First, you were talking about three years. Now, it's five to ten all of a sudden. Nice trick, but that does not change much, really.
While China will have a higher average income, most of the over a billion people will still be dirt poor by our standards. Same for India. And Africa. Eastern Russia, Mongolia etc as well. Look at how income develops, and at how it's being distributed, over time. Ten years is nothing for the poorest of the poor. They will still be in exactly the same situation.
Once demand for monochrome displays drops below the threshold that simply keeping old factories and machines on running is not a worthwhile endeavor any more, then yes, the same will happen as with DDR2.
But that will definitely not happen withing three years. And most likely not within ten years, either.
More than a monochrome display. And that is all that matters.
While obviously absolute costs need to go below a certain threshold so someone can afford a phone in the first place, cheaper is still better, especially for developing nations.
> You wouldn't have any issues with dependency graph if you wouldn't lump everything into a single namespace.
I disagree. But the point is moot as I prefer one version over half a dozen versions.
> Thats because you have been lucky, not because there is anything stopping one of those repositories from totally fucking up your system.
No, it's because I know what I install from where, only using sources I can reasonably trust which are maintained by people who know what _they_ are doing.
> What I am hinting at is a system where you could mix and match everything, enter Ubuntu in your sources.list, do "apt-get install ubuntu/gnome" and have that work fine without breaking your current Debian Gnome installation or the rest of the system.
Insane workload, almost no benefit and, due to the halting problem, impossible to get right for everyone. No thanks.
> The incentive? Cleanup up the distribution mess that Linux is and providing the best system possible?
What mess? And where? Debian? The whole Linux distribution ecosystem? Define "best"... what metric? Your notion of cleanliness?
> You are thinking to much in competitive terms. Just because other are worse doesn't mean Debian couldn't be a whole lot better.
The fact that it's the best (for _my_ metrics) does not mean that it can't be improved. Agreed. I just disagree that your ideas would make it better in any way I care about or can even come up with.
> Thats not two versions of the same package but two different packages. Small, but very important difference.
Indeed. And it ensures that proper dependencies graphs can be built. IMO, this is the better solution than a mix with unintended consequences.
> Yes, that flaw isn't a Debian issue, it affects basically every Linux distribution out there.
> To provide an analogy: Debian is basically the SVN of distributions, nice, solid, does 95% of what you want. I want the Git of distributions, that gives me all of that plus a lot of more freedom and less dependency on some single central repository.
Sorry if this sounds like an ad hominem, but you do not seem to be aware of how a Linux distribution works.
Package dependencies _must_ be a graph. No way around it.
Ignoring the fact that I have six different sources in my sources.list at this time and mix from them without problems, if you really want the git of distributions, I suggest you use the git of distributions: Exherbo.
> I get these from Googlemail users about once a month.
You're either getting them via the invite user function (manually triggered) or via a third party accessing those contacts.
As you claim they are _for_ GMail, I would say your contacts are not as reliable/tech-savvy as you thought. Did you try asking them if they sent those mails (by chance)?
> First of all, just because Debian should be able to provide a proper base system doesn't stop it from providing a full distribution.
I am intrigued. Define this "should" you are speaking of. What's the rationale? What's the incentive?
> Secondly, the whole "base system" thing wouldn't just benefit Ubuntu, but also Debian, as it could clean up the whole base stable vs unstable issues, like say you want to install a unstable Gnome on a stable system, currently that would be quite a mess, if there would be proper decoupling between components on the other side, such a thing might very well be possible.
What mess? Large installation or server: stable. Smaller installation or server: testing. Personal desktop: unstable. Works for me. Pick what works for you.
The claim that installing Gnome on Debian stable would be a mess proves that you didn't install Debian yet. Name any other desktop OS that has seen half as much stability testing as Debian and offers even half, or a tenth, the packages.
> And while at it, how about allowing to install two packages with different versions at the time?
Possible if the devs _want to_. Simply create package1 and package2 for versions 1 & 2. Libraries do this regularly. Debian's community does not like that, though. Especially the security team.
> This is really not just about making things easier for Ubuntu, but getting rid of the whole monolithic dependency tree in general.
Wait, what? It's a graph, but that is not the matter here.
Every distribution has a graph of dependencies. The fact that the one of Debian happens to be the most stable and reliable one is a disadvantage, somehow?
> For instance, I have a bunch of Canadian and European friends who are coming to the US this year to vacation. The weaker US dollar means their Canadian dollars and Euros will go much farther. And Tourism is awesome because it brings money from outside of the economy in.
You seem to be missing the part where less people visit the USA because of the security theater. I know several people who postponed their trips indefinitely.
> The flaw in GOOG and Yahoo and Hotmail? Social networking "features". They get the email address of every contact you have, and spam them from your address in spoofed headers. All without a login credential.
Google is far from perfect and we should stay wary. But the above is undeniably false, at least for Google.
If you stopped producing carriages, cars would still be made.
Ubuntu, Mint, etc depend on Debian. Without it, they would need to base on something else or invest incredible amounts of time into providing it themselves.
It's nice that you enjoy mooching off of Debian, but some of us like to use Debian directly. Implying that Debian should limit itself so that others can more easily mooch is... strange?
Anyway, shouldn't Ubuntu, Mint, etc make an effort to properly base upon Debian, not vice versa.
But then, thankfully, Debian & Ubuntu devs are actually collaborating instead of simply going around, ranting about how you need to mooch moar.
Those people the article talked about, the ones who think Debian is irrelevant? You're one of them. Congrats, I guess.
Where is the extensive cryptanalysis proving that this does not weaken the result, allowing me to make guesses about the input? And how do you know there aren't any shortcuts to get to the result?
I note that the initial paper is from 1999, a time when MD5 was state of the art.
Excuse me for doubting this scheme without further proof.
Slashdot Mirror
Yay!
A yottapixel picture of dirt would tell you a lot. A _lot_.
(Ugh, I glitched a line. It wasn't you who talked about three years. Still, the same applies.)
First, you were talking about three years. Now, it's five to ten all of a sudden. Nice trick, but that does not change much, really.
While China will have a higher average income, most of the over a billion people will still be dirt poor by our standards. Same for India. And Africa. Eastern Russia, Mongolia etc as well. Look at how income develops, and at how it's being distributed, over time. Ten years is nothing for the poorest of the poor. They will still be in exactly the same situation.
Once demand for monochrome displays drops below the threshold that simply keeping old factories and machines on running is not a worthwhile endeavor any more, then yes, the same will happen as with DDR2.
But that will definitely not happen withing three years. And most likely not within ten years, either.
...but "facepunch forum" is awesome.
> Within three years I think the dynamic duo will have surpassed Android in marketshare.
If you expressively limit that bet to Windows OS and not Meego, Android or other, I am willing to take you up on this bet.
Ideally, I'd suggest FOSDEM 2014 for exchanging of the goods, whatever they may be. Looking forward to it.
> And what will a similar screen cost in 3 years?
More than a monochrome display. And that is all that matters.
While obviously absolute costs need to go below a certain threshold so someone can afford a phone in the first place, cheaper is still better, especially for developing nations.
HTC is rising, and fast. MS is their stepping stone.
Nokia is going the other way and MS is their life-boat.
Now you're changing the subject to public friend lists? Decide what to rant about ;)
I note you are not listing Google any more, I can't say much about the other two either way.
You opt out by having sane friends which you educate about the issues.
Web 2.0, etc? OK. Social networks only? No.
Not that it would matter either way as IMAP, contact sync etc exist. No net difference.
> You wouldn't have any issues with dependency graph if you wouldn't lump everything into a single namespace.
I disagree. But the point is moot as I prefer one version over half a dozen versions.
> Thats because you have been lucky, not because there is anything stopping one of those repositories from totally fucking up your system.
No, it's because I know what I install from where, only using sources I can reasonably trust which are maintained by people who know what _they_ are doing.
> What I am hinting at is a system where you could mix and match everything, enter Ubuntu in your sources.list, do "apt-get install ubuntu/gnome" and have that work fine without breaking your current Debian Gnome installation or the rest of the system.
Insane workload, almost no benefit and, due to the halting problem, impossible to get right for everyone.
No thanks.
That's totally different to what you were claiming earlier.
How an API means social networking is beyond me. APIs predate those sites and they are useful without them.
> The incentive? Cleanup up the distribution mess that Linux is and providing the best system possible?
What mess? And where? Debian? The whole Linux distribution ecosystem?
Define "best"... what metric? Your notion of cleanliness?
> You are thinking to much in competitive terms. Just because other are worse doesn't mean Debian couldn't be a whole lot better.
The fact that it's the best (for _my_ metrics) does not mean that it can't be improved. Agreed. I just disagree that your ideas would make it better in any way I care about or can even come up with.
> Thats not two versions of the same package but two different packages. Small, but very important difference.
Indeed. And it ensures that proper dependencies graphs can be built. IMO, this is the better solution than a mix with unintended consequences.
> Yes, that flaw isn't a Debian issue, it affects basically every Linux distribution out there.
> To provide an analogy: Debian is basically the SVN of distributions, nice, solid, does 95% of what you want. I want the Git of distributions, that gives me all of that plus a lot of more freedom and less dependency on some single central repository.
Sorry if this sounds like an ad hominem, but you do not seem to be aware of how a Linux distribution works.
Package dependencies _must_ be a graph. No way around it.
Ignoring the fact that I have six different sources in my sources.list at this time and mix from them without problems, if you really want the git of distributions, I suggest you use the git of distributions: Exherbo.
> I get these from Googlemail users about once a month.
You're either getting them via the invite user function (manually triggered) or via a third party accessing those contacts.
As you claim they are _for_ GMail, I would say your contacts are not as reliable/tech-savvy as you thought. Did you try asking them if they sent those mails (by chance)?
> First of all, just because Debian should be able to provide a proper base system doesn't stop it from providing a full distribution.
I am intrigued. Define this "should" you are speaking of. What's the rationale? What's the incentive?
> Secondly, the whole "base system" thing wouldn't just benefit Ubuntu, but also Debian, as it could clean up the whole base stable vs unstable issues, like say you want to install a unstable Gnome on a stable system, currently that would be quite a mess, if there would be proper decoupling between components on the other side, such a thing might very well be possible.
What mess? Large installation or server: stable. Smaller installation or server: testing. Personal desktop: unstable. Works for me. Pick what works for you.
The claim that installing Gnome on Debian stable would be a mess proves that you didn't install Debian yet. Name any other desktop OS that has seen half as much stability testing as Debian and offers even half, or a tenth, the packages.
> And while at it, how about allowing to install two packages with different versions at the time?
Possible if the devs _want to_. Simply create package1 and package2 for versions 1 & 2. Libraries do this regularly. Debian's community does not like that, though. Especially the security team.
> This is really not just about making things easier for Ubuntu, but getting rid of the whole monolithic dependency tree in general.
Wait, what?
It's a graph, but that is not the matter here.
Every distribution has a graph of dependencies. The fact that the one of Debian happens to be the most stable and reliable one is a disadvantage, somehow?
Bollocks.
> They don't need to know that I donated to wikileaks and other projects.
Seems as if they don't need to worry about getting that info, any more.
> For instance, I have a bunch of Canadian and European friends who are coming to the US this year to vacation. The weaker US dollar means their Canadian dollars and Euros will go much farther. And Tourism is awesome because it brings money from outside of the economy in.
You seem to be missing the part where less people visit the USA because of the security theater. I know several people who postponed their trips indefinitely.
> The flaw in GOOG and Yahoo and Hotmail? Social networking "features". They get the email address of every contact you have, and spam them from your address in spoofed headers. All without a login credential.
Google is far from perfect and we should stay wary. But the above is undeniably false, at least for Google.
ITYM Facebook.
Back when they announced it for the paid Google Apps versions, I was enthusiastic.
Now, everyone except me has it. And it's not even a technical issue, merely an incentive to buy stuff from them.
> BTW, the build system in Gentoo smells suspiciously like the system that has been in place in the various *BSD distros for years
FreeBSD calls its package eco-system ports, Gentoo calls it portage. This is not a coincidence.
To maintain your style, oh trusty AC:
>16 years old in 2011
>installed Windows 98
obvioustrollisobvious.jpg
If you stopped producing carriages, cars would still be made.
Ubuntu, Mint, etc depend on Debian. Without it, they would need to base on something else or invest incredible amounts of time into providing it themselves.
What the...?
It's nice that you enjoy mooching off of Debian, but some of us like to use Debian directly. Implying that Debian should limit itself so that others can more easily mooch is... strange?
Anyway, shouldn't Ubuntu, Mint, etc make an effort to properly base upon Debian, not vice versa.
But then, thankfully, Debian & Ubuntu devs are actually collaborating instead of simply going around, ranting about how you need to mooch moar.
Those people the article talked about, the ones who think Debian is irrelevant? You're one of them. Congrats, I guess.
Detailed analysis, followed by a concise statement of disagreement.
As an aside, looking through your other comments is fun. You seem to be on a roll.
So, it introduces a work factor.
Where is the extensive cryptanalysis proving that this does not weaken the result, allowing me to make guesses about the input? And how do you know there aren't any shortcuts to get to the result?
I note that the initial paper is from 1999, a time when MD5 was state of the art.
Excuse me for doubting this scheme without further proof.