It does indeed sound pretty neat. I wonder whether the additional load due to the broadcasting could become an issue on large installations. Also I'm curious whether it can somehow work across switched segments (if you want to discover hosts in a neighbour net).
While I doubt (correct me if I'm wrong..) that the broadcast mechanism could scale enough to replace old fashioned DNS it's still a nice substitute for DHCP at least.
And it definately makes some nice playground for the p2p hackers. If you can get the auto-discovery feature basically by linking a lib and adding some syscalls then I bet we will soon see a lot of utility apps learn how to find and talk to each other over the wire.
Can't wait for the xscreensaver plugin that connects to all other xscreensavers around and uh.. launches a sproingies contest.
Can I use the mousebutton of another mac in the cluster as additional apple-key/right-mousebutton for the primary desktop? (maybe put it down on the floor and tap it with your foot) What about latency?
TCO is a catch phrase and a particularly stupid one, too. It's just yet another meaningless acronym that everybody can (and does) bend and twist to their advantage.
These acronyms were invented in order to dumb down complex equations to short formulas so they can be conviniently shuffled around on a PowerPoint slide or serve as back up for thin slogans in shiny advertisements.
The problem when dumbing something down is always the same: you lose information because you have to make assumptions.
Two of your three URLS gave me a 404. The one that works has pretty pie charts but doesnt tell much more about TCO than that it truly is a complicated thing to measure and that this has to be done seperately and individually for each and every "implementation" (as they call it). It then goes indepth to describe the one model that is preferred by compaq: the "Gartner Group TCO Lifecycle Model".
So was MS using the gartner definition for their comparison? I didn't see it mentioned anywhere.
And since your other URLS were 404 I will prolly never find out which models are preferred by ibm or others and how these compare to whatever MS used...
Well, sorry, but DNS doesn't work that way. Most users can't easily add stuff to their DNS record because they don't have access to the responsible DNS server (which is good, DNS was not designed for that kind of usage).
And, well, adding p2p support features to DNS and the involved protocols isn't really a much better idea than trying to do it with SMTP & friends.
It just doesn't belong there. We need to replace FTP with something better, that's all. The rest of the internet ain't broken, so don't try to fix it.;-)
Guess where I found all these useless catch-phrases? Yes, on the article page.
MS and you deal with these terms as if they were constants that could easily be pre-determined for any possible context. But that's not the case. And there is no definition for any of them available on the MS page. They keep telling stuff like "Developement Cost for fantasy-project $P equals $X on MS-ware and $Y with linux". Their definition of "Developement Cost" is usually hidden in the explanations of the figures $X and $Y. Often these turn out to be the sums of many strange assumptions like training costs, travel costs or expensive software licenses. Anything goes as long it distorts the formula enough in favor to $X*1.5<$Y (=BINGO).
Well, now it's your turn, I bet you can turn up at least four more bingos only from that MS FUD page! And I haven't even clicked the links, yet...
That's not nice as it kills some of the biggest advantages of E-Mail: being available at the same address, forever and from everywhere.
I want to give my (=one) E-Mail address to anyone without having to worry about getting my inbox rammed with spam soon.
I don't want to make up an alias everytime I give someone my address (or post it somewhere) and I don't want to have to manage all these aliases.
What if one alias that constantly receives important mail (e.g. mailing list, newsletter, whatever) starts getting spam? Kill the alias and re-subscribe to all the mailing lists?
What if the alias that you gave to your most important customer starts getting spam? Kill the alias, kill the customer?
Multiplying the problem with aliases doesn't solve it..
I use galeon and checked "Open Popup Windows in Tabs". Good enough for me. I don't miss sites that depend on popups or display useful stuff in them (those are still out there *sigh*) and spotting/closing the ads takes no time - the title bar gives it away.
There are a few really nice sites where the webmaster makes some extra bucks (or any at all..) through popup ads. I like the fact that I'm still supporting them a bit that way even tho I never see their ads (unless I miss the X on first click...).
I used to go there but watch out, you can only take so much. After reading a dozen articles or so I really had to rush for the big white phone...
SCO is like one of these large brown piles of recent dogshit that tingles your nose long before you walk past it and that will make you puke if you come too close.
The good news is that IBM has stepped into it. And IBM certainly doesn't like shit on their shoes.
They're not tablet, more a cross of a PDA and a mini-notebook (touchscreen w/ swivel). But they're "instant-on". Oh, and it's running linux. IMHO pretty close to what you were asking for but see for yourself.
The screen is too small for enjoying a full-length movie and the CPU too weak to decode most stuff that you'd want to watch anyways. AND the thing is so big and ugly that you wouldn't even buy it to feed your vanity.
The only positive aspect seems to be that it'll burn some MS money until they drop it.
Agreed, installation worked like a charme on any recent "user"-distro that I've tried (Mandrake, SuSE). Also the default install tends to be a lot more streamlined today than it once was. Unified desktop themes, less choices, sane menu-entries etc. The progress is very visible. I can very well remember my first contacts with older distros where often you'd be left with a very rudimentary X install and application menus that are cluttered with all kinds of stuff nobody would ever want there (basically/usr/bin/*). That time is long gone, we now have graphical installers way more sophisticated than anything windows ever had to offer and out-of-the-box installations that "just work".
IMHO Knoppix & co show where it's going next. Installation to disk will be optional. Settings & data will be saved to memory-stick or the like. You can boot your desktop, whereever you are, on any PC. You only have to carry your CD and your memstick with you...
Give it another year of progress at the current (incredible) pace and we'll have a linux desktop more mature and polished than that "OS with the worms, spyware and stuff" that everybody was using back in '04...
Please do yourself a favor and try blackbox (or fluxbox) on your machine. Use galeon (or links) for browsing. Maybe look into mutt for E-Mail. You'll like it.
I dunno why everybody keeps complaining about "old hardware". Esp. old notebooks can be so damn useful. Okay, KDE on a p133 is a bit *cough* optimistic. But who needs KDE anyhow..
Well, the problem is interoperability. There already are dozens of different solutions to file exchange (ftp, saft, p2p, every IM-system has their own way to do it, too).
But all this has nothing to do with E-Mail.
People simply need to learn that E-Mail was not designed to transfer large files (~33% overhead due to MIME).
If you want to send $bigfile then don't use E-Mail. As easy as that.
Inform the admin of the mailing list that his mailing list software is not behaving properly, once and politely.
If the mails don't stop (even after a couple of bounces) you could either:
a) Block their sending mx ip at your packet filter (not applicable if it varies) or b) Forward all mail coming from them back to them (postmaster@, listmaster@ etc.)
I think what you say is wrong. SSL/TLS is not vulnerable to MiTM when configured properly and used properly.
The main cause why MiTM on SSL can happen in the wild is that most browsers allow you to override SSL-warnings and establish a connection even tho the identify of the other end can't be guaranteed.
Whenever your browser presents you with a warning message (whatever it is) regarding the SSL-connection that it is about to establish then make sure to realize that you could as well switch back to plain http at that point.
Slashdot Mirror
It does indeed sound pretty neat.
I wonder whether the additional load due to the broadcasting could become an issue on large installations. Also I'm curious whether it can somehow work across switched segments (if you want to discover hosts in a neighbour net).
While I doubt (correct me if I'm wrong..) that the broadcast mechanism could scale enough to replace old fashioned DNS it's still a nice substitute for DHCP at least.
And it definately makes some nice playground for the p2p hackers. If you can get the auto-discovery feature basically by linking a lib and adding some syscalls then I bet we will soon see a lot of utility apps learn how to find and talk to each other over the wire.
Can't wait for the xscreensaver plugin that connects to all other xscreensavers around and uh.. launches a sproingies contest.
Thanks for the hint. I'll definately give the linux port a shot (can't be worse than dhcp i guess).
Can I use the mousebutton of another mac in the cluster as additional apple-key/right-mousebutton for the primary desktop?
(maybe put it down on the floor and tap it with your foot)
What about latency?
TCO is a catch phrase and a particularly stupid one, too. It's just yet another meaningless acronym that everybody can (and does) bend and twist to their advantage.
These acronyms were invented in order to dumb down complex equations to short formulas so they can be conviniently shuffled around on a PowerPoint slide or serve as back up for thin slogans in shiny advertisements.
The problem when dumbing something down is always the same: you lose information because you have to make assumptions.
Two of your three URLS gave me a 404.
The one that works has pretty pie charts but doesnt tell much more about TCO than that it truly is a complicated thing to measure and that this has to be done seperately and individually for each and every "implementation" (as they call it).
It then goes indepth to describe the one model that is preferred by compaq: the "Gartner Group TCO Lifecycle Model".
So was MS using the gartner definition for their comparison? I didn't see it mentioned anywhere.
And since your other URLS were 404 I will prolly never find out which models are preferred by ibm or others and how these compare to whatever MS used...
Well, sorry, but DNS doesn't work that way.
;-)
Most users can't easily add stuff to their DNS record because they don't have access to the responsible DNS server (which is good, DNS was not designed for that kind of usage).
And, well, adding p2p support features to DNS and the involved protocols isn't really a much better idea than trying to do it with SMTP & friends.
It just doesn't belong there.
We need to replace FTP with something better, that's all. The rest of the internet ain't broken, so don't try to fix it.
You say TCO?
So, you wanna play bullshit bingo?
OK, let's go!
Embedded Developement Costs
Staffing Costs
Dynamic Content
Throughput
TCO
BINGO!
Guess where I found all these useless catch-phrases? Yes, on the article page.
MS and you deal with these terms as if they were constants that could easily be pre-determined for any possible context.
But that's not the case. And there is no definition for any of them available on the MS page. They keep telling stuff like "Developement Cost for fantasy-project $P equals $X on MS-ware and $Y with linux". Their definition of "Developement Cost" is usually hidden in the explanations of the figures $X and $Y. Often these turn out to be the sums of many strange assumptions like training costs, travel costs or expensive software licenses.
Anything goes as long it distorts the formula enough in favor to $X*1.5<$Y (=BINGO).
Well, now it's your turn, I bet you can turn up at least four more bingos only from that MS FUD page!
And I haven't even clicked the links, yet...
That's not nice as it kills some of the biggest advantages of E-Mail: being available at the same address, forever and from everywhere.
I want to give my (=one) E-Mail address to anyone without having to worry about getting my inbox rammed with spam soon.
I don't want to make up an alias everytime I give someone my address (or post it somewhere) and I don't want to have to manage all these aliases.
What if one alias that constantly receives important mail (e.g. mailing list, newsletter, whatever) starts getting spam?
Kill the alias and re-subscribe to all the mailing lists?
What if the alias that you gave to your most important customer starts getting spam?
Kill the alias, kill the customer?
Multiplying the problem with aliases doesn't solve it..
I use galeon and checked "Open Popup Windows in Tabs". Good enough for me.
I don't miss sites that depend on popups or display useful stuff in them (those are still out there *sigh*) and spotting/closing the ads takes no time - the title bar gives it away.
There are a few really nice sites where the webmaster makes some extra bucks (or any at all..) through popup ads. I like the fact that I'm still supporting them a bit that way even tho I never see their ads (unless I miss the X on first click...).
Unforna
I generally like flash (still waiting for a site that puts it to good use, tho)...
But for this one, well.
After playing around with it for like 10 minutes I'd describe it as "Google for PowerPoint users".
Or in short: completely useless...
I don't see why they'd launch a budget iPod that isn't really a budget iPod.
Don't forget to adjust your definition of "budget" to apple's "vanity has its price"-scale, tho.
Glad you had <<
My first walkman didn't have << which kinda doubled the seek time.
>||
[] [^] (flip)
>>
[] [^] (flip)
>||
[] [^] (curse, flip)
etc.
Amen. (SCNR)
I used to go there but watch out, you can only take so much. After reading a dozen articles or so I really had to rush for the big white phone...
SCO is like one of these large brown piles of recent dogshit that tingles your nose long before you walk past it and that will make you puke if you come too close.
The good news is that IBM has stepped into it.
And IBM certainly doesn't like shit on their shoes.
Run McBride, while you can.
Yahoo! plans to dump Google as its primary search technology
Erm, excuse me. Is there any other "search technology" that can only remotely compare to google quality and coverage?
To me the above phrase sounds like:
Yahoo! plans to render their search functionality useless.
Have you seen these?
They're not tablet, more a cross of a PDA and a mini-notebook (touchscreen w/ swivel).
But they're "instant-on". Oh, and it's running linux.
IMHO pretty close to what you were asking for but see for yourself.
Amen.
The screen is too small for enjoying a full-length movie and the CPU too weak to decode most stuff that you'd want to watch anyways.
AND the thing is so big and ugly that you wouldn't even buy it to feed your vanity.
The only positive aspect seems to be that it'll burn some MS money until they drop it.
Agreed, installation worked like a charme on any recent "user"-distro that I've tried (Mandrake, SuSE). Also the default install tends to be a lot more streamlined today than it once was. Unified desktop themes, less choices, sane menu-entries etc. The progress is very visible. I can very well remember my first contacts with older distros where often you'd be left with a very rudimentary X install and application menus that are cluttered with all kinds of stuff nobody would ever want there (basically /usr/bin/*).
That time is long gone, we now have graphical installers way more sophisticated than anything windows ever had to offer and out-of-the-box installations that "just work".
IMHO Knoppix & co show where it's going next. Installation to disk will be optional. Settings & data will be saved to memory-stick or the like. You can boot your desktop, whereever you are, on any PC. You only have to carry your CD and your memstick with you...
Give it another year of progress at the current (incredible) pace and we'll have a linux desktop more mature and polished than that "OS with the worms, spyware and stuff" that everybody was using back in '04...
I was gonna say the same.
But in the end it doesn't really matter.
We want it both banned; copy protection AND celine. Don't we?
And if it's only to save a Mac or two...
Dude, here's a list of window managers for you.
Please do yourself a favor and try blackbox (or fluxbox) on your machine.
Use galeon (or links) for browsing.
Maybe look into mutt for E-Mail.
You'll like it.
I dunno why everybody keeps complaining about "old hardware". Esp. old notebooks can be so damn useful.
Okay, KDE on a p133 is a bit *cough* optimistic. But who needs KDE anyhow..
Well, the problem is interoperability.
There already are dozens of different solutions to file exchange (ftp, saft, p2p, every IM-system has their own way to do it, too).
But all this has nothing to do with E-Mail.
People simply need to learn that E-Mail was not designed to transfer large files (~33% overhead due to MIME).
If you want to send $bigfile then don't use E-Mail. As easy as that.
Blargh! sendmail.
Is there really still a distro that ships with sendmail as default-MTA?!
Someone invite them to 2004 and tell them about qmail and postfix.
Inform the admin of the mailing list that his mailing list software is not behaving properly, once and politely.
If the mails don't stop (even after a couple of bounces) you could either:
a) Block their sending mx ip at your packet filter (not applicable if it varies)
or
b) Forward all mail coming from them back to them (postmaster@, listmaster@ etc.)
Both a) and b) have worked for me in the past.
I think what you say is wrong.
SSL/TLS is not vulnerable to MiTM when configured properly and used properly.
The main cause why MiTM on SSL can happen in the wild is that most browsers allow you to override SSL-warnings and establish a connection even tho the identify of the other end can't be guaranteed.
Whenever your browser presents you with a warning message (whatever it is) regarding the SSL-connection that it is about to establish then make sure to realize that you could as well switch back to plain http at that point.
As you're interested in P2P and how its done, here's a good read for you and everybody else who wants to take a challenge..
(found in the dusty bookmarks)