klausmogensen
Don, I like what you create, but I believe you are wrong about how to test and bring out new versions of software products. As you probably know I've also been a software developer for 2-3 decades. And I think the way you describe testing is 20 years in the past. We don't even hire those kind of testers anymore in my day-job company
Testing is now mostly done automatically/scripted/by "software robots"
Testers (which are mostly also developers to some degree) build the scripts
Every build is automatically tested (unit tests, integrations tests and so on), and if those "goes to green", it's tested by developers. Then it's an alpha version, which may get more testing if it's a major change
If everything is ok it gets promoted to beta, and goes to the beta team. Those are super users
These steps are super fast for smaller changes / bug fixes - often 1-2 days from code change to beta
If all is well it goes to public beta/early release after days weeks of beta testing, which is for all users that want to spend time testing
If it survives that - and most builds don't, because users who likes to test will find all the "nice to have" and "1 in a million" kind of bugs - then it gets promoted to a public release
That is how modern software development is done, and should be done in my book
Because that means that you can have new releases once per month, which have only small increments to the software, instead of big bang releases once a year, where all hell sometimes break loose if there is a bug, and you know a fix is 3-6 months away
And (without me being an Atomix employee) that is how it seems to me VDJ is created and enhanced
(I'm in the beta tester team, so that is where I see the changes for the first time)
As far as I know even Serato - the old and tired elephant in the room - tries to do it this way now
Could there be more beta testers? Sure, there always could
Could there be a more controlled way of getting the public beta/early access releases tested by users? Sure, there always could
But I really do think the release model Atomix is using is the right one to use for software products in 2022
Testing is now mostly done automatically/scripted/by "software robots"
Testers (which are mostly also developers to some degree) build the scripts
Every build is automatically tested (unit tests, integrations tests and so on), and if those "goes to green", it's tested by developers. Then it's an alpha version, which may get more testing if it's a major change
If everything is ok it gets promoted to beta, and goes to the beta team. Those are super users
These steps are super fast for smaller changes / bug fixes - often 1-2 days from code change to beta
If all is well it goes to public beta/early release after days weeks of beta testing, which is for all users that want to spend time testing
If it survives that - and most builds don't, because users who likes to test will find all the "nice to have" and "1 in a million" kind of bugs - then it gets promoted to a public release
That is how modern software development is done, and should be done in my book
Because that means that you can have new releases once per month, which have only small increments to the software, instead of big bang releases once a year, where all hell sometimes break loose if there is a bug, and you know a fix is 3-6 months away
And (without me being an Atomix employee) that is how it seems to me VDJ is created and enhanced
(I'm in the beta tester team, so that is where I see the changes for the first time)
As far as I know even Serato - the old and tired elephant in the room - tries to do it this way now
Could there be more beta testers? Sure, there always could
Could there be a more controlled way of getting the public beta/early access releases tested by users? Sure, there always could
But I really do think the release model Atomix is using is the right one to use for software products in 2022
Posted Sat 22 Jan 22 @ 11:41 am
Don Moir
Klaus you earn respect because in general you do a pretty good job with the help videos and otherwise.
Also I don't say anything that is earth shaking to people who know about such things or that I can't back up. Pretty much common sense. I don't say anything that is not clearly a problem. I hope I have never ever given anyone any reason for me to make anything up. But they like to dismiss or discredit me as nonsense rather than do the obvious thing and ask me. Some users actually think I make this stuff up because the end users believe in you even though some times because of BS.
Larger and smaller companies that care and know enough still have testers because it is the best way to do it. It's all logged for reference in detail and presented to programmers to fix or comment on. For one company I worked for which started with about 10 people, one guy (just one) became responsible for testing. He did an amazing job and saved so much time and irritation. The company grew to over 300 people and the test group grew to about 10 people. They are well worth the money spent as it saves money, time, and more. It only takes someone who knows how a program is supposed to work and he can become skilled in time.
If the way you described actually worked then I would not be saying anything about bugs in the software. It's not like I just up and made it up. Like I said no one even bothers to ask me to show them the bugs I refer to. They don't address simple things that would take less then an hour to fix. Instead they are passed on as user error or similar. Just one easy case. When on external drive and database gets corrupted issue is blamed on user not ejecting the drive. I mentioned they should flush the file before closing file and that should fix. One function call less than 5 minutes to do. I wonder if they even do that. But it needs to be more robust then that but still very easy to make 1000 percent more secure. End users are not going to catch a lot of things. Some here think if they can't reproduce a problem then it must be the other guys fault. One problem took at least 6 months for you all to fix which was entirely your own fault. Lack of testing. Users began converting in some cases many thousands of files. I helped and it finally got fixed and called names because of. Truly I don't care about that. Keep in mind VDJ 8 was in bad shape and was tested I guess with your beta testers. One day it was declared that is not a beta any more and all bugs have been fixed. I still call it a beta not for my benefit but that's the way I see it as would many others.
Final testing is not usually left up to automation or developers. Developers always test a lot but focused on small details and better to have someone focused entirely on testing. I also use automation to test some things. I have some automatic code that will run through many thousands of videos. I have a collection of the worst or problem videos that I could find. Good videos pretty still can present problems but they pretty much take care of themselves. Automation for video testing will catch hangs and crashes but not many things. Video code is done by an outside group. It is large and complex and changes a lot. Every time I got new video code I tested it for about a month before releasing it.
A billion beta testers that don't know much about testing would still not work. It's not their thing to know. Besides why would you need more beta testers because like you say: "Testing is now mostly done automatically/scripted/by "software robots"" Some is but lots can't be done that way. Beta testers are mostly for usability testing. End users should never be counted on for real testing period. It would only take one guy who knows what he is doing to start real testing. He would have plenty to do initially. But then there must also be a desire to fix things. He might be diminished with BS from above. With the amount of time spent blaming other things you could have fixed things many times over. Releases that throw stuff up in your face (skin video or whatever), things that are removed that have no alternative that have been there for a long time, defaults that are changed to force things... I don't know how many had to ask how to turn that off. Well if user had turned it on he might know how to turn it off. If you want to show case these things it's easy. Have a message come up to turn them on if user has time to play. Sometimes these changes are not even mentioned in the release notes. Some times things have been changed foolishly and several complained. That's why I asked about the dedicated audio setting. We never know what you might change and things have been changed before that should have not. You never know. But I can't believe this is done to showcase features because no business in their right mind would do that with monthly releases.
It's true that a lot of software that is called professional is become more buggy and amateurish.
Why would I bother if it was one in a million? Things that don't work or have unpredictable behavior are bugs you know. No buddy knows how many bugs are in VDJ. There are most likely still many. To be honest I have not looked in awhile. Whats the point I think as not worth the time reporting them. Since no one ask me about it I will take the time to point out some problems that are obvious to me. It won't be a complete test because that would take too much time. Then you can decide about it yourself.
Of course small incremental releases have nothing at all to do with real testing. It was one of these incremental releases that caused certain videos to stop working. Many complained. It was partially fixed in a couple months but not completely. I stayed persistent with it and identified the reason. About 6 months later now from initial problem which had always worked before. The peanut galley making all kinds of excuses. When you all could reproduce the problem with the details I gave you (not elaborate at all) it was blamed on something else. No man it's your frekin fault. Like I said when the hell is it safe? no idea
You mention Serato. I have never been a fan. It's a far simpler program that does not change much over time. VDJ can change radically and you never know. A lot people only use the program when live. You must know people are going to use the program no matter what you say. But I ask when it is safe. No one knows and that is the truth. Things pop up and the blame goes elsewhere.
Also I don't say anything that is earth shaking to people who know about such things or that I can't back up. Pretty much common sense. I don't say anything that is not clearly a problem. I hope I have never ever given anyone any reason for me to make anything up. But they like to dismiss or discredit me as nonsense rather than do the obvious thing and ask me. Some users actually think I make this stuff up because the end users believe in you even though some times because of BS.
klausmogensen wrote :
Don, I like what you create, but I believe you are wrong about how to test and bring out new versions of software products. As you probably know I've also been a software developer for 2-3 decades. And I think the way you describe testing is 20 years in the past. We don't even hire those kind of testers anymore in my day-job company
Testing is now mostly done automatically/scripted/by "software robots"
Testers (which are mostly also developers to some degree) build the scripts
Testing is now mostly done automatically/scripted/by "software robots"
Testers (which are mostly also developers to some degree) build the scripts
Larger and smaller companies that care and know enough still have testers because it is the best way to do it. It's all logged for reference in detail and presented to programmers to fix or comment on. For one company I worked for which started with about 10 people, one guy (just one) became responsible for testing. He did an amazing job and saved so much time and irritation. The company grew to over 300 people and the test group grew to about 10 people. They are well worth the money spent as it saves money, time, and more. It only takes someone who knows how a program is supposed to work and he can become skilled in time.
If the way you described actually worked then I would not be saying anything about bugs in the software. It's not like I just up and made it up. Like I said no one even bothers to ask me to show them the bugs I refer to. They don't address simple things that would take less then an hour to fix. Instead they are passed on as user error or similar. Just one easy case. When on external drive and database gets corrupted issue is blamed on user not ejecting the drive. I mentioned they should flush the file before closing file and that should fix. One function call less than 5 minutes to do. I wonder if they even do that. But it needs to be more robust then that but still very easy to make 1000 percent more secure. End users are not going to catch a lot of things. Some here think if they can't reproduce a problem then it must be the other guys fault. One problem took at least 6 months for you all to fix which was entirely your own fault. Lack of testing. Users began converting in some cases many thousands of files. I helped and it finally got fixed and called names because of. Truly I don't care about that. Keep in mind VDJ 8 was in bad shape and was tested I guess with your beta testers. One day it was declared that is not a beta any more and all bugs have been fixed. I still call it a beta not for my benefit but that's the way I see it as would many others.
Final testing is not usually left up to automation or developers. Developers always test a lot but focused on small details and better to have someone focused entirely on testing. I also use automation to test some things. I have some automatic code that will run through many thousands of videos. I have a collection of the worst or problem videos that I could find. Good videos pretty still can present problems but they pretty much take care of themselves. Automation for video testing will catch hangs and crashes but not many things. Video code is done by an outside group. It is large and complex and changes a lot. Every time I got new video code I tested it for about a month before releasing it.
klausmogensen wrote :
Could there be more beta testers? Sure, there always could
Could there be a more controlled way of getting the public beta/early access releases tested by users? Sure, there always could
Could there be a more controlled way of getting the public beta/early access releases tested by users? Sure, there always could
A billion beta testers that don't know much about testing would still not work. It's not their thing to know. Besides why would you need more beta testers because like you say: "Testing is now mostly done automatically/scripted/by "software robots"" Some is but lots can't be done that way. Beta testers are mostly for usability testing. End users should never be counted on for real testing period. It would only take one guy who knows what he is doing to start real testing. He would have plenty to do initially. But then there must also be a desire to fix things. He might be diminished with BS from above. With the amount of time spent blaming other things you could have fixed things many times over. Releases that throw stuff up in your face (skin video or whatever), things that are removed that have no alternative that have been there for a long time, defaults that are changed to force things... I don't know how many had to ask how to turn that off. Well if user had turned it on he might know how to turn it off. If you want to show case these things it's easy. Have a message come up to turn them on if user has time to play. Sometimes these changes are not even mentioned in the release notes. Some times things have been changed foolishly and several complained. That's why I asked about the dedicated audio setting. We never know what you might change and things have been changed before that should have not. You never know. But I can't believe this is done to showcase features because no business in their right mind would do that with monthly releases.
It's true that a lot of software that is called professional is become more buggy and amateurish.
klausmogensen wrote :
"1 in a million" kind of bugs - then it gets promoted to a public release
Why would I bother if it was one in a million? Things that don't work or have unpredictable behavior are bugs you know. No buddy knows how many bugs are in VDJ. There are most likely still many. To be honest I have not looked in awhile. Whats the point I think as not worth the time reporting them. Since no one ask me about it I will take the time to point out some problems that are obvious to me. It won't be a complete test because that would take too much time. Then you can decide about it yourself.
klausmogensen wrote :
That is how modern software development is done, and should be done in my book
Because that means that you can have new releases once per month, which have only small increments to the software, instead of big bang releases once a year, where all hell sometimes break loose if there is a bug, and you know a fix is 3-6 months away
Because that means that you can have new releases once per month, which have only small increments to the software, instead of big bang releases once a year, where all hell sometimes break loose if there is a bug, and you know a fix is 3-6 months away
Of course small incremental releases have nothing at all to do with real testing. It was one of these incremental releases that caused certain videos to stop working. Many complained. It was partially fixed in a couple months but not completely. I stayed persistent with it and identified the reason. About 6 months later now from initial problem which had always worked before. The peanut galley making all kinds of excuses. When you all could reproduce the problem with the details I gave you (not elaborate at all) it was blamed on something else. No man it's your frekin fault. Like I said when the hell is it safe? no idea
You mention Serato. I have never been a fan. It's a far simpler program that does not change much over time. VDJ can change radically and you never know. A lot people only use the program when live. You must know people are going to use the program no matter what you say. But I ask when it is safe. No one knows and that is the truth. Things pop up and the blame goes elsewhere.
Posted Thu 27 Jan 22 @ 1:44 am
Don Moir
screw up please remove
Posted Thu 27 Jan 22 @ 1:46 am
Citras Flava
Out of the 5 months I have been using this software I can't get anything after build6732 to run more than an hour before it requires a restart. Some versions just flat crash while other versions have odd audio issues that no troubleshoot has fixed except for using older builds. I regularly bench test my pc to ensure I'm still up to optimization and build 6732 doesn't give me set ending issues. I try to take software with a pick your poison mindset. Whatever gives you the best outcome is what you want.
Posted Sat 29 Jan 22 @ 5:23 am
Adion
According to the crash reports the issue seems to be with the beatport plugin.
Try to remove beatport7_vdj_x64.dll from Documents\VirtualDJ\Plugins64
Try to remove beatport7_vdj_x64.dll from Documents\VirtualDJ\Plugins64
Posted Sat 29 Jan 22 @ 6:33 am
Citras Flava
Adion wrote :
According to the crash reports the issue seems to be with the beatport plugin.
Try to remove beatport7_vdj_x64.dll from Documents\VirtualDJ\Plugins64
Try to remove beatport7_vdj_x64.dll from Documents\VirtualDJ\Plugins64
Hey thanks I just updated and gave this a shot and it's working out great,
Posted Sat 29 Jan 22 @ 7:00 pm