When I bought my first LP in 1976, the music at the top of the charts was "Disco Duck". People tend the remember the best music from each era in hindsight, forget about all the terrible stuff that's always been popular.
CDs are physically analog too. There's pits and this spiral groove, a mechanical spinning thing, a reading assembly...there's a whole layer of analog technology at the bottom there. There is a layer at the top that converts all of this mess into a series of bits, with considerable uncertainty and potential data loss. Every seen a scratch cause a read hiccup? How about the laser not positioning accurately and reading the wrong things? All that stuff happens, and it's all analog. The illusion that even data CDs are digital is done not via direct digital reading, it's done by inserting markers into the stream of pits that let the read mechanism realize when it's gone off the rails, so it can reposition.
Kids use whatever is the easiest thing out there. When Napster ruled, it was by far the easiest way to get a digital copy of a song. Now, if you want to get a song on your iDevice etc., it's a whole lot easier to wander into the iTunes store to buy it than the navigate the mess of illegal downloading. It's not so much a generational gap as a market response. Distributors are finally selling what people were willing to buy all along: a song for $1, if it's available immediately and is easy to get onto players. The full album length market has been gone for over ten years, but for a while there labels kept dreaming it would come back anyway, and priced accordingly. Now they're pricing to where people find it easier to buy than steal, so they buy. It was always about convenience.
US banks don't gamble. When their trades pay off, they pay executives with the proceeds. When they don't, they get a government bailout. Gamble implies some possibility of a loss.
One of the things that's becoming clear in corporate governance is that an employee of a corporation faced with making a decision that benefits themselves versus a decision that benefits the organization will pick themselves almost every time.
Hint: this is also true of government too, especially at the highest levels. In the US there is this entire profession, "lobbyist", whose primary goal is to convey money from companies toward government officials who do them favors. The idea that corporations are dirty there while government isn't would be a bit naive.
The Tulip chips gained a lot of their fame as being one of the fastest build a Beowulf cluster cards going back to August of 2000. 3c905B cards didn't work right under Linux until kernel 2.2.17 in September 2000. I believe the Tulip came out first, then the 3c905B, but it was very close in time. Exactly when the original 3c905 came out relative to those two is even harder to place.
In 2000 I could afford 3c905 cards but still preferred Tulip ones. Before Linksys started screwing up the market by releasing both Tulip and knock-off versions, the card to buy was the Kingston KNT40T or KNE100TX. Those were much cheaper than a 3c905, and on Linux they were faster and more reliable too. Eventually Netgear and Linksys replaced Kingston as the Tulip vendors of choice, and then they started racing toward lower quality/cost with clone chipsets.
Linux will be pushed off computers because Microsoft has too much power? It's hard to take that idea seriously even as a bit of a joke here in 2013, as MS struggles to release any sort of compelling product. I've been making a living working on open-source projects for 15 years now, and there has never been this much wide-spread adoption of Linux. Everyone from giant enterprises to little web app startups have a Linux web server somewhere, just like they probably have an Exchange mail server. The rise of Android phones in particular has been a huge PR win for Linux.
As for high security environments, that crowd has been burned so many times now by zero-day Windows exploits and holes in desktop software that they're giving up on trying. I do database consulting, and I'm moving everything from anti-virus vendors to defense contractors over to Linux now. The terror of Windows 8 plus the new Office licensing terms were the last straw for a lot of them.
For normal office workers, Firefox/Chrome plus a web mail client work just fine now. Google Docs is fine for simple tasks, and OpenOffice is good enough for a lot of medium scale issues. (Real MS Office is only needed for the most complicated work) Linux has never been more competitive compared to Microsoft's offerings than now. Every time a useful app is written inside of a web browser, the reason to use Microsoft drops a little. And that's where a lot of business oriented development is moving now.
The OS X kernel has a number of limitations compared the Linux one. Terrible filesystem choices is at the top of that list. HFS+ sucks hard, starting with the handling of upper/lower case names. Concurrency at high core counts is much better on Linux. There are some other things that are oddly slow on OS X too, like some of the memory and message passing interfaces. There are a whole lot of people who work on making Linux faster on a variety of environments, while Apple focuses on a relatively narrow chunk of desktop user cases. OS X is a pretty bad memory hog compared to a well configured Linux too.
On the flip side, I really wish Linux had the OS X DTrace interface and its associated tools. RedHat and other vendors have dumped so much time into cloning that badly with Systemtap.
That wasn't about the average case. In the average case, both Windows and Linux will bring up a display that fills the screen, after you install the appropriate driver for the card. The point of that comment is that Linux can make the difficult cases easier to deal with than the typical Window GUI, one that tries to hide the inherent complexity in things like TV output. I've spent plenty of time struggling to get video configuration working correctly under both Linux and Windows. On average I'd say it's harder in Linux, but when it doesn't work in Windows there's little you can do about it. There's always something you can do in Linux, the only question is how many resources you're willing to bring to bear on a problem.
18 months ago I bought 4 $100 drives from Seagate, putting them into a pair of RAID-1 arrays in two boxes. They handle error correction badly, sometimes dropping from the array for no good reason. And periodically they go complety crazy due to a firmware bug. Later versions of the drive ship with a fix for the problem. That fix isn't available for the cheapo ones I have. Seagate just doesn't care about whether someone with a $100 drive works reliably or not. (I've had the same class of problem with Hitachi and Western Digital, just picking on Seagate as the most recent one)
If this were a Netapp system with a support contract I'd expect both help tracking down this weird problem and a real resolution at the end. I would also expect they would only ship hardware that goes through more QA for the purpose of going into a RAID array than a random consumer drive does. The drives I have were for example dropped from the HCL of the Netgear ReadyNAS I own, due to this bug in them. Their support documents are what helped lead me to the source of the bug. I've been on phone calls with very sharp Netapp staff tracking down odd problems before. It was nice to have that level of storage expert available in addition to what I knew.
Are these things worth the price premium? That part is debatable. Claiming that they have zero value is really ignorant though. The more you know about hard drives, the more you appreciate that they are not simple works/doesn't work devices. Building reliable systems out of multiple drives is complicated and has a number of subtle failure possibilities. And individual consumer drives are normally not optimized for that very well. WD's Red line is starting to show some promise for good error handling in RAID, but again you're back to only a small number of drives relative to the mass of mostly funky hardware on the market.
It's not just about pandering to casual gamers. The Wii did that partly through having an innovative and interesting input device. I see the success of the Wii as a lucky accident that way. Who would have expected I'd find old people playing Wii bowling? That it was possible for people who weren't into gaming at all to use a Wiimote. There's nothing uniquely interesting about the Wii U that way though.
The Wii was the right technology at the right price for its marketplace. They've tried to duplicate that but with nothing innovative this time. They're not going to get lucky and happen to pick up a whole new market (the casual gamer then) this time.
The problem is that advertisers are like cockroaches; you can't kill them with nukes. When all of civilization has been reduced to a post-apocalyptic wasteland, and mutant zombies roam the land, there will still be someone trying to sell you that one weird trick for losing belly fat.
I used those specific words to be funny, but there's truth behind them too. I have my bank account setup such that an overdraft mistake will hit a credit card instead of bouncing. And I sometimes expect I can charge things in advance of the cash being available. When the credit card I used for PSN was stolen and used by thieves, both mechanisms weren't available until the fraud was cleared up. It didn't actually cost me anything directly, but control over my normal finances was taken from me by someone else for a while. When someone hits you with a denial of service attack, they have asserted a form of control over a resource.
Sony is reponsible for their security issues because their storage measures for personal data were pathetic. What the fuck were they doing saving so much unencrypted data? The issue is not just that they were breached, which does happen to lots of companies; it was that they obviously didn't care at all about securing things until they were.
As for the rest of your speculation, the credit card I used for PSN purchases was used to make $3000 of purchases shortly after the breach. Are those two connected? I can't be sure, but the Sony incident was the only such breach in the time period before then. It didn't cost me anything, but I was definitely unable to use that card until a replacement showed up. If I couldn't use it, that means I lost control of the account to another party. Am I supposed to excuse Sony for that because the thieves didn't manage to steal all of my identity? No.
Don't call me a sheep when you're posting your rambling and poorly written bullshit as an AC. Your apologist fanboy crap is pathetic.
If the free market value for the property was $250K, there would be multiple bidders willing to pay that amount. This case is more like extortion than it is a market. "We gots this domain with your name on it...sure would be a shame if we said bad things about you there". It is domain squatting bullshit, and they're lucky the penalty is only losing the domain. This isn't sex.com or something that has some inherent market value to multiple buyers. It's only worth anything because of the association of Ron Paul, the politician, and it's only worth >$0 to him alone.
The sellers are playing their own semantic game, where they've only offered the sale of the domain plus their mailing list, for a large sum based on the list value. They only did that because the market value of the domain itself, the only thing Paul was interested in, is zero. They know it's not worth anything because they're bound by the domain agreement, which says you can't do the thing they did. This is more like a contract dispute than it is government involvement. The distinction is that you don't have to buy a.com domain and agree to its terms. But if you choose to do so, you then have to follow the rules you agreed to.
You'd lose that bet. The rate of future I/O operations is extremely difficult to work out at any price, with any quality of engineer. I would bet you a billion dollars it can't be done reliably in the case of HTTP downloads, and not lose a moment of sleep about losing. If you spent all of the money hiring the best developers in the world with that billion, all you'd have when you were done is a better understanding that it's impossible. In fact, I'd make a side-bet that someone tells you it's impossible before they even start!
I would have stopped at "yes" without the optimism, and it's not even a platform/processor issue. This is a fundamentally unsolvable thing, since in some cases it reduces to the halting problem. That best you can do is come up with an approximation.
But the approximations will be wrong too. Simple example: write speed. If you write a block to disk, and it fits into the operating system cache, a write happens almost instantly. If the cache is full, you might have to spend tens of seconds waiting before that write occurs. Here's what happens with installers, copy programs, and a large chunk of other things. You write until the cache fills, and those happen at memory speeds. Then, all of the sudden, you grind to a halt when the cache fills. You won't see any progress, sometimes for minutes--it can take a while to chug through gigabytes of random writes. I show an example of this on my blog on Linux, and this problem gets worse as memory increases, not better.
What does that look like to the user? They get a progress bar, it zooms along for a while, and then it cranks to a halt. Then it hangs for a bit, starts moving, and the whole thing completely changes scale. Does this sound familiar? That's what people complain about, right? You can't make that go away without building a model of the caching mechanism that's more accurate than the cache itself. After all, if you could predict this was coming, the OS could have done a better job scheduling I/O with that information, too. Think about that for a minute: to write a really good progress bar for write operations, you have to do a better job on I/O scheduling than Linux does.
If you step back and say "well let's approximate how long disk I/O takes then and base the bar on that", you'll discover that doesn't work either. There's over a 100:1 difference between the fastest and slowest storage on the market. Good luck modeling that accurately enough to predict the future, too.
A 2 hour movie can be fit into 9GB, because that's how much a single DVD can hold. 5 TB lets you transfer every byte of data for over 500 movies per month. And even if you're playing them 24x7, you can only fit 360 2-hour movies into a month.
You cannot overrun 5 TB with DVD quality playback unless you bloat the broadcast excessively. It sounds like your movie server doesn't care about how much bandwidth it uses. That's fine when it's free. If it isn't, you can do much better, and 5 TB is plenty for a month of standard definition video.
If the short-term math makes sense, he can deploy the home server and then see what happens. When the replacement parts get expensive, you can always switch to the VPS at that point. The short-term spending is a $100 upgrade to save about $20/month. The only lifecycle question worth asking is "is the system likely to last 6 months?" If it does, you can bank the savings until the home server dies, and then figure out what to do.
This sort of thinking is not an option is not available to business class service, because it assumes a day or two of downtime is acceptable. At home, you can do the "let's see how long this server lasts" game. As long as you're willing to jump to another option when you're facing another "capital expense", you can usually find a cheaper long term path.
Not necessarily. True story: I once had "Greg Smith Equipment Sales" try to buy my web site because I was messing with their search rank. What really killed both of us was a baseball player with that name though. Once ESPN started writing about him I was done with being in the top 3. I wouldn't recommend just any common name; what you really want is the name of a celebrity. I use "Michael Bolton" now.
The only embarrassing things I did online 15 years ago were posts to Usenet that seem a bit immature now. I think it was easier then to see that things you posted went out of your control and might be around forever though, because that distribution was itself sometimes a technical problem to be solved.
I've had the same e-mail address for over 15 years though. For the majority of the login/password combinations I created, I outlasted the site I created it on. I'm not exactly worried that I can't provide the credentials for the gopher account I signed up for once in the early 90's.
Slashdot Mirror
When I bought my first LP in 1976, the music at the top of the charts was "Disco Duck". People tend the remember the best music from each era in hindsight, forget about all the terrible stuff that's always been popular.
CDs are physically analog too. There's pits and this spiral groove, a mechanical spinning thing, a reading assembly...there's a whole layer of analog technology at the bottom there. There is a layer at the top that converts all of this mess into a series of bits, with considerable uncertainty and potential data loss. Every seen a scratch cause a read hiccup? How about the laser not positioning accurately and reading the wrong things? All that stuff happens, and it's all analog. The illusion that even data CDs are digital is done not via direct digital reading, it's done by inserting markers into the stream of pits that let the read mechanism realize when it's gone off the rails, so it can reposition.
Kids use whatever is the easiest thing out there. When Napster ruled, it was by far the easiest way to get a digital copy of a song. Now, if you want to get a song on your iDevice etc., it's a whole lot easier to wander into the iTunes store to buy it than the navigate the mess of illegal downloading. It's not so much a generational gap as a market response. Distributors are finally selling what people were willing to buy all along: a song for $1, if it's available immediately and is easy to get onto players. The full album length market has been gone for over ten years, but for a while there labels kept dreaming it would come back anyway, and priced accordingly. Now they're pricing to where people find it easier to buy than steal, so they buy. It was always about convenience.
Basically, they used a law that was never designed for a networked computing based world and applied it to a network computing based world
What the hell? Don't those losers know you're supposed to patent this sort of shit first?
US banks don't gamble. When their trades pay off, they pay executives with the proceeds. When they don't, they get a government bailout. Gamble implies some possibility of a loss.
One of the things that's becoming clear in corporate governance is that an employee of a corporation faced with making a decision that benefits themselves versus a decision that benefits the organization will pick themselves almost every time.
Hint: this is also true of government too, especially at the highest levels. In the US there is this entire profession, "lobbyist", whose primary goal is to convey money from companies toward government officials who do them favors. The idea that corporations are dirty there while government isn't would be a bit naive.
The Tulip chips gained a lot of their fame as being one of the fastest build a Beowulf cluster cards going back to August of 2000. 3c905B cards didn't work right under Linux until kernel 2.2.17 in September 2000. I believe the Tulip came out first, then the 3c905B, but it was very close in time. Exactly when the original 3c905 came out relative to those two is even harder to place.
In 2000 I could afford 3c905 cards but still preferred Tulip ones. Before Linksys started screwing up the market by releasing both Tulip and knock-off versions, the card to buy was the Kingston KNT40T or KNE100TX. Those were much cheaper than a 3c905, and on Linux they were faster and more reliable too. Eventually Netgear and Linksys replaced Kingston as the Tulip vendors of choice, and then they started racing toward lower quality/cost with clone chipsets.
Linux will be pushed off computers because Microsoft has too much power? It's hard to take that idea seriously even as a bit of a joke here in 2013, as MS struggles to release any sort of compelling product. I've been making a living working on open-source projects for 15 years now, and there has never been this much wide-spread adoption of Linux. Everyone from giant enterprises to little web app startups have a Linux web server somewhere, just like they probably have an Exchange mail server. The rise of Android phones in particular has been a huge PR win for Linux.
As for high security environments, that crowd has been burned so many times now by zero-day Windows exploits and holes in desktop software that they're giving up on trying. I do database consulting, and I'm moving everything from anti-virus vendors to defense contractors over to Linux now. The terror of Windows 8 plus the new Office licensing terms were the last straw for a lot of them.
For normal office workers, Firefox/Chrome plus a web mail client work just fine now. Google Docs is fine for simple tasks, and OpenOffice is good enough for a lot of medium scale issues. (Real MS Office is only needed for the most complicated work) Linux has never been more competitive compared to Microsoft's offerings than now. Every time a useful app is written inside of a web browser, the reason to use Microsoft drops a little. And that's where a lot of business oriented development is moving now.
The OS X kernel has a number of limitations compared the Linux one. Terrible filesystem choices is at the top of that list. HFS+ sucks hard, starting with the handling of upper/lower case names. Concurrency at high core counts is much better on Linux. There are some other things that are oddly slow on OS X too, like some of the memory and message passing interfaces. There are a whole lot of people who work on making Linux faster on a variety of environments, while Apple focuses on a relatively narrow chunk of desktop user cases. OS X is a pretty bad memory hog compared to a well configured Linux too.
On the flip side, I really wish Linux had the OS X DTrace interface and its associated tools. RedHat and other vendors have dumped so much time into cloning that badly with Systemtap.
That wasn't about the average case. In the average case, both Windows and Linux will bring up a display that fills the screen, after you install the appropriate driver for the card. The point of that comment is that Linux can make the difficult cases easier to deal with than the typical Window GUI, one that tries to hide the inherent complexity in things like TV output. I've spent plenty of time struggling to get video configuration working correctly under both Linux and Windows. On average I'd say it's harder in Linux, but when it doesn't work in Windows there's little you can do about it. There's always something you can do in Linux, the only question is how many resources you're willing to bring to bear on a problem.
18 months ago I bought 4 $100 drives from Seagate, putting them into a pair of RAID-1 arrays in two boxes. They handle error correction badly, sometimes dropping from the array for no good reason. And periodically they go complety crazy due to a firmware bug. Later versions of the drive ship with a fix for the problem. That fix isn't available for the cheapo ones I have. Seagate just doesn't care about whether someone with a $100 drive works reliably or not. (I've had the same class of problem with Hitachi and Western Digital, just picking on Seagate as the most recent one)
If this were a Netapp system with a support contract I'd expect both help tracking down this weird problem and a real resolution at the end. I would also expect they would only ship hardware that goes through more QA for the purpose of going into a RAID array than a random consumer drive does. The drives I have were for example dropped from the HCL of the Netgear ReadyNAS I own, due to this bug in them. Their support documents are what helped lead me to the source of the bug. I've been on phone calls with very sharp Netapp staff tracking down odd problems before. It was nice to have that level of storage expert available in addition to what I knew.
Are these things worth the price premium? That part is debatable. Claiming that they have zero value is really ignorant though. The more you know about hard drives, the more you appreciate that they are not simple works/doesn't work devices. Building reliable systems out of multiple drives is complicated and has a number of subtle failure possibilities. And individual consumer drives are normally not optimized for that very well. WD's Red line is starting to show some promise for good error handling in RAID, but again you're back to only a small number of drives relative to the mass of mostly funky hardware on the market.
It's not just about pandering to casual gamers. The Wii did that partly through having an innovative and interesting input device. I see the success of the Wii as a lucky accident that way. Who would have expected I'd find old people playing Wii bowling? That it was possible for people who weren't into gaming at all to use a Wiimote. There's nothing uniquely interesting about the Wii U that way though.
The Wii was the right technology at the right price for its marketplace. They've tried to duplicate that but with nothing innovative this time. They're not going to get lucky and happen to pick up a whole new market (the casual gamer then) this time.
The problem is that advertisers are like cockroaches; you can't kill them with nukes. When all of civilization has been reduced to a post-apocalyptic wasteland, and mutant zombies roam the land, there will still be someone trying to sell you that one weird trick for losing belly fat.
It makes me sad that someone could run up a $12K monthly electric bill without assigning an environmental cost to where that power was coming from.
I used those specific words to be funny, but there's truth behind them too. I have my bank account setup such that an overdraft mistake will hit a credit card instead of bouncing. And I sometimes expect I can charge things in advance of the cash being available. When the credit card I used for PSN was stolen and used by thieves, both mechanisms weren't available until the fraud was cleared up. It didn't actually cost me anything directly, but control over my normal finances was taken from me by someone else for a while. When someone hits you with a denial of service attack, they have asserted a form of control over a resource.
Sony is reponsible for their security issues because their storage measures for personal data were pathetic. What the fuck were they doing saving so much unencrypted data? The issue is not just that they were breached, which does happen to lots of companies; it was that they obviously didn't care at all about securing things until they were.
As for the rest of your speculation, the credit card I used for PSN purchases was used to make $3000 of purchases shortly after the breach. Are those two connected? I can't be sure, but the Sony incident was the only such breach in the time period before then. It didn't cost me anything, but I was definitely unable to use that card until a replacement showed up. If I couldn't use it, that means I lost control of the account to another party. Am I supposed to excuse Sony for that because the thieves didn't manage to steal all of my identity? No.
Don't call me a sheep when you're posting your rambling and poorly written bullshit as an AC. Your apologist fanboy crap is pathetic.
The PS3 already allowed non-friends to take control of my bank account.
If the free market value for the property was $250K, there would be multiple bidders willing to pay that amount. This case is more like extortion than it is a market. "We gots this domain with your name on it...sure would be a shame if we said bad things about you there". It is domain squatting bullshit, and they're lucky the penalty is only losing the domain. This isn't sex.com or something that has some inherent market value to multiple buyers. It's only worth anything because of the association of Ron Paul, the politician, and it's only worth >$0 to him alone.
The sellers are playing their own semantic game, where they've only offered the sale of the domain plus their mailing list, for a large sum based on the list value. They only did that because the market value of the domain itself, the only thing Paul was interested in, is zero. They know it's not worth anything because they're bound by the domain agreement, which says you can't do the thing they did. This is more like a contract dispute than it is government involvement. The distinction is that you don't have to buy a .com domain and agree to its terms. But if you choose to do so, you then have to follow the rules you agreed to.
You'd lose that bet. The rate of future I/O operations is extremely difficult to work out at any price, with any quality of engineer. I would bet you a billion dollars it can't be done reliably in the case of HTTP downloads, and not lose a moment of sleep about losing. If you spent all of the money hiring the best developers in the world with that billion, all you'd have when you were done is a better understanding that it's impossible. In fact, I'd make a side-bet that someone tells you it's impossible before they even start!
I would have stopped at "yes" without the optimism, and it's not even a platform/processor issue. This is a fundamentally unsolvable thing, since in some cases it reduces to the halting problem. That best you can do is come up with an approximation.
But the approximations will be wrong too. Simple example: write speed. If you write a block to disk, and it fits into the operating system cache, a write happens almost instantly. If the cache is full, you might have to spend tens of seconds waiting before that write occurs. Here's what happens with installers, copy programs, and a large chunk of other things. You write until the cache fills, and those happen at memory speeds. Then, all of the sudden, you grind to a halt when the cache fills. You won't see any progress, sometimes for minutes--it can take a while to chug through gigabytes of random writes. I show an example of this on my blog on Linux, and this problem gets worse as memory increases, not better.
What does that look like to the user? They get a progress bar, it zooms along for a while, and then it cranks to a halt. Then it hangs for a bit, starts moving, and the whole thing completely changes scale. Does this sound familiar? That's what people complain about, right? You can't make that go away without building a model of the caching mechanism that's more accurate than the cache itself. After all, if you could predict this was coming, the OS could have done a better job scheduling I/O with that information, too. Think about that for a minute: to write a really good progress bar for write operations, you have to do a better job on I/O scheduling than Linux does.
If you step back and say "well let's approximate how long disk I/O takes then and base the bar on that", you'll discover that doesn't work either. There's over a 100:1 difference between the fastest and slowest storage on the market. Good luck modeling that accurately enough to predict the future, too.
A 2 hour movie can be fit into 9GB, because that's how much a single DVD can hold. 5 TB lets you transfer every byte of data for over 500 movies per month. And even if you're playing them 24x7, you can only fit 360 2-hour movies into a month.
You cannot overrun 5 TB with DVD quality playback unless you bloat the broadcast excessively. It sounds like your movie server doesn't care about how much bandwidth it uses. That's fine when it's free. If it isn't, you can do much better, and 5 TB is plenty for a month of standard definition video.
If the short-term math makes sense, he can deploy the home server and then see what happens. When the replacement parts get expensive, you can always switch to the VPS at that point. The short-term spending is a $100 upgrade to save about $20/month. The only lifecycle question worth asking is "is the system likely to last 6 months?" If it does, you can bank the savings until the home server dies, and then figure out what to do.
This sort of thinking is not an option is not available to business class service, because it assumes a day or two of downtime is acceptable. At home, you can do the "let's see how long this server lasts" game. As long as you're willing to jump to another option when you're facing another "capital expense", you can usually find a cheaper long term path.
Not necessarily. True story: I once had "Greg Smith Equipment Sales" try to buy my web site because I was messing with their search rank. What really killed both of us was a baseball player with that name though. Once ESPN started writing about him I was done with being in the top 3. I wouldn't recommend just any common name; what you really want is the name of a celebrity. I use "Michael Bolton" now.
I'm not worried, my geocities forums will always be there.
The only embarrassing things I did online 15 years ago were posts to Usenet that seem a bit immature now. I think it was easier then to see that things you posted went out of your control and might be around forever though, because that distribution was itself sometimes a technical problem to be solved.
I've had the same e-mail address for over 15 years though. For the majority of the login/password combinations I created, I outlasted the site I created it on. I'm not exactly worried that I can't provide the credentials for the gopher account I signed up for once in the early 90's.