Hopefully, he does not develop software for NASA or a nuclear plant or an airplane.
Ever read a software license carefully? You just mentioned two items that specifically are mentioned in *MOST* licenses as areas where the software is NOT TO BE USED: nuclear power plants, and flight systems.
As for space exploration... you don't recall the metric/English fiasco recently with the Mars missions? How about the Ariane 5 that had to be destroyed due to a sign error in the navigation systems? You *betcha* they have bugs... and it's quite spectacular when they do.
I understand your point, but sorry, it doesn't hold water. The fact is that every commercial provider of consumer software ships product with known bugs. Always have, always will, IMO.
Which brings me to my next point: melgross, Adobe may have shipped software with no bugs that *you* were aware of, but trust me, internally they were tracking them. 'No known bugs' is a fairy tale told to beta testers and early adopters.
So you are saying that Windows is always released with every known bug fixed?
At some point you have to put it out there, and fix some things later.
As far as developers complaining. What I've heard and what was said on Ars Technica is that Apple has always said the way the kernal was set up was not permanent and would change. Many are confused as to why some developers were not prepared.
Changes in an OS that invalidates a third party's (or even in this case Apple's) software is fine if Apple stated that those functions, addresses, pointers, etc. couldn't be relied on for future use. This happens all the time. I couldn't blame Apple for that.
But that's not what I'm talking about.
And again, MS is one of the worst examples for anyone to use. They are notorious for lacking the ability to produce stable software.
I've been doing software and hardware product management for over 15 years, so I think I can add a bit to this discussion from the "corporate" perspective. In software development, there are usually two high-level questions: 1) when is the ship date, and 2) how many bugs of what severity level are acceptable for a released product. Nobody likes to ship buggy products, but the reality is that, at some point, you've got to say "ship it!"
Many software and hardware companies have moved to the "train leaving the station" model: the software will ship on (insert date here), and it will either be ready to ship by then or troublesome features will be removed to make it ready to ship (i.e., they will "miss the train"). The "train" model ensures that *something* is released on time to meet the expectations of customers, stockholders, the media, etc., and it also ensures that development resources get freed-up for other projects. For public companies, shipping late can have disastrous consequences for stock prices and can put a company at a competitive disadvantage if another company gets their product to market first.
Most product bugs are ranked by severity (or priority): P1 bugs prevent a product from shipping: they affect the primary function of the product for many users and cannot be worked around. Products should never ship with any P1 bugs. P2 bugs are somewhat less severe or have work-arounds; a small number of P2s are present in most shipping software. Priorities often go as deep as 9 levels (a P9 is often related to documentation or minor appearance flaws that have no affect on performance or usability, and is classified as an "enhancement").
I have never heard of any software product shipping without any bugs at all.
Beta testing is really hard for corporate developers. You generally want to keep betas under some control (i.e., not available to just anyone) to allow a bit of surprise to customers, to prevent late betas from being distributed illegally, and to allow the project managers to keep track of test results. Unfortunately, there are so many permutations of hardware and software, that it is literally impossible to develop test cases that cover everything. A problem is that 500 public beta testers may essentially test the same things, and none may hit the critical sequence that triggers a problem. Also, while many beta testers have good intentions, it is often true that they - as a group - don't do the thorough testing expected by the developers.
Some companies I've worked with are even considering paid beta programs: you get the software and a list of things to test; when you submit documentation on schedule that shows test results for specific tests, you get paid. Otherwise, you get a "thank you" and no payment.
I don't plan to upgrade to Tiger until at least 10.4.1, but I think anyone who's using 10.4.0 should be getting a fairly solid release, especially when used on unmodified hardware running common apps.
I agree with you to a large degree. The problem is that we have to look at what bugs are getting out. As I've said, bugs that don't affect function are not always bothered with.
I've also said the same thing about beta testing. you can't find all of the bugs because you can't test for everything. I fill out online forms when I do testing. It's just not the joy of getting a program before everyone else gets it . It's a lot of work. A lot of so called "testers" drop out when they relize that.
Also in-house development tends to be buggy. More than half of all in-house projects fail, and a large number of those that get to completion are never completely validated.
When a software company asks you to test a specific area, they want it tested to completion, or as close to that as is possible. Any bugs that are founf are squashed. As I've said, if the area had insurmountable problems, either development is extended, or the area is re-evaluated. Possibly turned off, or removed.
There do seem to be too many problems with this release though.
Ever read a software license carefully? You just mentioned two items that specifically are mentioned in *MOST* licenses as areas where the software is NOT TO BE USED: nuclear power plants, and flight systems.
As for space exploration... you don't recall the metric/English fiasco recently with the Mars missions? How about the Ariane 5 that had to be destroyed due to a sign error in the navigation systems? You *betcha* they have bugs... and it's quite spectacular when they do.
I understand your point, but sorry, it doesn't hold water. The fact is that every commercial provider of consumer software ships product with known bugs. Always have, always will, IMO.
Which brings me to my next point: melgross, Adobe may have shipped software with no bugs that *you* were aware of, but trust me, internally they were tracking them. 'No known bugs' is a fairy tale told to beta testers and early adopters.
Legalese. Windows NT is being used on out Ageis crusiers even it's failed several times, and they had to be towed into port. The license states that it shouldn't be used for "mission Critical support where lives may depend on it"! Whoa!
Oh please.
You think that NASA knew that they were using both inch and metric units and said "what the hell, let's do this anyway"?
You also think that the ESA knew that they had a sign error, but thought that it was only a plus or minus, so it wasn't important?
You don't think that it's possible that they DIDN'T know about those errors?
I don't know about fairy tales. Adobe doesn't officially tell us that there are no known bugs, or that there are lots of them. All I know is that when I speak to John, he says that we got 'em all. Remember, that means all that we know about.
It is continuous while they are still in pre-release. That's what pre-release development is. It can't be done any other way.
But beta testers aren't supposed to see any bugs in the final stage. I'm sorry, but that's just the way it is.
Interesting. But I have a question. What's the point of beta testing if there are no bugs in the software. I was under the impression that the whole point of beta-testing was to find the bugs in the software.
Quote:
Apple used to be known for very stable and bug free OS releases. Like it or not, this release is considered, by those developers you mentioned, to be the buggiest release since 10.0.
They were? Could you tell me when? For I remember all the problems with System 7, System 7.1, 7.5.1, 8, 9, 10.0, 10.1, 10.2, 10.3...
Systems 8 and 9 were semi-solid (but still had huge problems), but recall these were basically just minor updates to system 7. And since all of the 'classic' OSes were just one bad program from total system lockup, I have a hard time calling that system "solid".
Quote:
I work on a software project right now and we never release any software with a known bug. If it's known, it's fixed. I was kinda amused reading about all companies releasing software with known bugs.
Must be nice to work on a project where you have control over when software is released. But when software is released is dependent on so many factors, trying to generalize with "We never do this, so no one else should either" is just a farce. I work on a software project that's several years (and several versions) old, and I know it has bugs in it. But who's paying me the month's worth of time to fix a bug that maybe three user's might experience over the course of a couple of years? Who's going to tell management that we need to delay the product another couple of months to fix all the irritations, which causes issues with marketing, sales, purchasing, etc?
Oh, and finally, public betas sound like great ideas, but generally, the public are lousy testers. Sure, they can install the software, then complain that this or that doesn't work. Getting bug reports from people is easy. Getting decent bug reports, with info on how it occurred, whether you can recreate it, what hardware/software you have installed that might be interacting with it, etc, that's nearly impossible to get from the general public. I've literally received bugs that read "Received error xxx when running program". Not even an indication of where in the software they were, let alone what they did, what they tried to do, nothing. Which is just a little worse then the users who say they were doing one thing, but actually were doing something else. Or think they know what they were doing. Yeah, I'm sure Apple needs even MORE of those type of reports.
I can pretty much promise you Adobe lets software out with known bugs...
We've worked with Adobe pretty closely on a variety of issues we have had with their software (we're a rather large unit). Adobe has even sent techs to our site to see first hand what our issues are.
So far, we still see issues we reported early in the previous version of Creative Suite in the new one. We also have found issues in the first two days of using the new version that we may consider show stoppers until fixed.
This tells me either Adobe is doing a pretty crappy job of testing different environments or they realize they have to make a buck or two too and can fix some problems later.
I also agree Adobe simply turns off a variety of features that they can't get working well. It's a double edge sword. How often do you suppose they consider all the "planned" updates they cut when they set their pricing?
I use CS, but I only test Photoshop, so I can't speak to the rest.
But Adobe has a very good rep. about their software being bug free and stable. Is it all perfect, I doubt it.
I've found problems with software that was cause by other problems.
Look, for example, at Macfixit, or Macintouch. A lot of problems are reported there. I have had very few of them.
So where are those problems coming from, and who is responsible for them?
Many people, and even large organizations, set their systems up in a way that just begs problems to appear. Some problems are The OS vendors fault for not making things clear enough. Sometimes it's the 3rd party vendors. Sometimes it's the IT department. Often it's a combination of the above.
It's even the hardware. Marginal memory causes many problems that seem to be software related, etc.
All these combinations can't be tested for, and so incompatibilities escape. I find it hard to blame anyone for many of these.
But some combinations are inexcusably not tested.
Apple's Firewire data, and drive loss was one of those. It survived several updates. I hope it doesn't return in 10.4.
You think that NASA knew that they were using both inch and metric units and said "what the hell, let's do this anyway"?
You also think that the ESA knew that they had a sign error, but thought that it was only a plus or minus, so it wasn't important?
You don't think that it's possible that they DIDN'T know about those errors?
Well obviously they didn't know about them, but that wasn't wnurse's assertion. wnurse made a much stronger assertion that simply isn't true.
Quote:
I don't know about fairy tales. Adobe doesn't officially tell us that there are no known bugs, or that there are lots of them. All I know is that when I speak to John, he says that we got 'em all. Remember, that means all that we know about.
Or that are required to ship.
Look, we can argue this till Doomsday, and we're not going to sway each other, so why don't we just agree to disagree?
Interesting. But I have a question. What's the point of beta testing if there are no bugs in the software. I was under the impression that the whole point of beta-testing was to find the bugs in the software.
They were? Could you tell me when? For I remember all the problems with System 7, System 7.1, 7.5.1, 8, 9, 10.0, 10.1, 10.2, 10.3...
Systems 8 and 9 were semi-solid (but still had huge problems), but recall these were basically just minor updates to system 7. And since all of the 'classic' OSes were just one bad program from total system lockup, I have a hard time calling that system "solid".
Must be nice to work on a project where you have control over when software is released. But when software is released is dependent on so many factors, trying to generalize with "We never do this, so no one else should either" is just a farce. I work on a software project that's several years (and several versions) old, and I know it has bugs in it. But who's paying me the month's worth of time to fix a bug that maybe three user's might experience over the course of a couple of years? Who's going to tell management that we need to delay the product another couple of months to fix all the irritations, which causes issues with marketing, sales, purchasing, etc?
Oh, and finally, public betas sound like great ideas, but generally, the public are lousy testers. Sure, they can install the software, then complain that this or that doesn't work. Getting bug reports from people is easy. Getting decent bug reports, with info on how it occurred, whether you can recreate it, what hardware/software you have installed that might be interacting with it, etc, that's nearly impossible to get from the general public. I've literally received bugs that read "Received error xxx when running program". Not even an indication of where in the software they were, let alone what they did, what they tried to do, nothing. Which is just a little worse then the users who say they were doing one thing, but actually were doing something else. Or think they know what they were doing. Yeah, I'm sure Apple needs even MORE of those type of reports.
I think you missed the point on that first one. The key is "final". Re-read it.
If you compare Apple's old system software to MS's, you will find that indeed it was stable. 8 and 9 were a lot more than just updates to 7. 8.5 was very solid, and so was 9+.
Yes, I've said that 10.0 was a problem. It's been described as Apple's second (paid) public beta of X. 10.1 should have been the first release.
Sure, the public is lousy. But Apple has bug reporting in it's software right now. Also, if thousands of people report the same problem, and Apple didn't see it, like the VERY serious Firewire bug, don't you think they would look into it?
My experience is that software is released with bugs but that is no excuse for tiger. Yes, a company has to say at some point ship but increasingly, companies are using their customers as beta testers. Apple has resisted this trend for a while but i guess with OS X, they finally joined the crowd. What is infuriating with apple sometimes is that when you call them to report a bug, they act as if the bug does not exist. Nothing gets my blood boiling like when some cheeky customer service person tells me that nothing is wrong. This is when i get angry and think
1. I am a software developer
2. I was once a network administrator.
3. I once built a computer from the ground up.
4. I know myself around computers. I'm not some little old lady who doesnt know a keyboard from a mouse.
5. The customer service person must think i am an idiot.
6. Who the hell is this person telling me that the error message on screen does not exist. I'm looking at it!!!.
Nevertheless, i'm not a version X.0 adopter. I usually let the fanatics be the beta testers and then get version X.1 or maybe even wait for version X.2.
I think consumers as a whole should always reject version X.0 software. That would force all companies (apple included) to create quality products. This disturbing trend of using customers as beta testers must be reversed.
Anyone having trouble with peripherals (particularly scanners)? We added Tiger to two of eight machines at our organization and an Epson device is now no longer functioning (with Tiger).
Also, any word on when Norton Antivirus will be updated to work with 10.4?
My sense is that 10.3.9 is solid as a rock. I'm waiting for a while with my machine at home before upgrading there. The present idiosyncrasies are not worth the trouble, Tiger's compelling features notwithstanding.
look, anyone getting tiger 10.4.0 are early adopters for the most part. you take your chances. for those buying a new mac with 10.4.0 preinstalled, i assume that for most of them the factory default install is 'clean' enough for them to slowly get into the swing of things.
anyone running anything mission critical who primarily uses a mac will wait till about 10.4.2 before making the jump
I've also read (but not been able to test yet) that VPC networking is broken, and that there are PPTP problems too. I don't see any mention of those in this update :-/
Only the Switch part is broken, where you can assign a unique IP to the PC session. It still works if you want to share your Mac's IP address so networking still works and so does internet access. Since I only use VPC to test IE6 bugs, that's enough for me. If you've got more specialist windows only networking apps that may be a problem.
It broke because Apple changed the networking kernel extension and VPC wedged in to that. It just needs Microsoft to do a small change to fix it.
In 10.4 they changed the way most things interacted with the kernel so this kind of problem shouldn't happen again. There's probably similar issues with things like Norton, some USB devices and various VPN clients which wedged in as kexts also.
Only the Switch part is broken, where you can assign a unique IP to the PC session. It still works if you want to share your Mac's IP address so networking still works and so does internet access. Since I only use VPC to test IE6 bugs, that's enough for me. If you've got more specialist windows only networking apps that may be a problem.
It broke because Apple changed the networking kernel extension and VPC wedged in to that. It just needs Microsoft to do a small change to fix it.
In 10.4 they changed the way most things interacted with the kernel so this kind of problem shouldn't happen again. There's probably similar issues with things like Norton, some USB devices and various VPN clients which wedged in as kexts also.
What Apple has done is actually a big improvement. This way, changes to the kernel in the future won't cause problems for developers. We should be seeing less 3rd party updates after a new upgrade comes out.
[B]Anyone having trouble with peripherals (particularly scanners)? We added Tiger to two of eight machines at our organization and an Epson device is now no longer functioning (with Tiger).
Actually, I'd not tried my Epson 2580 scanner since the upgrade but now I've noticed the Epson Scan software no longer works. It was very flaky before anyway and was constantly hanging, needing a reboot. However, pressing the scan button now opens Image Capture which seems to be much better than it was and more stable than the old Epson software anyway!
Quote:
Originally posted by Sybaritic
[B]Also, any word on when Norton Antivirus will be updated to work with 10.4?
Why?
a) There are no viruses other than PC ones.
b) ClamAV is updated quicker than Norton and is free and doesn't rely on terrible kernel extensions to wedge it's way in to your system.
Quote:
Originally posted by Sybaritic
My sense is that 10.3.9 is solid as a rock. I'm waiting for a while with my machine at home before upgrading there. The present idiosyncrasies are not worth the trouble, Tiger's compelling features notwithstanding.
I miss 10.3.9's Mail and DivX movie playback but that's about all.
b) ClamAV is updated quicker than Norton and is free and doesn't rely on terrible kernel extensions to wedge it's way in to your system.
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
What Apple has done is actually a big improvement. This way, changes to the kernel in the future won't cause problems for developers. We should be seeing less 3rd party updates after a new upgrade comes out.
Strong agreement. The new KPIs are (finally) the official way for apps and services to hook into the kernel. Previously, apps had to rather wing it, and while most old approaches still work (but are firmly and 100% deprecated), some don't. Fortunately, Apple warned developers about this since 10.1, that the APIs would likely change, and not to rely on them, and then gave them the KPIs in the various builds of Tiger this past year. Some are just tardy to the party.
It never really made sense to polish up a KPI suite on the old kernel, what with the lack of fine-grain locks, minimal threading, and such.
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
Why not? They passed them to me first...
Agreed though - if you're in a mixed environment where you are likely to be getting infected PC files, and passing them on, then it's a social nicety... but that's all.
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
And how exactly are you going to pass on a PC virus from a Mac when the PC virus doesn't run on the Mac?
I suppose you could email the virus yourself manually after downloading it from a PC or Kazaa embedded in another PC application but that would be pretty dumb just emailing on PC applications you've not tried yourself to your friends. And Norton won't stop you from doing that anyway.
Agreed though - if you're in a mixed environment where you are likely to be getting infected PC files, and passing them on, then it's a social nicety... but that's all.
mac os 10.4.1 should have a database of PC viruses that it spreads to PC users that way mac os 10.4.1 becomes a big PC virus in its own way but in a very subtle way because it cripples PCs and makes people more and more frustrated with their PC rubbish... think about it... several million 10.4.1 machines on the 'net, spreading the joy of PC viruses
Comments
Originally posted by wnurse
Hopefully, he does not develop software for NASA or a nuclear plant or an airplane.
Ever read a software license carefully? You just mentioned two items that specifically are mentioned in *MOST* licenses as areas where the software is NOT TO BE USED: nuclear power plants, and flight systems.
As for space exploration... you don't recall the metric/English fiasco recently with the Mars missions? How about the Ariane 5 that had to be destroyed due to a sign error in the navigation systems? You *betcha* they have bugs... and it's quite spectacular when they do.
I understand your point, but sorry, it doesn't hold water. The fact is that every commercial provider of consumer software ships product with known bugs. Always have, always will, IMO.
Which brings me to my next point: melgross, Adobe may have shipped software with no bugs that *you* were aware of, but trust me, internally they were tracking them. 'No known bugs' is a fairy tale told to beta testers and early adopters.
Originally posted by TenoBell
So you are saying that Windows is always released with every known bug fixed?
At some point you have to put it out there, and fix some things later.
As far as developers complaining. What I've heard and what was said on Ars Technica is that Apple has always said the way the kernal was set up was not permanent and would change. Many are confused as to why some developers were not prepared.
Changes in an OS that invalidates a third party's (or even in this case Apple's) software is fine if Apple stated that those functions, addresses, pointers, etc. couldn't be relied on for future use. This happens all the time. I couldn't blame Apple for that.
But that's not what I'm talking about.
And again, MS is one of the worst examples for anyone to use. They are notorious for lacking the ability to produce stable software.
Originally posted by Voxapps
I've been doing software and hardware product management for over 15 years, so I think I can add a bit to this discussion from the "corporate" perspective. In software development, there are usually two high-level questions: 1) when is the ship date, and 2) how many bugs of what severity level are acceptable for a released product. Nobody likes to ship buggy products, but the reality is that, at some point, you've got to say "ship it!"
Many software and hardware companies have moved to the "train leaving the station" model: the software will ship on (insert date here), and it will either be ready to ship by then or troublesome features will be removed to make it ready to ship (i.e., they will "miss the train"). The "train" model ensures that *something* is released on time to meet the expectations of customers, stockholders, the media, etc., and it also ensures that development resources get freed-up for other projects. For public companies, shipping late can have disastrous consequences for stock prices and can put a company at a competitive disadvantage if another company gets their product to market first.
Most product bugs are ranked by severity (or priority): P1 bugs prevent a product from shipping: they affect the primary function of the product for many users and cannot be worked around. Products should never ship with any P1 bugs. P2 bugs are somewhat less severe or have work-arounds; a small number of P2s are present in most shipping software. Priorities often go as deep as 9 levels (a P9 is often related to documentation or minor appearance flaws that have no affect on performance or usability, and is classified as an "enhancement").
I have never heard of any software product shipping without any bugs at all.
Beta testing is really hard for corporate developers. You generally want to keep betas under some control (i.e., not available to just anyone) to allow a bit of surprise to customers, to prevent late betas from being distributed illegally, and to allow the project managers to keep track of test results. Unfortunately, there are so many permutations of hardware and software, that it is literally impossible to develop test cases that cover everything. A problem is that 500 public beta testers may essentially test the same things, and none may hit the critical sequence that triggers a problem. Also, while many beta testers have good intentions, it is often true that they - as a group - don't do the thorough testing expected by the developers.
Some companies I've worked with are even considering paid beta programs: you get the software and a list of things to test; when you submit documentation on schedule that shows test results for specific tests, you get paid. Otherwise, you get a "thank you" and no payment.
I don't plan to upgrade to Tiger until at least 10.4.1, but I think anyone who's using 10.4.0 should be getting a fairly solid release, especially when used on unmodified hardware running common apps.
I agree with you to a large degree. The problem is that we have to look at what bugs are getting out. As I've said, bugs that don't affect function are not always bothered with.
I've also said the same thing about beta testing. you can't find all of the bugs because you can't test for everything. I fill out online forms when I do testing. It's just not the joy of getting a program before everyone else gets it . It's a lot of work. A lot of so called "testers" drop out when they relize that.
Also in-house development tends to be buggy. More than half of all in-house projects fail, and a large number of those that get to completion are never completely validated.
When a software company asks you to test a specific area, they want it tested to completion, or as close to that as is possible. Any bugs that are founf are squashed. As I've said, if the area had insurmountable problems, either development is extended, or the area is re-evaluated. Possibly turned off, or removed.
There do seem to be too many problems with this release though.
Originally posted by Kickaha
Ever read a software license carefully? You just mentioned two items that specifically are mentioned in *MOST* licenses as areas where the software is NOT TO BE USED: nuclear power plants, and flight systems.
As for space exploration... you don't recall the metric/English fiasco recently with the Mars missions? How about the Ariane 5 that had to be destroyed due to a sign error in the navigation systems? You *betcha* they have bugs... and it's quite spectacular when they do.
I understand your point, but sorry, it doesn't hold water. The fact is that every commercial provider of consumer software ships product with known bugs. Always have, always will, IMO.
Which brings me to my next point: melgross, Adobe may have shipped software with no bugs that *you* were aware of, but trust me, internally they were tracking them. 'No known bugs' is a fairy tale told to beta testers and early adopters.
Legalese. Windows NT is being used on out Ageis crusiers even it's failed several times, and they had to be towed into port. The license states that it shouldn't be used for "mission Critical support where lives may depend on it"! Whoa!
Oh please.
You think that NASA knew that they were using both inch and metric units and said "what the hell, let's do this anyway"?
You also think that the ESA knew that they had a sign error, but thought that it was only a plus or minus, so it wasn't important?
You don't think that it's possible that they DIDN'T know about those errors?
I don't know about fairy tales. Adobe doesn't officially tell us that there are no known bugs, or that there are lots of them. All I know is that when I speak to John, he says that we got 'em all. Remember, that means all that we know about.
It is continuous while they are still in pre-release. That's what pre-release development is. It can't be done any other way.
But beta testers aren't supposed to see any bugs in the final stage. I'm sorry, but that's just the way it is.
Interesting. But I have a question. What's the point of beta testing if there are no bugs in the software. I was under the impression that the whole point of beta-testing was to find the bugs in the software.
Apple used to be known for very stable and bug free OS releases. Like it or not, this release is considered, by those developers you mentioned, to be the buggiest release since 10.0.
They were? Could you tell me when? For I remember all the problems with System 7, System 7.1, 7.5.1, 8, 9, 10.0, 10.1, 10.2, 10.3...
Systems 8 and 9 were semi-solid (but still had huge problems), but recall these were basically just minor updates to system 7. And since all of the 'classic' OSes were just one bad program from total system lockup, I have a hard time calling that system "solid".
I work on a software project right now and we never release any software with a known bug. If it's known, it's fixed. I was kinda amused reading about all companies releasing software with known bugs.
Must be nice to work on a project where you have control over when software is released. But when software is released is dependent on so many factors, trying to generalize with "We never do this, so no one else should either" is just a farce. I work on a software project that's several years (and several versions) old, and I know it has bugs in it. But who's paying me the month's worth of time to fix a bug that maybe three user's might experience over the course of a couple of years? Who's going to tell management that we need to delay the product another couple of months to fix all the irritations, which causes issues with marketing, sales, purchasing, etc?
Oh, and finally, public betas sound like great ideas, but generally, the public are lousy testers. Sure, they can install the software, then complain that this or that doesn't work. Getting bug reports from people is easy. Getting decent bug reports, with info on how it occurred, whether you can recreate it, what hardware/software you have installed that might be interacting with it, etc, that's nearly impossible to get from the general public. I've literally received bugs that read "Received error xxx when running program". Not even an indication of where in the software they were, let alone what they did, what they tried to do, nothing. Which is just a little worse then the users who say they were doing one thing, but actually were doing something else. Or think they know what they were doing. Yeah, I'm sure Apple needs even MORE of those type of reports.
Originally posted by Tomahawk
I can pretty much promise you Adobe lets software out with known bugs...
We've worked with Adobe pretty closely on a variety of issues we have had with their software (we're a rather large unit). Adobe has even sent techs to our site to see first hand what our issues are.
So far, we still see issues we reported early in the previous version of Creative Suite in the new one. We also have found issues in the first two days of using the new version that we may consider show stoppers until fixed.
This tells me either Adobe is doing a pretty crappy job of testing different environments or they realize they have to make a buck or two too and can fix some problems later.
I also agree Adobe simply turns off a variety of features that they can't get working well. It's a double edge sword. How often do you suppose they consider all the "planned" updates they cut when they set their pricing?
I use CS, but I only test Photoshop, so I can't speak to the rest.
But Adobe has a very good rep. about their software being bug free and stable. Is it all perfect, I doubt it.
I've found problems with software that was cause by other problems.
Look, for example, at Macfixit, or Macintouch. A lot of problems are reported there. I have had very few of them.
So where are those problems coming from, and who is responsible for them?
Many people, and even large organizations, set their systems up in a way that just begs problems to appear. Some problems are The OS vendors fault for not making things clear enough. Sometimes it's the 3rd party vendors. Sometimes it's the IT department. Often it's a combination of the above.
It's even the hardware. Marginal memory causes many problems that seem to be software related, etc.
All these combinations can't be tested for, and so incompatibilities escape. I find it hard to blame anyone for many of these.
But some combinations are inexcusably not tested.
Apple's Firewire data, and drive loss was one of those. It survived several updates. I hope it doesn't return in 10.4.
Originally posted by melgross
Oh please.
You think that NASA knew that they were using both inch and metric units and said "what the hell, let's do this anyway"?
You also think that the ESA knew that they had a sign error, but thought that it was only a plus or minus, so it wasn't important?
You don't think that it's possible that they DIDN'T know about those errors?
Well obviously they didn't know about them, but that wasn't wnurse's assertion. wnurse made a much stronger assertion that simply isn't true.
I don't know about fairy tales. Adobe doesn't officially tell us that there are no known bugs, or that there are lots of them. All I know is that when I speak to John, he says that we got 'em all. Remember, that means all that we know about.
Or that are required to ship.
Look, we can argue this till Doomsday, and we're not going to sway each other, so why don't we just agree to disagree?
Originally posted by Louzer
Sorry, couple of points I thought were hillarious
Interesting. But I have a question. What's the point of beta testing if there are no bugs in the software. I was under the impression that the whole point of beta-testing was to find the bugs in the software.
They were? Could you tell me when? For I remember all the problems with System 7, System 7.1, 7.5.1, 8, 9, 10.0, 10.1, 10.2, 10.3...
Systems 8 and 9 were semi-solid (but still had huge problems), but recall these were basically just minor updates to system 7. And since all of the 'classic' OSes were just one bad program from total system lockup, I have a hard time calling that system "solid".
Must be nice to work on a project where you have control over when software is released. But when software is released is dependent on so many factors, trying to generalize with "We never do this, so no one else should either" is just a farce. I work on a software project that's several years (and several versions) old, and I know it has bugs in it. But who's paying me the month's worth of time to fix a bug that maybe three user's might experience over the course of a couple of years? Who's going to tell management that we need to delay the product another couple of months to fix all the irritations, which causes issues with marketing, sales, purchasing, etc?
Oh, and finally, public betas sound like great ideas, but generally, the public are lousy testers. Sure, they can install the software, then complain that this or that doesn't work. Getting bug reports from people is easy. Getting decent bug reports, with info on how it occurred, whether you can recreate it, what hardware/software you have installed that might be interacting with it, etc, that's nearly impossible to get from the general public. I've literally received bugs that read "Received error xxx when running program". Not even an indication of where in the software they were, let alone what they did, what they tried to do, nothing. Which is just a little worse then the users who say they were doing one thing, but actually were doing something else. Or think they know what they were doing. Yeah, I'm sure Apple needs even MORE of those type of reports.
I think you missed the point on that first one. The key is "final". Re-read it.
If you compare Apple's old system software to MS's, you will find that indeed it was stable. 8 and 9 were a lot more than just updates to 7. 8.5 was very solid, and so was 9+.
Yes, I've said that 10.0 was a problem. It's been described as Apple's second (paid) public beta of X. 10.1 should have been the first release.
Sure, the public is lousy. But Apple has bug reporting in it's software right now. Also, if thousands of people report the same problem, and Apple didn't see it, like the VERY serious Firewire bug, don't you think they would look into it?
1. I am a software developer
2. I was once a network administrator.
3. I once built a computer from the ground up.
4. I know myself around computers. I'm not some little old lady who doesnt know a keyboard from a mouse.
5. The customer service person must think i am an idiot.
6. Who the hell is this person telling me that the error message on screen does not exist. I'm looking at it!!!.
Nevertheless, i'm not a version X.0 adopter. I usually let the fanatics be the beta testers and then get version X.1 or maybe even wait for version X.2.
I think consumers as a whole should always reject version X.0 software. That would force all companies (apple included) to create quality products. This disturbing trend of using customers as beta testers must be reversed.
Also, any word on when Norton Antivirus will be updated to work with 10.4?
My sense is that 10.3.9 is solid as a rock. I'm waiting for a while with my machine at home before upgrading there. The present idiosyncrasies are not worth the trouble, Tiger's compelling features notwithstanding.
anyone running anything mission critical who primarily uses a mac will wait till about 10.4.2 before making the jump
Originally posted by cubist
I've also read (but not been able to test yet) that VPC networking is broken, and that there are PPTP problems too. I don't see any mention of those in this update :-/
Only the Switch part is broken, where you can assign a unique IP to the PC session. It still works if you want to share your Mac's IP address so networking still works and so does internet access. Since I only use VPC to test IE6 bugs, that's enough for me. If you've got more specialist windows only networking apps that may be a problem.
It broke because Apple changed the networking kernel extension and VPC wedged in to that. It just needs Microsoft to do a small change to fix it.
In 10.4 they changed the way most things interacted with the kernel so this kind of problem shouldn't happen again. There's probably similar issues with things like Norton, some USB devices and various VPN clients which wedged in as kexts also.
Originally posted by aegisdesign
Only the Switch part is broken, where you can assign a unique IP to the PC session. It still works if you want to share your Mac's IP address so networking still works and so does internet access. Since I only use VPC to test IE6 bugs, that's enough for me. If you've got more specialist windows only networking apps that may be a problem.
It broke because Apple changed the networking kernel extension and VPC wedged in to that. It just needs Microsoft to do a small change to fix it.
In 10.4 they changed the way most things interacted with the kernel so this kind of problem shouldn't happen again. There's probably similar issues with things like Norton, some USB devices and various VPN clients which wedged in as kexts also.
What Apple has done is actually a big improvement. This way, changes to the kernel in the future won't cause problems for developers. We should be seeing less 3rd party updates after a new upgrade comes out.
Originally posted by Sybaritic
[B]Anyone having trouble with peripherals (particularly scanners)? We added Tiger to two of eight machines at our organization and an Epson device is now no longer functioning (with Tiger).
Actually, I'd not tried my Epson 2580 scanner since the upgrade but now I've noticed the Epson Scan software no longer works. It was very flaky before anyway and was constantly hanging, needing a reboot. However, pressing the scan button now opens Image Capture which seems to be much better than it was and more stable than the old Epson software anyway!
Originally posted by Sybaritic
[B]Also, any word on when Norton Antivirus will be updated to work with 10.4?
Why?
a) There are no viruses other than PC ones.
b) ClamAV is updated quicker than Norton and is free and doesn't rely on terrible kernel extensions to wedge it's way in to your system.
Originally posted by Sybaritic
My sense is that 10.3.9 is solid as a rock. I'm waiting for a while with my machine at home before upgrading there. The present idiosyncrasies are not worth the trouble, Tiger's compelling features notwithstanding.
I miss 10.3.9's Mail and DivX movie playback but that's about all.
Originally posted by aegisdesign
a) There are no viruses other than PC ones.
b) ClamAV is updated quicker than Norton and is free and doesn't rely on terrible kernel extensions to wedge it's way in to your system.
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
Originally posted by melgross
What Apple has done is actually a big improvement. This way, changes to the kernel in the future won't cause problems for developers. We should be seeing less 3rd party updates after a new upgrade comes out.
Strong agreement. The new KPIs are (finally) the official way for apps and services to hook into the kernel. Previously, apps had to rather wing it, and while most old approaches still work (but are firmly and 100% deprecated), some don't. Fortunately, Apple warned developers about this since 10.1, that the APIs would likely change, and not to rely on them, and then gave them the KPIs in the various builds of Tiger this past year. Some are just tardy to the party.
It never really made sense to polish up a KPI suite on the old kernel, what with the lack of fine-grain locks, minimal threading, and such.
Originally posted by melgross
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
Why not? They passed them to me first...
Agreed though - if you're in a mixed environment where you are likely to be getting infected PC files, and passing them on, then it's a social nicety... but that's all.
Originally posted by melgross
I find Norton's to work quite well. I have no complaints.
The point to anti-virus software for the Mac up to this point, at least, isn't so much for Mac virii, but rather for catching PC ones.
I don't know about you and others, but I, as well as many other people, feel that if you have PC using friends and colleagues (you do don't you?), it's not nice to pass virii on to them.
And how exactly are you going to pass on a PC virus from a Mac when the PC virus doesn't run on the Mac?
I suppose you could email the virus yourself manually after downloading it from a PC or Kazaa embedded in another PC application but that would be pretty dumb just emailing on PC applications you've not tried yourself to your friends. And Norton won't stop you from doing that anyway.
btw plural of virus is viruses
Originally posted by aegisdesign
And how exactly are you going to pass on a PC virus from a Mac when the PC virus doesn't run on the Mac?
Forwarding (manually) a mail you just received... or sharing (infected) files...
Originally posted by Kickaha
Why not? They passed them to me first...
Agreed though - if you're in a mixed environment where you are likely to be getting infected PC files, and passing them on, then it's a social nicety... but that's all.
mac os 10.4.1 should have a database of PC viruses that it spreads to PC users
damn thats evil man.... bwah ha ha ha ha ha ha ha