may be a function of the way I use computers, or because I happen to be a sophisticated user, or because I don't happen to do the things (like FPS games) that Windows has an edge on. I'll add that I'm agreeing with both sides here. on one hand, I recognize that apple did less than their normally stellar job of preparing 10.5 for release (though I'll remind everyone that the first shipping of os 9 just didn't frigging work). on the other hand, I'm pretty damned impressed with 10.5, bugs and all. it's a nice piece of engineering. I'm sure they'll get 85% of the bugs fumigated over the next few months, and there is always going to be that stubborn 5% that won't get resolved until 10.6, if then.

in short, it's all good. :)

>>>> Reply to Craig Arko

Well, I was merely attempting to explain personal experiences which influence my definitions of high quality and low quality. Feel free to discount that if you wish.

However, the (non-personal) examples of aircraft s/w, financial s/w, other industrial s/w, etc., are credible examples of high quality software deliveries for everyday, common-experience cases.

Does the fact that that s/w is probably proprietary and not available to the general public to evaluate also invalidate them as examples of what I consider to be high quality s/w in comparison with "personal computer" s/w?

Flying a plane is certainly complex, but I doubt it compares to the almost limitless possibilities facing anyone developing any OS. An operating system is used by many different people for many different purposes, and while an application may do some impressive things, it is limited to a much narrower field, so any comparison will not be apples to apples.

I'll point out that there is quite a difference between designing a piece of software that's designed to do a particular task (no matter how complex), and designing on operating system, where you cannot know in advance (except in a general way) what the goals and intentions of its users are, what software they will have access to, what conditions they will be working under, or even what machine they are going to be running it on. I bet if you surveyed state governments, you'd find that the machines they have processing our taxes and monitoring us in our sleep were mostly build back in the 80s and 90s, chugging away on Fortran programs. you don't need to be up to date if you've got a system that does exactly what it's supposed to already.

dagnab it, you type too fast!

I think you are underestimating the difference in difficulty level of developing software that will be used by relatively few people, for relatively few purposes, to given specifications, and the difficulty of developing a commercial operating system that will be used by millions of people, who use their computers to do an uncountable number of things, who have not given the developers any specifications. The latter is far more difficult than the former, even if neither are easy.

If you think Apple is doing such a terrible job, and that is feasible to release perfect software in the initial release, then you should be developing consumer software instead of specialty software. I expect that, best case, you would have similar very minor bugs as what Apple has in Leopard. There are many users who have long awaited the perfect OS, so feel free to prove me wrong if you think it is as easily done as you imply; there's definitely a market.

Yes, this is not a valid comparison. High-reliability (not to be confused with bug-free) software is developed using an economic model which is not suitable for assessing consumer grade software. I wouldn't, for example, enjoy paying a thousand or more dollars for each OS upgrade, after waiting a few extra years for them to go through qualification committee testing.

These are not life and death products.

Here is a real-world example of how release decisions get made for consumer grade software:

http://arstechnica.com/news.ars/post...-squashed.html

Well, I made the jump to leopard last night. Initially I tried the upgrade install, which failed very near the end. Went back, started over, and did a clean install. It took forever on a 1.25 ghz G4 Mini (forever, in this case, was a little short of two hours) but eventually it finished, rebooted, updated, and started the fun task of reinstalling software. Luckily I had saved the disk images of most things on my windows box (and sharing folders in leopard is a snap). Two hours later, the system was all set up just like I want it, synergy was working nicely, and the clipboard is even synching between machines. The original problem I had with that was because, apparently, there is more than one version of Synergy's GUI.

However I digress... The point of the post was, so far, I haven't run into a single snag with Leopard. Everything works just like it did with Tiger, only faster.

To those of you who are Windows bashing, whoever made that comment about "no such thing as a perfect OS" is quite right. I don't "like" a lot of things about Windows, and yes, there are some things OS X does better, and some things you just can't do on a Mac. Yeah I know, playing Utopia isn't that important to most people, but the truth (this is an *fact*) is, my Mac (even better in Leopard than Tiger) integrates quite flawlessly with my Windows network, giving me (this is an *opinion*) the best of both worlds.