Personally, I wouldn't - I'd just buy the cheap crap one if there was a choice, unless the price on the better one was sufficiently reasonable.
That was intended to be part of my point - some things are worth paying top dollar for, some aren't. Perhaps I failed to make said point - it was 4:00am when I posted that after all.
Only for work. Please look in my comment history for discussion on that matter, including why I'm not interested in conducting it on Slashdot - especially with an AC.
Agreed. That's exactly why I said that while I'll pay for quality in some things, computer equipment is not one of those things. Not for personal use, anyway - work is different because the cost of failure is more than "Bugger, off to the PC shop again tomorrow."
I'm not sure it's possible to make a modern console for US$300 that'll work like your old Atari.
Modern consoles are a _lot_ more complex, unfortunately. Some, like the xbox, also borrow from PC tech - including the "cheap and nasty" aspect.
Of course, I suppose the Atari was pretty cutting edge and complex for the time too, so perhaps I'm just talking hot air.
While it would be interesting to see an "Xbox Premium" - think good quality SCSI disk set to operate at lower RPMs, quality DVD-ROM, faster CPU running under spec to keep it cool and quiet, etc - I doubt many would buy it. Almost certainly not enough to cover the costs of making the two different variants.
Worse, MS couldn't form a viable market around a quality console - volume and market share are king, and if you want those you _must_ win on price.
As for PATA HDDs... you lucky bastard. I use SATA disks in a RAID array at work for our main storage array (fairly static "live" archival data, backed up). I had a 60% failure rate on the first batch. When I recently upgraded the array to 250GB disks, I again had a 60% failure rate. I lost three disks in the first two weeks. I've also had less than stellar results at home. I've since discovered that the first round was a bad batch of disks - the three with sequential serial no#s died, the other two were fine, and that batch was confirmed to be bad. The more recent round... nfi.
All the disks were well cooled and carefully installed. All were confirmed to have serious media failures using S.M.A.R.T and mfgr disk tools.
With a failure rate like this, we'd be better off spending 4-5x as much to get something like WD Raptors. Of course, I'll probably use consumer SATA again next time because I won't be given the budget for anything else. Grr.
Bike in this case == pushbike , not motorcycle. Sorry.
Possibly language differences AU <> US<> UK ?
As for how it can fail... manufacturing problems mostly if the frame is good. Lower quality frames skimp on materials and use cheaper construction techniques that can make them weaker than one would expect and more prone to failure.
Having snapped my first bike's frame (the rear wheel attachment point fell off - a clean break in the metal on both sides - while accelerating across an intersection), I care about this stuff. Also, given the option I like to buy quality.
In addition to all this, a better frame tends to be lighter - which is always nice.
Unfortunately, that model doesn't seem to actually work. The reason everything used to be well made and often sold with lifetime warranties - but rarely is now - is that people started buying the cheaper option no matter what.
People buy on price. So long as that persists, there's only a small market for quality products.
Would you really pay probably five times as much for a system you can expect to run vastly more reliably and run for, say, ten years? If so, would you, by the end of the second year why you didn't buy the cheap one?
I do agree that it's gone too far (6 month HDD warranties and 90 day warranties on products that cost several hundered - WTF?!) but usually where it makes sense you have the option of a decent warranty. You also, again where it's practical to make one, usually have the option of a better made product that will last longer.
So tell me, do you have enterprise class SCSI or SATA drives in your PC? Or did you buy the cheap 120GB PATA ones?
I'll put my money where my mouth is when I care. I just bought a very nice bike with a lifetime warranty on the frame - and don't ever expect to have to call in the warranty. It doesn't cost that much more to make a bike frame _much_ better, and it's worth the premium.
On the other hand, if someone told me SUN were making dual Opteron workstations with a lifetime warranty (and build quality such that SUN didn't expect it to be used much) so why don't I get one - it's "only" $10,000 - I'd laugh in their face. To me, my PC dying is acceptable compared to the price tradeoff I get in exchange for that risk - I have a functional brain, I keep backups and know how to recover quickly so it's not that big a deal.
I often wish it was possible to buy better quality products, and am increasingly irritated by the tendency of quality to equate with overpriced wank - to pay for quality, often you're forced to pay for stupid wank factor too. Sometimes you just can't seem to get quality (I've been through three kettles recently, and have just given up hope of finding one that doesn't suck). Still, indestructable and expensive isn't for every situation, much as crap with a warranty or cheap with no warranty are both also unsuitable for many situations.
I really don't see a stable driver API happening, though. One can write a stable API in C as well as C++ - but the kernel folks explicitly choose not to. I can understand why, though I won't get into the endless arguments about the merits of the decision.
Also, regarding the difficulty of getting drivers - it's not, to me, a matter of having them easy to install. It's that they're/already there/. No 'net access required, no stuffing about with boot-time driver disks, it just works. I suppose one could provide tools to download and build all the drivers, but I just don't see the attraction vs the current scheme while the drivers are so closely bound to the kernel core.
Seriously. The fact that one can simply build and install a Linux kernel and _already_ have all the drivers required is IMO very important.
I really don't like the idea of having to go online and download drivers, and if the drivers were availible as another bundled download, really what would be the point?
Anyway, I don't really see how separating drivers from the kernel core could make sense given the Linux development policy on stable driver APIs (that is, there aren't any).
Why not simply use the `nv' driver and not use a 3D screensaver? Reliability is excellent. If you need 3D, the card this guy was talking about wouldn't do you any good anyway.
On the other hand, I use the `nvidia' driver on my home PC and find it 100% dead reliable as well, so I guess results vary.
The guy who proposed this was not talking about a high-performance card - or a card with any 3D acceleration.
Remember the S3 Trio32 and Trio64? It sounds like it'll be something like that - though presumably with more video RAM, a faster DAC or DVI, and better 2D acceleration. Maybe more like a FreeG450;-)
A cleanroom rewrite probably wouldn't do it. It's entirely likely that they have other people's/patented/ processes in use without license. Worse, they probably don't even know what they might be infringing on - the joy of software patents being what they are.
If they don't release their code, it arguably makes it harder for patent litigation companies (or competitors) to go trawling.
Additionally, there's a good chance they've licensed other people's code for use as well, and they'd have to re-implement that before releasing such a driver.
That's a lot of work, and a lot of risk, for what would be little gain for them. Much as it'd be nice to assume that the "open source community" would help maintain and improve the drivers, I'm not convinced that would be the case for these specialised extreme-performance bits of software that are closely tied to rapidly changing hardware.
There's also the possibility that their competititors could learn more about their drivers, and their hardware, from the code than from the disassembly of the drivers that they no doubt already do. One would certainly hope so, since code that's no easier to study than code generated from disassembled object code doesn't strike me as fun to work with.
For these reasons, I just don't see it happening. The Linux and BSD user market is small, and the percentage of that market with money and strong opinions about open source or free software is smaller. The proportion with the strong opinions and the willingless to being their wallet out to back those opinions is probably a lot smaller still.
Right now, it's probably much more worth their while to mainain the drivers in house and track whatever the kernel folks break today;-)
Re:EC over IP I have been doing this for years.
on
Replacing TCP?
·
· Score: 1
"it was a 64K shared line with 90% packet loss, I received 60Kbps for the video stream. "
That makes no sense. You cannot reconstruct lost data out of nowhere. Using an ECC scheme works by sacrificing peak _normal_ bandwidth in exchange for error tolerance. With an ECC scheme capable of withstanding 90% packet loss you can't fit a 60kbps video through a 64kbps link even at zero packet loss. For that matter, it wouldn't be easy to do with TCP/IP, either.
Are you sure you don't mean a 64kilobit/second video through a 64kilobyte/second (~512k) link? That verges on possible, though it's still really pushing your luck at 90% loss. Alternately, I could understand it being vaguely possible if you mean 60KB/s _after_ encoding with the ECC data, though again that'd be a bit of a stretch IMO.
Re:Not gonna work if encumbered
on
Replacing TCP?
·
· Score: 1
They may simply mean that they consider the GPL non-commercial by its very nature, that is that it is not generally possible to deeply embed GPL code in a larger product. If your view of commercial use is limited to embedding something in a closed source product, I suppose that makes sense.
Of course, if it's based on a GPL product they'll have a rather harder time offering "commercial" (presumably conventional embed in a closed source for a fee style) licenses than they might be expecting;-)
Aggressive quarantining is great. I actually quarantine zip files - there's a small false positive rate, but the work caused retrieving the occasional quarantined file is minimal compared to the time and effort saved by blocking them.
I'm lucky enough to have most of my users either smart enough not to open suspicious attachments, or on computers where it just doesn't matter. The LTSP users, obviously, can do whatever they want - though they're encouraged to follow sensible security rules anyway, both at home and at work. Ditto the MacOS 9 users. For the win98 and XP users, it looks like the message got through on something like the eighteenth try for most of them, and they now tend to act fairly sensibly.
Given that I have the class of user who will call me and say "my email is broken" when somebody (one person) said they sent a message five minutes ago and it hasn't arrived yet, I wonder at your organisation's choice of staff;-)
I don't presently even run a proxy server. The business has perfectly good EXISTING ways of detecting abuse by staff, much the same as for excessive personal phone calls.
To date it simply has not been an issue - folks will use 'net banking, sometimes do a bit of research, or look up something they've been sent (yeah, security, I know...). If they get their jobs done well, nobody cares.
That said, I'm not 100% against filters, I just think they should be used appropriately and after thinking about what the real problem is. An administrator who uses filters is not automatically incompitent (and in fact is more likely to have been forced into using some product by management).
I'm also unconvinced that unrestricted 'net access is good for productivity. Some staff really just don't need it. That said, at my work we've decided to simply accept that, expect them to behave responsibly, and deal with it though normal means if they don't.
The show is utter drivel. What frightens me is that everything else is moving down to its level (not that most TV content had far to fall). TV "news" is worse every year, we get "reality" TV now, etc.
Given how bad even the sections of the media that at least pretend to have some form of integrity are, such as the papers, are, I decided several years back that I had better things to do than watch that drek. (Of course, this argument is somewhat undermined by being in the process of reading SlashDot, but anyway...).
Even the ABC (different to USA; ABC is the gov't funded TV channel) news is pretty terrible these days.
It seems that any form of depth, coverage of both sides of an issue, or even *gasp* admitting that there's no "right" answer is just forbidden by the TV media. Yay.
One small point: If you're "disgusted with the crappy tv we have to put up with" it may be worth recognising that you do not, in fact, have to put up with it. You're perfectly free not to watch it.
There will be some, or a lot, of PPC ASM used in performance critical parts of Aqua. That'll need rewriting. Remember that Darwin doesn't support any of the OS/X GUI stuff - certainly not on x86.
They would also have to address driver differences, especially with video. Remember that Apple uses video cards with custom BIOSes.
That said, I'd be a little surprised if there wasn't a skunkworks project at Apple to build a version of OS/X on x86 that could at least stumble along.
If OS/X is just a souped-up version of NextSTEP, then Win2k3 is probably "just" a souped-up version of DOS, and Solaris is just 4.4BSD with a few tweaks.
Sorry, I just can't agree. OS/X is _not_ UNIX, but it's hardly just a souped up NextSTEP - the *STEP api is just one of three core APIs for the OS. The others are the OS/9 compat Carbon API, and the POSIX API. I'd be more inclined to call OS/X what happens when you put OS/9, NextSTEP, and an old version of FreeBSD in a blender;-)
The credit to Microsoft for assistance is interesting. They're clearly taking more than one approach to attempt make good on their "stamp out spam" promise.
This particular tack is one that MS is uniquely positioned for, given their rather strong contacts in government (hmm) and impressive financial and personell strengths.
Hell, I wish 'em luck. It'd be nice if they'd stop with the "gain control of eMail" angle, but this approach is useful. Even if it's not overly effective or efficient, it'll be one more thing that makes spamming less worthwhile, and that can only be good.
<p>Last I checked it comes free with the hardware and 99$ upgrades. cheaper than winblows</p> </i></blockquote>
<p>Perhaps the pricing is different in the USA. However, while the initial install may be free, upgrades to MacOS X are annual and cost (in Australia) AU$229. That adds up to nearly AU$700 over three years. No upgrade pricing is offered. By comparison, WinXP costs AU$289 (OEM, shipped on the machine just like MacOS) and doesn't have to be re-purchased annually. MS also tend to offer upgrade pricing from old OS versions.</p>
<p>What Apple does have going for them in this regard is that AppleCare, which is AU$419, also includes OS/X updates. It's still more expensive than Windows, enough to matter for a business on a budget, but not as bad.</p>
<p>Now, I'm no fan of MS - but I find Apple's software pricing a bit over the top given the premium one already pays for their hardware.</p>
<blockquote><i> My PowerBook hasn't crashed in two years. Can you say that about a PC? </i></blockquote>
<p>A Windows PC - hell, no. Any MacOS X machine I've ever used - hell, no, though they seem much more stable than XP. A Linux box - hell no, but it seems more stable than both OS/X and Windows.</p>
<blockquote><i> Have you heard of coctail or onyx? guess what you can do all of the above. </i></blockquote>
<p>The tip is appreciated. If Apple will support users using those utilities, that even makes a difference. I'm very surprised they don't offer these basic options built into the operating system, though - there's nothing wrong with an "advanced settings" menu.</p>
<blockquote><i> The x86 arch is the worst architecture in the world. I pray that Apple would never use it. If they went with Spark or POWER4 or even Opteron sans the x86 instructions that would be ok. </i></blockquote>
<p>I agree - I think x86 is pretty shocking in design, it's hack layered upon hack. The hacks add up to make it work OK, but it's still pretty gross. Unfortunately, it's also cheap, mass-market, and fast.</p>
<p>Your comments about architecture choice leave me a bit puzzled. First, I can't see any point in Apple going to Spark - it'd be an instruction set change for little purpose, as the spark is also relatively expensive and niche market. As for POWER4, Apple is already using it in a cut down and slightly extended form - the G5.</p>
<p>Finally, regarding the Opteron, I believe you are seriously mistaken. The Opteron, you see, <i>is</i> x86 through and through. The 64bit support and extra instructions do not compose a whole new independent instruction set, but simply some extensions that also clean up and fix a few things like the pathetic lack of registers in x86. An Opteron without the x86 instrution set does nothing.</p>
<blockquote><i> As for CD-ROMs, a little hole and a paperclip work in a pinch. </i></blockquote>
<p>My point is that it's absurd that that is required, when there's a perfectly good eject button that the user is being prevented from accessing.</p>
<blockquote></i> Reset - theres a key sequence called force quit. </i></blockquote>
<p>Yes, and sometimes it works. However, if the OS crashes (OS9 does this quite a bit, and OSX occasionally too), that does you no good at all. What's wrong with having the reset switch there? That's what I don't get - why make life harder for the user?</p>
<blockquote><i> Mouse - I have a Logitech 3 button wheel trackball that works great. </i></blockquote>
<p>Yep. All the macs at work have MS optical wheel mice and they work great too. I just don't get why I have to buy them in the first place. I've given
Agreed regarding stability. I have recently found that our win9x machines in fact quite stable. They do play up sometimes, but in all cases a reset cures it - no random preferences corruption, no weird registry stuff-ups, nothing that just doesn't go away on its own.
As someone who utterly loathes Win9x, having suffered my time on it, this irritates me greatly;-)
Our (one) XP box has caused much more trouble, usually with broken software updates. In part it's my fault - I've been stupid enough to try to have the system used day-to-day on a restricted user account, and it just doesn't f**ing work.
Then there's MacOS 9, which I'd rather not talk about. *shudder*.
I rather like OS/X in some ways, but a few things would bother me:
(a) It's even more expensive than Windows
(b) Apple can't even make it totally stable on a hardware platform they entirely control, so how would it far on x86?
(c) The OS doesn't like to give the user control of a lot of aspects of the user interface. Don't want various animations? Well... too bad.
(d) Chances are that if Apple ever ported OS/X they'd do it on a custom x86 sub-architecture with a custom BIOS etc. We all know how overdue x86 is for a good revamp, and the idea of x86 with OpenFirmware sounds mighty good to me. The idea of an x86 box at Apple prices with Apple hardware design principles does not.
I also dislike Apple's general attitude to the user. Think removed reset switches (Why? Because "MacOS X doesn't crash?" Riiiight), no eject buttons on CD-ROM and Zip drives, and that infernal mouse. I could fix the hardware side with OS/X on x86, but not where that attitude makes its self known in the software.
Security would also be an issue. I remain unconvinced that Apple has security under control in OS/X - with MacOS 7 to MacOS 9 they largely ignored it, mostly because the systems had basically no externally visible services. There are signs they still don't quite "get it," like extremely long delays on updates, even later updates for older versions, etc.
I support a mixed network of Win9x, WinXP, MacOS 9, and MacOS/X, so I'm not just talking hot air. I rather strongly dislike all operating systems, just some more than others;-) . MacOS X does pretty well overall, so just because I have some major issues with it doesn't mean I don't think it's an OK OS.
I guess overall I'd use it in preference to WinXP, _if_ application vendors got on the ball and started building for x86 as well. There would really need to be a fat binary scheme similar to what was done with m68k and PPC for that transition. However, WinXP is not my only choice, and for my personal needs OS/X loses against a well-customised Linux box.
For my users... maybe. Apple would have to provide saner pricing, or an upgrade scheme that was a bit more reasonable than what they currently have. Upgrade pricing on the OS would be a good start. If the pricing issue was addressed, I'd provide OS/X for my users in a second - even if I had to buy overpriced apple hardware to do it. x86? Who cares, I'd put a bunch of eMacs in right now if they fixed their OS pricing (though the eMac's shoddy build quality does bother me). No more MSIE security holes - yay. Of course, there would still be network security holes, random crashes, and weird system and preference corruption (something Macs have been doing since MacOS 7), but that seems to be life with both major OS options.
Slashdot Mirror
Personally, I wouldn't - I'd just buy the cheap crap one if there was a choice, unless the price on the better one was sufficiently reasonable.
That was intended to be part of my point - some things are worth paying top dollar for, some aren't. Perhaps I failed to make said point - it was 4:00am when I posted that after all.
Only for work. Please look in my comment history for discussion on that matter, including why I'm not interested in conducting it on Slashdot - especially with an AC.
Agreed. That's exactly why I said that while I'll pay for quality in some things, computer equipment is not one of those things. Not for personal use, anyway - work is different because the cost of failure is more than "Bugger, off to the PC shop again tomorrow."
I'm not sure it's possible to make a modern console for US$300 that'll work like your old Atari.
... nfi.
Modern consoles are a _lot_ more complex, unfortunately. Some, like the xbox, also borrow from PC tech - including the "cheap and nasty" aspect.
Of course, I suppose the Atari was pretty cutting edge and complex for the time too, so perhaps I'm just talking hot air.
While it would be interesting to see an "Xbox Premium" - think good quality SCSI disk set to operate at lower RPMs, quality DVD-ROM, faster CPU running under spec to keep it cool and quiet, etc - I doubt many would buy it. Almost certainly not enough to cover the costs of making the two different variants.
Worse, MS couldn't form a viable market around a quality console - volume and market share are king, and if you want those you _must_ win on price.
As for PATA HDDs... you lucky bastard. I use SATA disks in a RAID array at work for our main storage array (fairly static "live" archival data, backed up). I had a 60% failure rate on the first batch. When I recently upgraded the array to 250GB disks, I again had a 60% failure rate. I lost three disks in the first two weeks. I've also had less than stellar results at home. I've since discovered that the first round was a bad batch of disks - the three with sequential serial no#s died, the other two were fine, and that batch was confirmed to be bad. The more recent round
All the disks were well cooled and carefully installed. All were confirmed to have serious media failures using S.M.A.R.T and mfgr disk tools.
With a failure rate like this, we'd be better off spending 4-5x as much to get something like WD Raptors. Of course, I'll probably use consumer SATA again next time because I won't be given the budget for anything else. Grr.
Bike in this case == pushbike , not motorcycle. Sorry.
... manufacturing problems mostly if the frame is good. Lower quality frames skimp on materials and use cheaper construction techniques that can make them weaker than one would expect and more prone to failure.
Possibly language differences AU <> US<> UK ?
As for how it can fail
Having snapped my first bike's frame (the rear wheel attachment point fell off - a clean break in the metal on both sides - while accelerating across an intersection), I care about this stuff. Also, given the option I like to buy quality.
In addition to all this, a better frame tends to be lighter - which is always nice.
Unfortunately, that model doesn't seem to actually work. The reason everything used to be well made and often sold with lifetime warranties - but rarely is now - is that people started buying the cheaper option no matter what.
People buy on price. So long as that persists, there's only a small market for quality products.
Would you really pay probably five times as much for a system you can expect to run vastly more reliably and run for, say, ten years? If so, would you, by the end of the second year why you didn't buy the cheap one?
I do agree that it's gone too far (6 month HDD warranties and 90 day warranties on products that cost several hundered - WTF?!) but usually where it makes sense you have the option of a decent warranty. You also, again where it's practical to make one, usually have the option of a better made product that will last longer.
So tell me, do you have enterprise class SCSI or SATA drives in your PC? Or did you buy the cheap 120GB PATA ones?
I'll put my money where my mouth is when I care. I just bought a very nice bike with a lifetime warranty on the frame - and don't ever expect to have to call in the warranty. It doesn't cost that much more to make a bike frame _much_ better, and it's worth the premium.
On the other hand, if someone told me SUN were making dual Opteron workstations with a lifetime warranty (and build quality such that SUN didn't expect it to be used much) so why don't I get one - it's "only" $10,000 - I'd laugh in their face. To me, my PC dying is acceptable compared to the price tradeoff I get in exchange for that risk - I have a functional brain, I keep backups and know how to recover quickly so it's not that big a deal.
I often wish it was possible to buy better quality products, and am increasingly irritated by the tendency of quality to equate with overpriced wank - to pay for quality, often you're forced to pay for stupid wank factor too. Sometimes you just can't seem to get quality (I've been through three kettles recently, and have just given up hope of finding one that doesn't suck). Still, indestructable and expensive isn't for every situation, much as crap with a warranty or cheap with no warranty are both also unsuitable for many situations.
I really don't see a stable driver API happening, though. One can write a stable API in C as well as C++ - but the kernel folks explicitly choose not to. I can understand why, though I won't get into the endless arguments about the merits of the decision.
/already there/. No 'net access required, no stuffing about with boot-time driver disks, it just works. I suppose one could provide tools to download and build all the drivers, but I just don't see the attraction vs the current scheme while the drivers are so closely bound to the kernel core.
Also, regarding the difficulty of getting drivers - it's not, to me, a matter of having them easy to install. It's that they're
Seriously. The fact that one can simply build and install a Linux kernel and _already_ have all the drivers required is IMO very important.
I really don't like the idea of having to go online and download drivers, and if the drivers were availible as another bundled download, really what would be the point?
Anyway, I don't really see how separating drivers from the kernel core could make sense given the Linux development policy on stable driver APIs (that is, there aren't any).
That's something I rarely see here - a post that _deserves_ the "insightful" label.
;-)
Of course, the fact that I entirely agree with you (btw, I'm Australian, and _so_ ashamed to be right now) may have something to do with this
Why not simply use the `nv' driver and not use a 3D screensaver? Reliability is excellent. If you need 3D, the card this guy was talking about wouldn't do you any good anyway.
On the other hand, I use the `nvidia' driver on my home PC and find it 100% dead reliable as well, so I guess results vary.
--
Craig Ringer
The guy who proposed this was not talking about a high-performance card - or a card with any 3D acceleration.
;-)
Remember the S3 Trio32 and Trio64? It sounds like it'll be something like that - though presumably with more video RAM, a faster DAC or DVI, and better 2D acceleration. Maybe more like a FreeG450
A cleanroom rewrite probably wouldn't do it. It's entirely likely that they have other people's /patented/ processes in use without license. Worse, they probably don't even know what they might be infringing on - the joy of software patents being what they are.
;-)
If they don't release their code, it arguably makes it harder for patent litigation companies (or competitors) to go trawling.
Additionally, there's a good chance they've licensed other people's code for use as well, and they'd have to re-implement that before releasing such a driver.
That's a lot of work, and a lot of risk, for what would be little gain for them. Much as it'd be nice to assume that the "open source community" would help maintain and improve the drivers, I'm not convinced that would be the case for these specialised extreme-performance bits of software that are closely tied to rapidly changing hardware.
There's also the possibility that their competititors could learn more about their drivers, and their hardware, from the code than from the disassembly of the drivers that they no doubt already do. One would certainly hope so, since code that's no easier to study than code generated from disassembled object code doesn't strike me as fun to work with.
For these reasons, I just don't see it happening. The Linux and BSD user market is small, and the percentage of that market with money and strong opinions about open source or free software is smaller. The proportion with the strong opinions and the willingless to being their wallet out to back those opinions is probably a lot smaller still.
Right now, it's probably much more worth their while to mainain the drivers in house and track whatever the kernel folks break today
"it was a 64K shared line with 90% packet loss, I received 60Kbps for the video stream. "
That makes no sense. You cannot reconstruct lost data out of nowhere. Using an ECC scheme works by sacrificing peak _normal_ bandwidth in exchange for error tolerance. With an ECC scheme capable of withstanding 90% packet loss you can't fit a 60kbps video through a 64kbps link even at zero packet loss. For that matter, it wouldn't be easy to do with TCP/IP, either.
Are you sure you don't mean a 64kilobit/second video through a 64kilobyte/second (~512k) link? That verges on possible, though it's still really pushing your luck at 90% loss. Alternately, I could understand it being vaguely possible if you mean 60KB/s _after_ encoding with the ECC data, though again that'd be a bit of a stretch IMO.
They may simply mean that they consider the GPL non-commercial by its very nature, that is that it is not generally possible to deeply embed GPL code in a larger product. If your view of commercial use is limited to embedding something in a closed source product, I suppose that makes sense.
;-)
Of course, if it's based on a GPL product they'll have a rather harder time offering "commercial" (presumably conventional embed in a closed source for a fee style) licenses than they might be expecting
Aggressive quarantining is great. I actually quarantine zip files - there's a small false positive rate, but the work caused retrieving the occasional quarantined file is minimal compared to the time and effort saved by blocking them.
;-)
I'm lucky enough to have most of my users either smart enough not to open suspicious attachments, or on computers where it just doesn't matter. The LTSP users, obviously, can do whatever they want - though they're encouraged to follow sensible security rules anyway, both at home and at work. Ditto the MacOS 9 users. For the win98 and XP users, it looks like the message got through on something like the eighteenth try for most of them, and they now tend to act fairly sensibly.
Given that I have the class of user who will call me and say "my email is broken" when somebody (one person) said they sent a message five minutes ago and it hasn't arrived yet, I wonder at your organisation's choice of staff
Agreed.
I don't presently even run a proxy server. The business has perfectly good EXISTING ways of detecting abuse by staff, much the same as for excessive personal phone calls.
To date it simply has not been an issue - folks will use 'net banking, sometimes do a bit of research, or look up something they've been sent (yeah, security, I know...). If they get their jobs done well, nobody cares.
That said, I'm not 100% against filters, I just think they should be used appropriately and after thinking about what the real problem is. An administrator who uses filters is not automatically incompitent (and in fact is more likely to have been forced into using some product by management).
I'm also unconvinced that unrestricted 'net access is good for productivity. Some staff really just don't need it. That said, at my work we've decided to simply accept that, expect them to behave responsibly, and deal with it though normal means if they don't.
The show is utter drivel. What frightens me is that everything else is moving down to its level (not that most TV content had far to fall). TV "news" is worse every year, we get "reality" TV now, etc.
Given how bad even the sections of the media that at least pretend to have some form of integrity are, such as the papers, are, I decided several years back that I had better things to do than watch that drek. (Of course, this argument is somewhat undermined by being in the process of reading SlashDot, but anyway...).
Even the ABC (different to USA; ABC is the gov't funded TV channel) news is pretty terrible these days.
It seems that any form of depth, coverage of both sides of an issue, or even *gasp* admitting that there's no "right" answer is just forbidden by the TV media. Yay.
One small point: If you're "disgusted with the crappy tv we have to put up with" it may be worth recognising that you do not, in fact, have to put up with it. You're perfectly free not to watch it.
Mono, a free .NET implementation, runs on PPC, x86, Sparc, and other architectures. It will run on Linux, Solaris, OS/X, and other OSes.
.NET CLR is only availible for x86 Windows.
The _Microsoft_
There will be some, or a lot, of PPC ASM used in performance critical parts of Aqua. That'll need rewriting. Remember that Darwin doesn't support any of the OS/X GUI stuff - certainly not on x86.
They would also have to address driver differences, especially with video. Remember that Apple uses video cards with custom BIOSes.
That said, I'd be a little surprised if there wasn't a skunkworks project at Apple to build a version of OS/X on x86 that could at least stumble along.
If OS/X is just a souped-up version of NextSTEP, then Win2k3 is probably "just" a souped-up version of DOS, and Solaris is just 4.4BSD with a few tweaks.
;-)
Sorry, I just can't agree. OS/X is _not_ UNIX, but it's hardly just a souped up NextSTEP - the *STEP api is just one of three core APIs for the OS. The others are the OS/9 compat Carbon API, and the POSIX API. I'd be more inclined to call OS/X what happens when you put OS/9, NextSTEP, and an old version of FreeBSD in a blender
The credit to Microsoft for assistance is interesting. They're clearly taking more than one approach to attempt make good on their "stamp out spam" promise.
This particular tack is one that MS is uniquely positioned for, given their rather strong contacts in government (hmm) and impressive financial and personell strengths.
Hell, I wish 'em luck. It'd be nice if they'd stop with the "gain control of eMail" angle, but this approach is useful. Even if it's not overly effective or efficient, it'll be one more thing that makes spamming less worthwhile, and that can only be good.
Gee, wasn't that stupid. All that, and I (a) forget to preview and (b) forget to enable HTML.
Colour me an idiot.
Agreed regarding stability. I have recently found that our win9x machines in fact quite stable. They do play up sometimes, but in all cases a reset cures it - no random preferences corruption, no weird registry stuff-ups, nothing that just doesn't go away on its own.
;-)
As someone who utterly loathes Win9x, having suffered my time on it, this irritates me greatly
Our (one) XP box has caused much more trouble, usually with broken software updates. In part it's my fault - I've been stupid enough to try to have the system used day-to-day on a restricted user account, and it just doesn't f**ing work.
Then there's MacOS 9, which I'd rather not talk about. *shudder*.
I rather like OS/X in some ways, but a few things would bother me:
;-) . MacOS X does pretty well overall, so just because I have some major issues with it doesn't mean I don't think it's an OK OS.
... maybe. Apple would have to provide saner pricing, or an upgrade scheme that was a bit more reasonable than what they currently have. Upgrade pricing on the OS would be a good start. If the pricing issue was addressed, I'd provide OS/X for my users in a second - even if I had to buy overpriced apple hardware to do it. x86? Who cares, I'd put a bunch of eMacs in right now if they fixed their OS pricing (though the eMac's shoddy build quality does bother me). No more MSIE security holes - yay. Of course, there would still be network security holes, random crashes, and weird system and preference corruption (something Macs have been doing since MacOS 7), but that seems to be life with both major OS options.
(a) It's even more expensive than Windows
(b) Apple can't even make it totally stable on a hardware platform they entirely control, so how would it far on x86?
(c) The OS doesn't like to give the user control of a lot of aspects of the user interface. Don't want various animations? Well... too bad.
(d) Chances are that if Apple ever ported OS/X they'd do it on a custom x86 sub-architecture with a custom BIOS etc. We all know how overdue x86 is for a good revamp, and the idea of x86 with OpenFirmware sounds mighty good to me. The idea of an x86 box at Apple prices with Apple hardware design principles does not.
I also dislike Apple's general attitude to the user. Think removed reset switches (Why? Because "MacOS X doesn't crash?" Riiiight), no eject buttons on CD-ROM and Zip drives, and that infernal mouse. I could fix the hardware side with OS/X on x86, but not where that attitude makes its self known in the software.
Security would also be an issue. I remain unconvinced that Apple has security under control in OS/X - with MacOS 7 to MacOS 9 they largely ignored it, mostly because the systems had basically no externally visible services. There are signs they still don't quite "get it," like extremely long delays on updates, even later updates for older versions, etc.
I support a mixed network of Win9x, WinXP, MacOS 9, and MacOS/X, so I'm not just talking hot air. I rather strongly dislike all operating systems, just some more than others
I guess overall I'd use it in preference to WinXP, _if_ application vendors got on the ball and started building for x86 as well. There would really need to be a fat binary scheme similar to what was done with m68k and PPC for that transition. However, WinXP is not my only choice, and for my personal needs OS/X loses against a well-customised Linux box.
For my users