Idea: Add a stability mode option to next RC

[Hagbard]

I'm wondering if any other Parallels users think this would be a good idea: add an option to future versions of Parallels to disable many of the features introduced in the post-1970 builds and run in an "enhanced stability/performance" mode. There seems to be a persistent debate on the boards between users who want more and more features added (and love coherence and easier drag+drop) and users who want to maximize stability, performance, and security.

I think Parallels could maximize user satisfaction if it added an option to run Parallels in "Build 1970 mode" -- so users who wanted to see if coherence, etc, worked well for them could test those features without having to uninstall and downgrade if they experienced unacceptable stability issues. The size of the Parallels program itself is fairly trivial relative to the system resources needed by virtual machines, so there is no real penalty for rolling the build 1970 code into the ongoing builds.

Perhaps this is a terrible idea, and I suppose it does fall into the category of a kludge, but in the real world, a few good kludges can save an incredible amount of time and effort. In fact, with a bit of very simple scripting, this could be turned into a selling point: suppose when Parallels was launched, it looked at its logs to see if it had experienced a crash, and if so, offered the user the option to deactivate nonessential features.

I assume Parallels receives an unbelievable amount of comments/bug reports already, so I wanted to see if any other users would appreciate this feature. If it seems like a good ide to anyone else, I'll send the idea into the team.

 

[drval]

I'm wondering if any other Parallels users think this would be a good idea: add an option to future versions of Parallels to disable many of the features introduced in the post-1970 builds and run in an "enhanced stability/performance" mode. There seems to be a persistent debate on the boards between users who want more and more features added (and love coherence and easier drag+drop) and users who want to maximize stability, performance, and security.

I think Parallels could maximize user satisfaction if it added an option to run Parallels in "Build 1970 mode" -- so users who wanted to see if coherence, etc, worked well for them could test those features without having to uninstall and downgrade if they experienced unacceptable stability issues. The size of the Parallels program itself is fairly trivial relative to the system resources needed by virtual machines, so there is no real penalty for rolling the build 1970 code into the ongoing builds.

Perhaps this is a terrible idea, and I suppose it does fall into the category of a kludge, but in the real world, a few good kludges can save an incredible amount of time and effort. In fact, with a bit of very simple scripting, this could be turned into a selling point: suppose when Parallels was launched, it looked at its logs to see if it had experienced a crash, and if so, offered the user the option to deactivate nonessential features.

I assume Parallels receives an unbelievable amount of comments/bug reports already, so I wanted to see if any other users would appreciate this feature. If it seems like a good ide to anyone else, I'll send the idea into the team.

FWIW, and if I'm following you, this already exists in the form of version 1970. Just install and use that version if you want a "stable" version without any of the other features. My strong sense -- and hope! -- is that the next official release of Parallels will be as stable as 1970 - if not more so -- and will include all of these features. That is the way that the development process should work IMO.

 

[joem]

I think making feature optional is always a good idea, and making the defaults the safest, disabling potentially dangerous features is good practice. I think some of the functionality of 1970 needs improving, particularly USB support, and hopefully firewire support as well, but these seem to me to be core features, not extras.

I certainly won't be enabling "coherence", and would love to see drag and drop implemented safely without exposing the host filesystem to guest programs.

Some folks see Parallels as a way to integrate a Windows app or two into the Mac, and some (like me) really want to run Windows (or another OS) in a completely isolated sandbox so we can deliberately run malware to analyze it, knowing that the host OS is safe.

 

[joem]

FWIW, and if I'm following you, this already exists in the form of version 1970. Just install and use that version if you want a "stable" version without any of the other features. My strong sense -- and hope! -- is that the next official release of Parallels will be as stable as 1970 - if not more so -- and will include all of these features. That is the way that the development process should work IMO.

That really isn't a solution because some 1970 features are not completely working. Disabling potentially unstable or dangerous features, preferably by default, is, IMNSHO, the way to go. This is based on many years of experience in security.

Microsoft has for years had the attitude that features should trump safety, and the result is the mess we are in today with millions of compromised machines. They have changed their attitude, and Vista now errs in the side of safety. They did this because of pressure to fix their lack of security. Going backwards seems like a really bad idea.

I think the idea that the default configuration should be what "most people" want, is a really, really bad idea. Most people don't bother turning on the Windows firewall (a truly stupid mistake). Most people aren't technical enough to make these decisions, and satisfying most people is sometimes a bad idea.

Here's an example: We are considering the universe of Parallels users on Macs in these discussions. Let's, for a moment, consider the universe of cell phone users who drive cars, an arguably larger group. "Most people" would like to be able to use their phones whenever they like even while driving. "Most people" can do this safely, but a few can't so it's the law in several states including the one where I live, that this practice is prohibited unless a hands free device is used. This goes against the wishes of the majority, but is the law nonetheless, and probably makes the roads safer even though the majority can use phones safely while driving.

And to address the obvious counter argument that the phone issue affects others negatively, I'll point to seat belt and helmet laws that protect only the individual who might not otherwise choose protection. This is done because it's good for the community. Protecting computers from compromise is good for the community too, IMO.

 

[drval]

That really isn't a solution because some 1970 features are not completely working. Disabling potentially unstable or dangerous features, preferably by default, is, IMNSHO, the way to go. This is based on many years of experience in security.

Microsoft has for years had the attitude that features should trump safety, and the result is the mess we are in today with millions of compromised machines. They have changed their attitude, and Vista now errs in the side of safety. They did this because of pressure to fix their lack of security. Going backwards seems like a really bad idea.

I think the idea that the default configuration should be what "most people" want, is a really, really bad idea. Most people don't bother turning on the Windows firewall (a truly stupid mistake). Most people aren't technical enough to make these decisions, and satisfying most people is sometimes a bad idea.

Here's an example: We are considering the universe of Parallels users on Macs in these discussions. Let's, for a moment, consider the universe of cell phone users who drive cars, an arguably larger group. "Most people" would like to be able to use their phones whenever they like even while driving. "Most people" can do this safely, but a few can't so it's the law in several states including the one where I live, that this practice is prohibited unless a hands free device is used. This goes against the wishes of the majority, but is the law nonetheless, and probably makes the roads safer even though the majority can use phones safely while driving.

And to address the obvious counter argument that the phone issue affects others negatively, I'll point to seat belt and helmet laws that protect only the individual who might not otherwise choose protection. This is done because it's good for the community. Protecting computers from compromise is good for the community too, IMO.

I, too, have been involved in security related issues in computing for over two decades -- we simply have very different perspectives on features. BTW, the Windows firewall was turned OFF by default after users complained.

This is an old issue which has many legs, as I pointed out in the note bringing together various URLs related to safety. Part of the collision here is between an end-user oriented product -- like Parallels is largely -- and an IT department oriented product -- like VMWare is largely. End users want convenience AND safety. IT departments heavily weigh in on the side of safety -- wisely so -- but also in terms of control. They like things to, in general, be "their way or the highway". That certainly promotes safety but can be a real problem re: innovation, as well as ease of use.

Perhap the installation dialogues should step you through each of the various options and FORCE YOU to make a choice.

But I'm not hearing this as an acceptable option in the minds of those committed to removing the "evil" of Global Sharing. If it CAN be disabled then those who WANT TO DISABLE it can do so, while those of us -- apparently in the majority -- can choose to use it as designed because of its ease, or whatever other reason appeals to us.

There seems to be just a bit of an "extra edge" to the presentations about "safety" being promoted for the sake of "the rest of us", to gloss Apple's famous phrase. To gloss Shakespeare: "Methinks the <critics> doth protest too much."

What you don't like, disable.

 

[dkp]

If it CAN be disabled then those who WANT TO DISABLE it can do so, while those of us -- apparently in the majority -- can choose to use it as designed because of its ease, or whatever other reason appeals to us.{/QUOTE]

Cite please - I've not seen a poll, scientific or otherwise.

There seems to be just a bit of an "extra edge" to the presentations about "safety" being promoted for the sake of "the rest of us", to gloss Apple's famous phrase. To gloss Shakespeare: "Methinks the <critics> doth protest too much."

After the absurd accusations and fabrications you made against me, my purpose, and intentions here, you have the gall to suggest others have an edge? Classy stuff, Val.

What you don't like, disable.

Or what you like, enable. Even if you know it to be unsafe.