In addition to the obvious 'your system will likely corrupt itself in day-to-day usage', I'm fairly sure when anyone comes up with a scheme *like* this and says "this is entirely legal", that they probably are wrong. In a scenario in which you are compelled to give up your password (e.g. a country where that is the law), I imagine they could slap you with something to the effect 'conspiracy to tamper with evidence'.
Some people treat the law as a pure logic problem, whereas my impression is a fair amount of 'common sense' is employed. The technical 'truth' that the key was not on disk probably counts for little in a court of law, anyone with half a brain cell can see clearly your intent to tamper with evidence in the event of a warrant.
I don't want an optical drive in my laptop. It's added weight and a little noise on reboot. For me this is no problem, I never ever use the optical drive and my question doesn't apply to me since I run linux exclusively anyway, but do SecuROM gimped games work with USB attached optical drives? I could see that as a major inhibitor to a lot of people.
Their pricing is at least respectable for computer components., though any particular product can usually be found elsewhere for less. If I'm going to buy a set of stuff, I haven't found another place that has all of what I want and where the sum of all the items is lower (except cables where monoprice really shines). Between laziness and shipping costs, I tend not to buy different pieces from different places.
I also question the sentiment about selection. I haven't found another place where selection is better (at least for computer components).
For me, what drives a lot of reason for using them for components even if in theory I could find them on amazon, their product categorization is pretty specific and it's much easier to identify sets of things with traits that a place like amazon would not make trivial to select by.
I will say I have casually glanced at the other products (e.g. HDTVs, laptops, random stuff) and haven't seen any particularly compelling reason to go with newegg for that sort of stuff.
But basically, B&N has nothing to loose. If they loose they have to pay the same fee as if they didn't. It is not as if MS can hurt them in any other way.
Well, they do have something to lose. If they lose, Microsoft will seek a more drastic action. I think the assumption is they feel confident they can win, and if other companies reached the same conclusion that they could 'win' in court, they would still 'lose' on their business relationship.
Think about the companies they went after to date. Samsung and HTC both partner with microsoft on endeavours outside of Android (laptops and windows mobile). For both of those companies, caving in may have been considered a safer move from a business relationship move.
Now B&N has absolutely no worries about being penalized on a business relationship they simply do not have. It's worth it for them to fight.
Copyright law a bit tortured here. In the case where copyright material came on a physical good, it worked fine as the container was meaningfully transferable. In a case where the material *never* comes as a good, it's a non-trivial problem to ascertain what is fair. We can't pretend the CD (rather than the music) was the thing of intrinsic value when sold to a used music store, but then on the other hand we say the music cannot retain value without being on an 'original' CD.
The problem being that you *can't* really even pretend to 'give away' digital content without DRM. In the event of death, I think you can easily say the previous owner will not likely use their copies again, so inheriting makes sense (though should not count in any way in terms of 'estate tax'). In a divorce, I think practically both get the 'property' (maybe a *smidge* unfair to the music industry, but I doubt this constitutes much 'loss'). For purposes of bankruptcy, I'd say it's fair to call the music library valueless (I've never gone through bankruptcy, but I was under the impression that things on the order of CDs didn't really count anyway except in extraordinary circumstances.
Now, reselling a license without any even vague enforcement, I could see the RIAA getting rightfully worried that they have no hope at all of reselling really being a transfer rather than copy. I can't really imagine a good way of resolving this without DRM. I'd rather have no 'secondary' market than one made possible through DRM,.
Of course, doing that complicates the whole DRM benefit. Given the only thing that makes DRM even an annoyance is 'security through obscurity', an open implementation would be rather problematic to hulu and others.
Here's the thing, *before* people would buy the authoring tools either based on the merits of the tools *or* simply because they were the logical choice given their ironclad control of the ecosystem. Killing flash for them means the latter benefit evaporates and now it must sell solely on its own merits. Their position is simply weaker than it was. The users of the free flash plugin were, in essence, a valuable part of the offering Adobe had. If it's authoring HTML5/Javascript voodoo, well Adobe hardly has a lock on it and all the major browser vendors are likely candidates for competition with the edge of actually writing popular end-user implementations of the standard, while Adobe will have zero control over that.
Note also of course that they haven't given up on Flash on the desktop, they've only given up on Flash on mobile devices.
The thing is, mobile/tablet usage is a non-trivial part of the market. Why buy into Adobe's stuff when a good chunk of clients will be unable to read it?
If the flash development tools have gone bust completely, I can see it, but I just really wonder if Adobe really has lost so much on it relative to other endeavours......
De-emphasizing Flash is probably the worst move they could have made as a business. The user experience may come out better in the end as a result, but at the expense of Adobe's bottom line. They pretty well *still* have control over much of the web (particularly streaming sites with DRM demands).
I think the key factor is Adobe trying to emphasize a strategy to make nice with iOS, but they are trying to do so at the expense of a pretty robust core and I simply don't see them succeeding in the IOS world, with or without flash.
By some stretch it isn't keyboardless, as gobs of time people end up with the soft-keyboard up, providing no tactile feedback, no tangible cues as to where the keys are, and covering up the very screen you interact with.
I think if you find yourself with the keyboard up a lot of the time, you are not a good user for the tablet experience. If you are one of the consume-only people in the world, I suppose that works, but I'm personally failing to see why a tablet when a phone is pretty well serviceable for that.
Last I heard, the start menu is replaced with Metro. Even in 'legacy' view, the start button was the piece of real estate they nominated as 'go back to metro'.
I don't think a fictitious portrayal of a tablet us reflects the real practicality. One, it's a prop and shown only briefly on screen. You don't feel the discomfort of supporting it without propping it up like naturally occurs with a laptop while seated or a sort of stand. You don't think about how much dictation is awkward in most people's interaction with systems. I would go nuts if the 8 people within earshot at work were constantly dictating to their computers. Voice interaction is great in the car, but limited applicability other than that. As to neural scan interface, maybe one day, but that's not even remotely a practical thing today (I'm aware of some 'toy' that does something like that, but it's not a practical thing by any means last I saw).
The plasma desktop still acts 'weird' to me. They've restored much of the capability (adding plasmoids to the panel plasma, resize), but somehow it still doesn't feel as straightforward as the panels that preceded it (perhaps for one, when I change screen resolutions occasionally the panel plasma decided to retain the previous resolutions width). It certainly isn't touch happy, but I think a bit too 'gadget' happy.
Talk about an article just asking for rants. I'll chip in my rant...
I think the challenge is the UI paradigm preceding this generation is just too mature and way too many UI developers really have a hard time justifying their continued work. The MATE and Trinity projects forked out of an apparent strong desire to keep things as they are and have some confidence it won't magically bit-rot away, but they are far from 'glamorous' and really don't have much of substance to actually *do*, the job is pretty much already complete.
Now a whole generation of UI designers are largely pretending that computers *didn't* catch on every where and that some mythical large mass of people cannot cope with the UIs that all evidence suggests are working just fine. For a time they were sated with the genuine issue of UI design not scaling down to ~4" screens, but they are seized with the silly notion that there must be *one* UI to rule all form factors. MS decides their Metro UI is the answer for phones/tablets/desktops (despite not even making sufficient headway in the handset arena to prove that out even in the most likely case). Nearly every review of use of the Metro-UI in Windows 8 suggests a degree of awkwardness in the laptop and desktop case. Apple decides the iOS experience should dominate the OSX world (Apple is a bit of a special case, they can pretty much do *anything* and their loyal userbase will lap it up, it's more like a fashion brand and they probably see minimal difference in business results between the times they truly deliver an enriching experience and when they make missteps). Gnome 3 pisses away tons of screen real estate on oversized default titlebars to accommodate inprecise touch interaction regardless of context whilst also hiding their 'dock' for fear of wasting real estate.
A large part of this is what I think is a bad assumption that tablets will just logically displace all laptops/desktops. iPad has seen commercial success (for reasons I think are more fanboy than a 'genuine' revolution) and now a ton of companies are wondering why they can't reproduce those results and get people off their laptops and assume something must be 'wrong' since tablets are *obviously* the way of the future.
Anyway, if you want the UI paradigm to continue as it has been, throw your weight behind MATE (or see if MGSE successfully decrapifies Gnome 3) or Trinity. Elect not to upgrade from Windows 7 if you prefer that (though you are at the mercy of MS in that scenario and you cannot force them to keep Windows 7 going). Alternatively prove me wrong by embracing KDE4, Gnome3, Metro, full-screen OSX apps as you get off my lawn.
The ability to rapidly change content vs. the *need* to actively require power to maintain screen contents are two distinct points. In practice, those two points are in conflict with today's tech. eInk has steady-state properties that allow the display to be 'off' and still readable, but changing the state takes effort and incurs a large time penalty. If eInk had the steady-state property *and* could change between any possible states in under a millisecond, that would still have the battery-saving properties and be responsive.
Now I understand where the 2% desktop share comes from, seriously, our attitude (and taste?) toward USER experience and UI is bad
There could be all sorts of arguments made about the subjective assessment of 'good' v 'bad', but at *worst* 'our' attitude/taste is different, not bad.
If there are only 2% of the world sharing my tastes, I'd rather not piss away my only platform to boost that number. Market share is not worth giving up the whole damn point of existence.
I use SPICE, have used NX and, in a pinch, VNC. X 'the way it was designed' disintegrates in the face of network latency. At low latency, X is servicable and a lot more seamless than the alternatives (though NX rootless can do *much* (not all) of the seamless bits of X). If my laptop is going to switch from wired to wireless or hop from one access point to another X is going to kill the clients mercilessly.
In short, X network capability has been neglected far too long and sadly is not adapting to new requirements well. NX was promising for me, but the session seemed to die too often destroying the state persistence benefit and didn't cope with 'tray icons' in rootless mode. Now I've resigned myself to SPICE, it still sucks in terms of seamless integration into local applications, but at least the performance is good.
Huh? gnome-terminal --disable-factory works in my gnome 3 and gnome 2 system identically. I'm unaware of another way to select all-in-one or distinct processes in either environment...
which is a disaster for people who need to use the command line a lot.
I get uncomfortable with the reliability implication of all my terminals being beholden to a single process as complex as gnome-terminal, but calling it a 'disaster' is a bit much. I currently have about 70 terminals open under a single process and it hasn't broken me. There *was* a file descriptor leak that was pretty nasty at one point, but with that addressed I haven't seen anything that afflicted me in practice, even if in theory it's a little less isolated than I would like.
1. Ctrl-Alt-N 2. Alt-[key-above-tab], though two-tiered tabbing is a bit weirder, it can be more powerful in the face of many heterogenous windows. 3. This is not new to Gnome 3, gnome-terminal has been that way for a long time. I don't necessarily agree with this.
gnome-tweak-tool, despite being well out of the way, still offers very little in the way of customization.
customization requires people to put on a developer hat and write 'extensions'.
Despite all this time no one has restored 'search by window title' functionality (there is one, but it doesn't interact with the window preview view, which is still well behind the state of KDE or compiz). We also still don't have a 'preview all windows belonging to a single app' despite the lengths of having a 'dock' group windows together that provides an intuitive trigger for such a behavior (this behavior is in KDE and compiz).
I honestly would not mind the experience given a rich set of themes and those two particular behaviors added. On the flipside, I do know many people consider the overhead of the graphical strategy to be too much, and being told an even more resource intensive software OpenGL rendering engine is going to be the answer is just putting salt in the wound.
The abstract concept of 'CA' is not the issue, the issue is the x509 strategies of 'a single CA is sufficient' and 'any CA is good enough'.
In the DNSSEC case, the 'any CA is good enough' is taken out, it *must* be the server in the specific trust chain. 'sketchyauthority.net' getting compromised would do nothing for 'importantbank.com'..
The reason I said to encode which CAs you elect to use (that must *also* be in the set of 'trusted' CA certs in browsers) is to take care of the other part. By requiring a compromise of both your DNSSEC chain of trust *and* a CA at the same time, that significantly mitigates the risk. Being able to dictate multiple CAs is also good. The catch is you must assure your DNSSEC authorities have *no* control over *any* trusted CAs, or else they could point your DNS record at their CA (though that could be mitigated if the client implementations require 3 distinct CAs before indicating a secure connection).
My issue with Convergence is while taking care of the 'a single CA is sufficent' problem, it doesn't sufficiently address the 'any CA is good enough' issue and greatly weakens the strength of validation in practice compared to a working-as-intended CA relationship. Instead of critically considering the specific weaknesses and evolving a system that addresses those issues while preserving the 'good' in status-quo, these strategies seek to burn everything to the ground and start over.
What does this gain you over storing the cert signature itself in DNSSEC?
Ideally (perhaps forced by policy), the DNSSEC chain of trust has no overlap with x509 CAs. If you compromise DNSSEC chain of a target, you still have to compromise a CA before they figure it out and fix it. If you compromise a CA, then you have to break DNSSEC before TLS implementations revoke the CA.
Nothing prevents a notary from taking extra steps to verify the authenticity of a certificate.
That's a fairly weak 'assurance'. 'other methods of authentication can be added in a modular way' I read as 'more ways to induce a false-positive'. *Ideally* In the CA relationship, you would at least have assurance that the site being validated worked explicitly with a trustworthy CA. In the reputation system, the site being validated didn't work with anyone and has no way to authoritatively 'tell' someone they got compromised.
Slashdot Mirror
In addition to the obvious 'your system will likely corrupt itself in day-to-day usage', I'm fairly sure when anyone comes up with a scheme *like* this and says "this is entirely legal", that they probably are wrong. In a scenario in which you are compelled to give up your password (e.g. a country where that is the law), I imagine they could slap you with something to the effect 'conspiracy to tamper with evidence'.
Some people treat the law as a pure logic problem, whereas my impression is a fair amount of 'common sense' is employed. The technical 'truth' that the key was not on disk probably counts for little in a court of law, anyone with half a brain cell can see clearly your intent to tamper with evidence in the event of a warrant.
I don't want an optical drive in my laptop. It's added weight and a little noise on reboot. For me this is no problem, I never ever use the optical drive and my question doesn't apply to me since I run linux exclusively anyway, but do SecuROM gimped games work with USB attached optical drives? I could see that as a major inhibitor to a lot of people.
Their pricing is at least respectable for computer components., though any particular product can usually be found elsewhere for less. If I'm going to buy a set of stuff, I haven't found another place that has all of what I want and where the sum of all the items is lower (except cables where monoprice really shines). Between laziness and shipping costs, I tend not to buy different pieces from different places.
I also question the sentiment about selection. I haven't found another place where selection is better (at least for computer components).
For me, what drives a lot of reason for using them for components even if in theory I could find them on amazon, their product categorization is pretty specific and it's much easier to identify sets of things with traits that a place like amazon would not make trivial to select by.
I will say I have casually glanced at the other products (e.g. HDTVs, laptops, random stuff) and haven't seen any particularly compelling reason to go with newegg for that sort of stuff.
But basically, B&N has nothing to loose. If they loose they have to pay the same fee as if they didn't. It is not as if MS can hurt them in any other way.
Well, they do have something to lose. If they lose, Microsoft will seek a more drastic action. I think the assumption is they feel confident they can win, and if other companies reached the same conclusion that they could 'win' in court, they would still 'lose' on their business relationship.
Think about the companies they went after to date. Samsung and HTC both partner with microsoft on endeavours outside of Android (laptops and windows mobile). For both of those companies, caving in may have been considered a safer move from a business relationship move.
Now B&N has absolutely no worries about being penalized on a business relationship they simply do not have. It's worth it for them to fight.
Copyright law a bit tortured here. In the case where copyright material came on a physical good, it worked fine as the container was meaningfully transferable. In a case where the material *never* comes as a good, it's a non-trivial problem to ascertain what is fair. We can't pretend the CD (rather than the music) was the thing of intrinsic value when sold to a used music store, but then on the other hand we say the music cannot retain value without being on an 'original' CD.
The problem being that you *can't* really even pretend to 'give away' digital content without DRM. In the event of death, I think you can easily say the previous owner will not likely use their copies again, so inheriting makes sense (though should not count in any way in terms of 'estate tax'). In a divorce, I think practically both get the 'property' (maybe a *smidge* unfair to the music industry, but I doubt this constitutes much 'loss'). For purposes of bankruptcy, I'd say it's fair to call the music library valueless (I've never gone through bankruptcy, but I was under the impression that things on the order of CDs didn't really count anyway except in extraordinary circumstances.
Now, reselling a license without any even vague enforcement, I could see the RIAA getting rightfully worried that they have no hope at all of reselling really being a transfer rather than copy. I can't really imagine a good way of resolving this without DRM. I'd rather have no 'secondary' market than one made possible through DRM,.
horrible UX on mobile
If only Adobe had some sort of ability to refine that UX, like perhaps feature enhancement releases of the mobile flash plugin...
Of course, doing that complicates the whole DRM benefit. Given the only thing that makes DRM even an annoyance is 'security through obscurity', an open implementation would be rather problematic to hulu and others.
Here's the thing, *before* people would buy the authoring tools either based on the merits of the tools *or* simply because they were the logical choice given their ironclad control of the ecosystem. Killing flash for them means the latter benefit evaporates and now it must sell solely on its own merits. Their position is simply weaker than it was. The users of the free flash plugin were, in essence, a valuable part of the offering Adobe had. If it's authoring HTML5/Javascript voodoo, well Adobe hardly has a lock on it and all the major browser vendors are likely candidates for competition with the edge of actually writing popular end-user implementations of the standard, while Adobe will have zero control over that.
Note also of course that they haven't given up on Flash on the desktop, they've only given up on Flash on mobile devices.
The thing is, mobile/tablet usage is a non-trivial part of the market. Why buy into Adobe's stuff when a good chunk of clients will be unable to read it?
If the flash development tools have gone bust completely, I can see it, but I just really wonder if Adobe really has lost so much on it relative to other endeavours......
De-emphasizing Flash is probably the worst move they could have made as a business. The user experience may come out better in the end as a result, but at the expense of Adobe's bottom line. They pretty well *still* have control over much of the web (particularly streaming sites with DRM demands).
I think the key factor is Adobe trying to emphasize a strategy to make nice with iOS, but they are trying to do so at the expense of a pretty robust core and I simply don't see them succeeding in the IOS world, with or without flash.
By some stretch it isn't keyboardless, as gobs of time people end up with the soft-keyboard up, providing no tactile feedback, no tangible cues as to where the keys are, and covering up the very screen you interact with.
I think if you find yourself with the keyboard up a lot of the time, you are not a good user for the tablet experience. If you are one of the consume-only people in the world, I suppose that works, but I'm personally failing to see why a tablet when a phone is pretty well serviceable for that.
Last I heard, the start menu is replaced with Metro. Even in 'legacy' view, the start button was the piece of real estate they nominated as 'go back to metro'.
I don't think a fictitious portrayal of a tablet us reflects the real practicality. One, it's a prop and shown only briefly on screen. You don't feel the discomfort of supporting it without propping it up like naturally occurs with a laptop while seated or a sort of stand. You don't think about how much dictation is awkward in most people's interaction with systems. I would go nuts if the 8 people within earshot at work were constantly dictating to their computers. Voice interaction is great in the car, but limited applicability other than that. As to neural scan interface, maybe one day, but that's not even remotely a practical thing today (I'm aware of some 'toy' that does something like that, but it's not a practical thing by any means last I saw).
The plasma desktop still acts 'weird' to me. They've restored much of the capability (adding plasmoids to the panel plasma, resize), but somehow it still doesn't feel as straightforward as the panels that preceded it (perhaps for one, when I change screen resolutions occasionally the panel plasma decided to retain the previous resolutions width). It certainly isn't touch happy, but I think a bit too 'gadget' happy.
Talk about an article just asking for rants. I'll chip in my rant...
I think the challenge is the UI paradigm preceding this generation is just too mature and way too many UI developers really have a hard time justifying their continued work. The MATE and Trinity projects forked out of an apparent strong desire to keep things as they are and have some confidence it won't magically bit-rot away, but they are far from 'glamorous' and really don't have much of substance to actually *do*, the job is pretty much already complete.
Now a whole generation of UI designers are largely pretending that computers *didn't* catch on every where and that some mythical large mass of people cannot cope with the UIs that all evidence suggests are working just fine. For a time they were sated with the genuine issue of UI design not scaling down to ~4" screens, but they are seized with the silly notion that there must be *one* UI to rule all form factors. MS decides their Metro UI is the answer for phones/tablets/desktops (despite not even making sufficient headway in the handset arena to prove that out even in the most likely case). Nearly every review of use of the Metro-UI in Windows 8 suggests a degree of awkwardness in the laptop and desktop case. Apple decides the iOS experience should dominate the OSX world (Apple is a bit of a special case, they can pretty much do *anything* and their loyal userbase will lap it up, it's more like a fashion brand and they probably see minimal difference in business results between the times they truly deliver an enriching experience and when they make missteps). Gnome 3 pisses away tons of screen real estate on oversized default titlebars to accommodate inprecise touch interaction regardless of context whilst also hiding their 'dock' for fear of wasting real estate.
A large part of this is what I think is a bad assumption that tablets will just logically displace all laptops/desktops. iPad has seen commercial success (for reasons I think are more fanboy than a 'genuine' revolution) and now a ton of companies are wondering why they can't reproduce those results and get people off their laptops and assume something must be 'wrong' since tablets are *obviously* the way of the future.
Anyway, if you want the UI paradigm to continue as it has been, throw your weight behind MATE (or see if MGSE successfully decrapifies Gnome 3) or Trinity. Elect not to upgrade from Windows 7 if you prefer that (though you are at the mercy of MS in that scenario and you cannot force them to keep Windows 7 going). Alternatively prove me wrong by embracing KDE4, Gnome3, Metro, full-screen OSX apps as you get off my lawn.
The ability to rapidly change content vs. the *need* to actively require power to maintain screen contents are two distinct points. In practice, those two points are in conflict with today's tech. eInk has steady-state properties that allow the display to be 'off' and still readable, but changing the state takes effort and incurs a large time penalty. If eInk had the steady-state property *and* could change between any possible states in under a millisecond, that would still have the battery-saving properties and be responsive.
Now I understand where the 2% desktop share comes from, seriously, our attitude (and taste?) toward USER experience and UI is bad
There could be all sorts of arguments made about the subjective assessment of 'good' v 'bad', but at *worst* 'our' attitude/taste is different, not bad.
If there are only 2% of the world sharing my tastes, I'd rather not piss away my only platform to boost that number. Market share is not worth giving up the whole damn point of existence.
I use SPICE, have used NX and, in a pinch, VNC. X 'the way it was designed' disintegrates in the face of network latency. At low latency, X is servicable and a lot more seamless than the alternatives (though NX rootless can do *much* (not all) of the seamless bits of X). If my laptop is going to switch from wired to wireless or hop from one access point to another X is going to kill the clients mercilessly.
In short, X network capability has been neglected far too long and sadly is not adapting to new requirements well. NX was promising for me, but the session seemed to die too often destroying the state persistence benefit and didn't cope with 'tray icons' in rootless mode. Now I've resigned myself to SPICE, it still sucks in terms of seamless integration into local applications, but at least the performance is good.
In Gnome 3 you're forced to open just one,
Huh? gnome-terminal --disable-factory works in my gnome 3 and gnome 2 system identically. I'm unaware of another way to select all-in-one or distinct processes in either environment...
which is a disaster for people who need to use the command line a lot.
I get uncomfortable with the reliability implication of all my terminals being beholden to a single process as complex as gnome-terminal, but calling it a 'disaster' is a bit much. I currently have about 70 terminals open under a single process and it hasn't broken me. There *was* a file descriptor leak that was pretty nasty at one point, but with that addressed I haven't seen anything that afflicted me in practice, even if in theory it's a little less isolated than I would like.
1. Ctrl-Alt-N
2. Alt-[key-above-tab], though two-tiered tabbing is a bit weirder, it can be more powerful in the face of many heterogenous windows.
3. This is not new to Gnome 3, gnome-terminal has been that way for a long time. I don't necessarily agree with this.
gnome-tweak-tool, despite being well out of the way, still offers very little in the way of customization.
customization requires people to put on a developer hat and write 'extensions'.
Despite all this time no one has restored 'search by window title' functionality (there is one, but it doesn't interact with the window preview view, which is still well behind the state of KDE or compiz). We also still don't have a 'preview all windows belonging to a single app' despite the lengths of having a 'dock' group windows together that provides an intuitive trigger for such a behavior (this behavior is in KDE and compiz).
I honestly would not mind the experience given a rich set of themes and those two particular behaviors added. On the flipside, I do know many people consider the overhead of the graphical strategy to be too much, and being told an even more resource intensive software OpenGL rendering engine is going to be the answer is just putting salt in the wound.
That wouldn't be so bad, but so many advocates say 'ditch the CA system and replace with reputation'. Augmenting is rarely ever suggested.
The abstract concept of 'CA' is not the issue, the issue is the x509 strategies of 'a single CA is sufficient' and 'any CA is good enough'.
In the DNSSEC case, the 'any CA is good enough' is taken out, it *must* be the server in the specific trust chain. 'sketchyauthority.net' getting compromised would do nothing for 'importantbank.com'..
The reason I said to encode which CAs you elect to use (that must *also* be in the set of 'trusted' CA certs in browsers) is to take care of the other part. By requiring a compromise of both your DNSSEC chain of trust *and* a CA at the same time, that significantly mitigates the risk. Being able to dictate multiple CAs is also good. The catch is you must assure your DNSSEC authorities have *no* control over *any* trusted CAs, or else they could point your DNS record at their CA (though that could be mitigated if the client implementations require 3 distinct CAs before indicating a secure connection).
My issue with Convergence is while taking care of the 'a single CA is sufficent' problem, it doesn't sufficiently address the 'any CA is good enough' issue and greatly weakens the strength of validation in practice compared to a working-as-intended CA relationship. Instead of critically considering the specific weaknesses and evolving a system that addresses those issues while preserving the 'good' in status-quo, these strategies seek to burn everything to the ground and start over.
What does this gain you over storing the cert signature itself in DNSSEC?
Ideally (perhaps forced by policy), the DNSSEC chain of trust has no overlap with x509 CAs. If you compromise DNSSEC chain of a target, you still have to compromise a CA before they figure it out and fix it. If you compromise a CA, then you have to break DNSSEC before TLS implementations revoke the CA.
Nothing prevents a notary from taking extra steps to verify the authenticity of a certificate.
That's a fairly weak 'assurance'. 'other methods of authentication can be added in a modular way' I read as 'more ways to induce a false-positive'. *Ideally* In the CA relationship, you would at least have assurance that the site being validated worked explicitly with a trustworthy CA. In the reputation system, the site being validated didn't work with anyone and has no way to authoritatively 'tell' someone they got compromised.