Actually, you may be on to something. WinME could be beaten into 100% stability, but it *couldn't* be taught not to squander the resource heap -- so it really wasn't practical as a multitasking OS. And I've noticed that WinXP doesn't multitask as smoothly as Win95/98. (At the moment I don't run Win2K as an everyday OS, so can't comment on how it compares here, but I don't remember it as having any special issues in this regard.) XPHome is worse about it than XPPro, and I wonder if it's a side effect of having some of the networking code pulled out by the roots. Yanking out even more code may have introduced issues that make it unstable under heavy multitasking.
A three app limit does seem really lame, tho -- especially since running a decent two-way firewall is going to eat one of them. I know a lot of users have this bad habit of closing and restarting apps every time they want to switch from one to another, but why forcibly entrench the habit of beginner behaviour -- unless it's to ensure the new user such a miserable experience that they upgrade to a real multitasking OS out of sheer frustration. If M$ hopes to get 'em to pony up for XPHome or Pro, I think they're in for a major market LARTing.
I just skimmed the transcription, but I got the impression that he knows all too well how free software works, and has worked out exactly what FUD will best besmirch the entire concept. Frex, he pointed out how spyware, malware, and the like are also "free software", with the implication that "if it's free, it's probably tainted".
Yep, far as I remember (I RTFA yesterday, and by now it's fallen out of my brain:) Progressive's notion didn't seem out of line -- geared only toward penalizing the habitually reckless, such as those who speed well beyond the local norm on a daily basis. (Obviously, *some* "over limit" driving has to be allowed, if only for passing on two-lane roads.)
Say, have you noticed that we trust our fellow men with our very lives?? -- Don't think so? You drive on two-lane roads, don't you?:)
As to big-brotherism, that IS out of line; tracking so precise as to know when you're speeding on a localized basis is also sufficient for anyone who can tap into it to follow your movements whenever they wish. Even tho my tinfoil hat was recently refitted, I can still see how it's a short step from that sort of monitoring by insurance companies, to gov't requiring that the data be turned over for use in "criminal and terrorist investigations", or in the case of a more overtly repressive gov't, for harrassment and suppression of dissenting groups.
Yeah, that's one possibility (and I'd sure be thinking along the lines of "shortsighted local agenda" anywhere Beverly Hills or Burbank are involved!), tho all those examples I've seen are within a single city, or are rural. But zoning is another matter -- frex, residential zoning might allow one max speed limit, while business zoning might allow a different max, and each side of the street can be in another zoning district. I think that's what the CHP dude was getting at, tho he sounded a bit confused himself. Anyway, this sure makes it hard to remember what the speed limit is on some streets, especially with signs buried in the trees often as not so you're liable to miss them.
Southern California has a lot of roads where the speed limit differs depending on which direction you're going, I shit you not (frex, the speed limit northbound might be 40mph, and southbound *on the same street* might be 50mph). When I asked the CHP about this, they said something to the effect that it's a zoning thing.
Also, some areas have different speed limits and lane restrictions depending on the time of day.
Raw GPS position isn't enough; the direction of travel has to be taken into account. It would need to be so specific as to know all the posted speed limits AND all the odd quirks such as the above.
It's not always that easy. Here in north L.A. county, the local police and fire depts. close up shop at 5pm, so if I can't get 911, I have to call their downtown Los Angeles centers, and that number may or may not answer (in my experience, it doesn't always).
That's an interesting insight. Against a corporation, you can vote with your wallet; only rarely are you left with no alternatives at all. But against legislation that happens beyond your power to prevent, you can only vote with your feet, and it's not all that easy to find a new life somewhere else -- and the evil regulations you're trying to escape may well have preceeded you. Or that more-free country may not want any more immigrants. Etc.
I agree with the parent, tho -- what do you do when you call someone who uses a tapped VoIP system? I guess the first thing you do is ask what phone system they use, then decide how much you need to restrict your conversation.
Terrorists and criminals will simply revert to hand-delivered communications, rendering all this snooping worthless anyway.
Occurs to me that the obvious solution is to run every call against a keyword filter, and discard those that don't trigger the filter. That way there would be far less data to store in the first place (probably 99% of calls would be discarded up front).
This filtering might be doable in realtime, or data could be queued for off-peak processing. Hence at any given time, they'd need storage for no more than a couple days worth of raw data.
Of course, this means that privacy-minded folk will start using encrypted language... which naturally will trigger the filter as "unknown but suspicious".
That reminds me of a tale (reputed true) from way back in the era of the Iron Curtain: The dad still lived in the USSR, but the son had escaped to the West. As was common among the scientist community back then, both were fluent in Latin. Occasionally they'd speak on the phone.. and by the time the Soviet censors found someone who could translate spoken Latin, they'd be done with the political scuttlebutt and would revert to Russian.
Why, yes, I *did* mean to invoke a historical comparison...
Time to equalize the bargaining power... amend your contract to say, "For every month that my employment is restricted by your noncompete clause, you are prohibited from selling any product which I had a hand in developing.";)
Has anyone with good knowledge of the relevant laws worked up a chunk of legal boilerplate that folk so-threatened could use for such situations? Seems to me that would save a lot of worry and acrimony.
The late WebTechniques magazine had a regular legal column that gnawed on such issues. It was frequently the most interesting read in the entire issue, and a good chunk of why WT wasn't consigned to "skim and pitch" status.
[eyes present company's blog, backs up to safe distance, tries again, extracts visual sample for analysis...]
I?m packing my bags for a quick house call to eWEEK.com?s
<?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags"/>San Francisco...Now that's ugly:)
Oh yeah, the nominal topic... glad to know that (as a longtime subscriber) my first thought, "WTF?? *that* doesn't sound like eWeek!!" was correct:)
Sometimes it refuses to acknowledge the new login, tho. When it happens, it'll usually affect one or two subdomains, but not all of slashdot. Very annoying.:(
Back when I was occasionally sticking my hands inside a Mac, all the HDs were Quantums, and IIRC they were SCSIs at that. Boy, do I feel old:)
I've noticed that that every time there's a major new size threshold, *someone* has a spate of bad drives (not necessarily the first made in that size range, either). After the flurry of RMAs dies down, most of what's left are pretty reliable on average.
My clients tend to keep 'em til they show signs of impending death, or until a major upgrade requires a fresh HD. Most aren't accumulating enough data to keep a 4GB HD properly exercised. The oldest HD in continuous service among 'em dates to 1992 or thereabouts.
[eyeing HDs in immediate view] Unlike my own, where there is never enough disk space...
Sounds like you hit W.D. in one of their Ooops releases... there've been a few, just like everyone else, where chances of getting a bad one are seriously high. I've had a few myself.
However... FAT32 partitions over 32GB can experience a data wrapping bug in WinXP (and possibly previous Win32 versions, not sure) -- I've seen that myself, and before I knew what caused the data corruption and vanished files, I'd RMA'd the HD. Later I found info on the issue, and realised it wasn't the HD at all, but a filesystem bug of some sort.:( The bug does not occur on NTFS partitions. -- W.D. had the first widely consumer-available HDs over 20GB, and at the time NTFS wasn't yet used much in home/SOHO systems, so I expect a LOT of perfectly good HDs got RMA'd due to this issue. It really does *look* like the HD has failed, when it happens.
I tend to need 5-6 years out of HDs, either for myself or for clients, and Maxtors just don't cut it for that. Most that have come my way were already dead (in fact, only in the past year have I seen *any* live Maxtors from salvage, but 50% of the Maxtors in that heap were dead and some of the rest are suspect), but I know a guy who has run into the "sudden death" issue over and over while doing business networking support.:(
Conversely, I've yet to see a dead W.D. in salvaged boxes, and some have 7-8 years of heavy use behind 'em.
In my observation, used HDs *offered for sale* have essentially zero reliability regardless of brand or type -- I've yet to see one last more than 90 days (whether IDE or SCSI). I suspect a lot of this is due to rough handling by used-HD brokers, because I have piles of far-older *salvaged* HDs that work fine and show no sign of impending death.
I don't buy Maxtor HDs myself, due to learning from others' mistakes.:) In my experience, W.D. are most likely to have a long lifespan (and concomitantly, are least likely to die young), followed by Seagate. Maxtor is at the other end of the scale. Everyone has occasional bad batches, but they have 'em all the time:/
BTW same observation about used CDROMs/CDRWs and their kin -- I've yet to see a used unit *offered for sale* that had any reasonable lifespan; random salvage is a better bet (tho they're still more likely to be dead than are older HDs).
Conversely, I've been buying used components -- video, sound, SCSI, NIC, and some mainboards -- for years, with close to 100% being good from any source. But they're a lot tougher than HDs.
The smallest HD here that still sees some use is a lowly 850mb bearing Win98, which I use when I need a quick OS to check out some pile of random hardware. Why subject a newer drive, with realworld value, to being thumped around all the time, when I can use an old one of little worth, and not get too upset if I happen to accidentally drop and kill it? -- Otherwise, I figure 1GB and up are still useful for testing, for systems that are expected to only do limited jobs, etc. Why lay out for a new HD for that firewall or test box, when an old HD from the salvage box will do just as well?
I drag home a lot of salvage, including HDs of every size, age, and type, some of which naturally prove dead. So I have piles of IDE HDs here dating from 1991 to 2004. In my experience, the vast majority of long-survival HDs are Western Digitals. Some are Seagates (except for their spate of rebadged Conners, which are found pretty uniformly sick or dead). A few are Quantums. And Maxtors are the *least* likely to still be alive.
Every manufacturer has the occasional bad batch, even W.D. has suffered from this. But you're right -- Maxtors do have a habit of dying shortly after the warranty runs out, and it's too consistent to be merely an occasional bad production run.
The other issue is that when a Maxtor dies, it typically just QUITS without warning. Conversely, W.D. usually give you plenty of warning (complaints from "S.M.A.R.T.", funny noises, or the odd bad sector).
My main issue with Seagate over the years is that they've been so much slower than concurrent W.D., and run quite a bit hotter/louder (tho I can't speak for their newest models). Otherwise, for reliability Seagate would likely be my next choice. Whereas I don't buy Maxtors at all.
My own experience is that among generally good-quality HDs (excluding certain ones that do tend to die after a year or two), the typical HD running 24/7 either dies within 30-60 days due to a factory defect, or else lasts a bit over 5 years in perfect condition (but becomes at risk to die without notice after that). Likewise, I've found if they last 3 years, they're gonna last 5 years.
A five year warranty gives me confidence that the manufacturer *expects* what I consider a *normal* lifespan from their HDs. And frankly I would much rather have any that plan to die, get it over with ASAP rather than waiting a year and giving me a false sense of security, so to speak.
So I agree with you entirely... woulda modded you up instead of playing echo (have points today), but stupid slashdot has developed a habit of logging me out on half the comment pages, including this one (and ignoring subsequent login attempts until I make a post):(
Ah, you mean for the.NET *app*. What about the.NET runtime engine itself? that's the part that concerns me, having had some truly evil experiences on Win9* with both.MSI installers and with IE versions later than 5.0.
Personally, I *like* apps I can just drag from one machine to the next... Frex, this here copy of Netscape was installed once 6-7 years ago and never again, and just gets dumped onto the next machine when I need it. It's convenient and it stays configured how I want it, without effort.
I don't remember hotmail's earliest incarnation, but I've had an account there since 1998, and have seen it go from nicely usable to crap, and have since relegated the account to the status of spamtrap.:(
Someone gift me an invite, so I've been fiddling some with GMail... found and reported 3 bugs, at least one of which has since been fixed. It'll be more useful to me for real mail once they have a plain HTML interface available.
Yeah, but how WELL does it run on pre-XP Windows? My expectation is that, given its relative modernity, it's liable to be to some degree ill-mannered on any Windows that doesn't play 100% nice with the.MSI installer, which means 98SE and before (not sure about W2K, tho ME should be okay). [I've noticed a general change in behaviour as of the advent of.MSI, so tend to use it as a dividing point.]
But... I see "Pop Goes the GMail" (206k) requires the.NET runtimes (24mb), which presumably are not going to work on older Windows, let alone linux, not to mention being a painful download on dialup. Does any such utility exist compiled as a standalone, either for Win32 or *NIX?
Slashdot Mirror
What buggy crap -- you forgot the "AND CATCH FIRE" part of the instruction!!
Actually, you may be on to something. WinME could be beaten into 100% stability, but it *couldn't* be taught not to squander the resource heap -- so it really wasn't practical as a multitasking OS. And I've noticed that WinXP doesn't multitask as smoothly as Win95/98. (At the moment I don't run Win2K as an everyday OS, so can't comment on how it compares here, but I don't remember it as having any special issues in this regard.) XPHome is worse about it than XPPro, and I wonder if it's a side effect of having some of the networking code pulled out by the roots. Yanking out even more code may have introduced issues that make it unstable under heavy multitasking.
A three app limit does seem really lame, tho -- especially since running a decent two-way firewall is going to eat one of them. I know a lot of users have this bad habit of closing and restarting apps every time they want to switch from one to another, but why forcibly entrench the habit of beginner behaviour -- unless it's to ensure the new user such a miserable experience that they upgrade to a real multitasking OS out of sheer frustration. If M$ hopes to get 'em to pony up for XPHome or Pro, I think they're in for a major market LARTing.
Nah, I think this Rob Enderle guy is probably the Giver. ;)
I just skimmed the transcription, but I got the impression that he knows all too well how free software works, and has worked out exactly what FUD will best besmirch the entire concept. Frex, he pointed out how spyware, malware, and the like are also "free software", with the implication that "if it's free, it's probably tainted".
I'd say he's a very skilled troll.
Yep, far as I remember (I RTFA yesterday, and by now it's fallen out of my brain :) Progressive's notion didn't seem out of line -- geared only toward penalizing the habitually reckless, such as those who speed well beyond the local norm on a daily basis. (Obviously, *some* "over limit" driving has to be allowed, if only for passing on two-lane roads.)
:)
Say, have you noticed that we trust our fellow men with our very lives?? -- Don't think so? You drive on two-lane roads, don't you?
As to big-brotherism, that IS out of line; tracking so precise as to know when you're speeding on a localized basis is also sufficient for anyone who can tap into it to follow your movements whenever they wish. Even tho my tinfoil hat was recently refitted, I can still see how it's a short step from that sort of monitoring by insurance companies, to gov't requiring that the data be turned over for use in "criminal and terrorist investigations", or in the case of a more overtly repressive gov't, for harrassment and suppression of dissenting groups.
Yeah, that's one possibility (and I'd sure be thinking along the lines of "shortsighted local agenda" anywhere Beverly Hills or Burbank are involved!), tho all those examples I've seen are within a single city, or are rural. But zoning is another matter -- frex, residential zoning might allow one max speed limit, while business zoning might allow a different max, and each side of the street can be in another zoning district. I think that's what the CHP dude was getting at, tho he sounded a bit confused himself. Anyway, this sure makes it hard to remember what the speed limit is on some streets, especially with signs buried in the trees often as not so you're liable to miss them.
Southern California has a lot of roads where the speed limit differs depending on which direction you're going, I shit you not (frex, the speed limit northbound might be 40mph, and southbound *on the same street* might be 50mph). When I asked the CHP about this, they said something to the effect that it's a zoning thing.
Also, some areas have different speed limits and lane restrictions depending on the time of day.
Raw GPS position isn't enough; the direction of travel has to be taken into account. It would need to be so specific as to know all the posted speed limits AND all the odd quirks such as the above.
It's not always that easy. Here in north L.A. county, the local police and fire depts. close up shop at 5pm, so if I can't get 911, I have to call their downtown Los Angeles centers, and that number may or may not answer (in my experience, it doesn't always).
That's an interesting insight. Against a corporation, you can vote with your wallet; only rarely are you left with no alternatives at all. But against legislation that happens beyond your power to prevent, you can only vote with your feet, and it's not all that easy to find a new life somewhere else -- and the evil regulations you're trying to escape may well have preceeded you. Or that more-free country may not want any more immigrants. Etc.
I agree with the parent, tho -- what do you do when you call someone who uses a tapped VoIP system? I guess the first thing you do is ask what phone system they use, then decide how much you need to restrict your conversation.
Terrorists and criminals will simply revert to hand-delivered communications, rendering all this snooping worthless anyway.
Occurs to me that the obvious solution is to run every call against a keyword filter, and discard those that don't trigger the filter. That way there would be far less data to store in the first place (probably 99% of calls would be discarded up front).
... which naturally will trigger the filter as "unknown but suspicious".
This filtering might be doable in realtime, or data could be queued for off-peak processing. Hence at any given time, they'd need storage for no more than a couple days worth of raw data.
Of course, this means that privacy-minded folk will start using encrypted language
That reminds me of a tale (reputed true) from way back in the era of the Iron Curtain: The dad still lived in the USSR, but the son had escaped to the West. As was common among the scientist community back then, both were fluent in Latin. Occasionally they'd speak on the phone.. and by the time the Soviet censors found someone who could translate spoken Latin, they'd be done with the political scuttlebutt and would revert to Russian.
Why, yes, I *did* mean to invoke a historical comparison...
Time to equalize the bargaining power ... amend your contract to say, "For every month that my employment is restricted by your noncompete clause, you are prohibited from selling any product which I had a hand in developing." ;)
Has anyone with good knowledge of the relevant laws worked up a chunk of legal boilerplate that folk so-threatened could use for such situations? Seems to me that would save a lot of worry and acrimony.
The late WebTechniques magazine had a regular legal column that gnawed on such issues. It was frequently the most interesting read in the entire issue, and a good chunk of why WT wasn't consigned to "skim and pitch" status.
/>San Francisco ...Now that's ugly :)
:)
[eyes present company's blog, backs up to safe distance, tries again, extracts visual sample for analysis...]
I?m packing my bags for a quick house call to eWEEK.com?s
<?xml:namespace prefix = st1 ns =
"urn:schemas-microsoft-com:office:smarttags"
Oh yeah, the nominal topic... glad to know that (as a longtime subscriber) my first thought, "WTF?? *that* doesn't sound like eWeek!!" was correct
Sometimes it refuses to acknowledge the new login, tho. When it happens, it'll usually affect one or two subdomains, but not all of slashdot. Very annoying. :(
Back when I was occasionally sticking my hands inside a Mac, all the HDs were Quantums, and IIRC they were SCSIs at that. Boy, do I feel old :)
I've noticed that that every time there's a major new size threshold, *someone* has a spate of bad drives (not necessarily the first made in that size range, either). After the flurry of RMAs dies down, most of what's left are pretty reliable on average.
My clients tend to keep 'em til they show signs of impending death, or until a major upgrade requires a fresh HD. Most aren't accumulating enough data to keep a 4GB HD properly exercised. The oldest HD in continuous service among 'em dates to 1992 or thereabouts.
[eyeing HDs in immediate view] Unlike my own, where there is never enough disk space...
Sounds like you hit W.D. in one of their Ooops releases... there've been a few, just like everyone else, where chances of getting a bad one are seriously high. I've had a few myself.
:( The bug does not occur on NTFS partitions. -- W.D. had the first widely consumer-available HDs over 20GB, and at the time NTFS wasn't yet used much in home/SOHO systems, so I expect a LOT of perfectly good HDs got RMA'd due to this issue. It really does *look* like the HD has failed, when it happens.
:(
However... FAT32 partitions over 32GB can experience a data wrapping bug in WinXP (and possibly previous Win32 versions, not sure) -- I've seen that myself, and before I knew what caused the data corruption and vanished files, I'd RMA'd the HD. Later I found info on the issue, and realised it wasn't the HD at all, but a filesystem bug of some sort.
I tend to need 5-6 years out of HDs, either for myself or for clients, and Maxtors just don't cut it for that. Most that have come my way were already dead (in fact, only in the past year have I seen *any* live Maxtors from salvage, but 50% of the Maxtors in that heap were dead and some of the rest are suspect), but I know a guy who has run into the "sudden death" issue over and over while doing business networking support.
Conversely, I've yet to see a dead W.D. in salvaged boxes, and some have 7-8 years of heavy use behind 'em.
In my observation, used HDs *offered for sale* have essentially zero reliability regardless of brand or type -- I've yet to see one last more than 90 days (whether IDE or SCSI). I suspect a lot of this is due to rough handling by used-HD brokers, because I have piles of far-older *salvaged* HDs that work fine and show no sign of impending death.
:) In my experience, W.D. are most likely to have a long lifespan (and concomitantly, are least likely to die young), followed by Seagate. Maxtor is at the other end of the scale. Everyone has occasional bad batches, but they have 'em all the time :/
I don't buy Maxtor HDs myself, due to learning from others' mistakes.
BTW same observation about used CDROMs/CDRWs and their kin -- I've yet to see a used unit *offered for sale* that had any reasonable lifespan; random salvage is a better bet (tho they're still more likely to be dead than are older HDs).
Conversely, I've been buying used components -- video, sound, SCSI, NIC, and some mainboards -- for years, with close to 100% being good from any source. But they're a lot tougher than HDs.
The smallest HD here that still sees some use is a lowly 850mb bearing Win98, which I use when I need a quick OS to check out some pile of random hardware. Why subject a newer drive, with realworld value, to being thumped around all the time, when I can use an old one of little worth, and not get too upset if I happen to accidentally drop and kill it? -- Otherwise, I figure 1GB and up are still useful for testing, for systems that are expected to only do limited jobs, etc. Why lay out for a new HD for that firewall or test box, when an old HD from the salvage box will do just as well?
I drag home a lot of salvage, including HDs of every size, age, and type, some of which naturally prove dead. So I have piles of IDE HDs here dating from 1991 to 2004. In my experience, the vast majority of long-survival HDs are Western Digitals. Some are Seagates (except for their spate of rebadged Conners, which are found pretty uniformly sick or dead). A few are Quantums. And Maxtors are the *least* likely to still be alive.
Every manufacturer has the occasional bad batch, even W.D. has suffered from this. But you're right -- Maxtors do have a habit of dying shortly after the warranty runs out, and it's too consistent to be merely an occasional bad production run.
The other issue is that when a Maxtor dies, it typically just QUITS without warning. Conversely, W.D. usually give you plenty of warning (complaints from "S.M.A.R.T.", funny noises, or the odd bad sector).
My main issue with Seagate over the years is that they've been so much slower than concurrent W.D., and run quite a bit hotter/louder (tho I can't speak for their newest models). Otherwise, for reliability Seagate would likely be my next choice. Whereas I don't buy Maxtors at all.
My own experience is that among generally good-quality HDs (excluding certain ones that do tend to die after a year or two), the typical HD running 24/7 either dies within 30-60 days due to a factory defect, or else lasts a bit over 5 years in perfect condition (but becomes at risk to die without notice after that). Likewise, I've found if they last 3 years, they're gonna last 5 years.
:(
A five year warranty gives me confidence that the manufacturer *expects* what I consider a *normal* lifespan from their HDs. And frankly I would much rather have any that plan to die, get it over with ASAP rather than waiting a year and giving me a false sense of security, so to speak.
So I agree with you entirely... woulda modded you up instead of playing echo (have points today), but stupid slashdot has developed a habit of logging me out on half the comment pages, including this one (and ignoring subsequent login attempts until I make a post)
Ah, you mean for the .NET *app*. What about the .NET runtime engine itself? that's the part that concerns me, having had some truly evil experiences on Win9* with both .MSI installers and with IE versions later than 5.0.
Personally, I *like* apps I can just drag from one machine to the next... Frex, this here copy of Netscape was installed once 6-7 years ago and never again, and just gets dumped onto the next machine when I need it. It's convenient and it stays configured how I want it, without effort.
I don't remember hotmail's earliest incarnation, but I've had an account there since 1998, and have seen it go from nicely usable to crap, and have since relegated the account to the status of spamtrap. :(
Someone gift me an invite, so I've been fiddling some with GMail... found and reported 3 bugs, at least one of which has since been fixed. It'll be more useful to me for real mail once they have a plain HTML interface available.
Yeah, but how WELL does it run on pre-XP Windows? My expectation is that, given its relative modernity, it's liable to be to some degree ill-mannered on any Windows that doesn't play 100% nice with the .MSI installer, which means 98SE and before (not sure about W2K, tho ME should be okay). [I've noticed a general change in behaviour as of the advent of .MSI, so tend to use it as a dividing point.]
.... no shit!
Tho I still have clients on Win3.1
[laughing] That's one weirdly ironic post. :)
But... I see "Pop Goes the GMail" (206k) requires the .NET runtimes (24mb), which presumably are not going to work on older Windows, let alone linux, not to mention being a painful download on dialup. Does any such utility exist compiled as a standalone, either for Win32 or *NIX?