You need to understand that Slot machines, unlike roulette wheels, are not purely random. They contain logic that pre-decides where the wheels will spin to and discard outlier values that would take the machine outside programmed min and max payout rates (as well as do other things designed to hook and encourage play). These payout rates are frequently set by law. Thus if a machine is reaching the low point there actually is a higher chance of a win as the machine discards more losing combinations.
This one does work on iPhone. Slow as a wet week, of course, but very impressive. Crashed first time, but then ran fine. Perhaps if I had an iPad 2 with it's super processor I might start vi...
See Oztivo where they've hacked the series 1 frimware to get it's updates over the Internet from a community run guide service. I'm in NZ and using the New Zealand variant on an English Series 1 TiVo (the kind we're discussing here) to good effect. TiVo have resisted people doing this in countries where they're selling the guide service - perhaps now is the time for the community to have a go?
I'm a little suspicious of the emission claims though. How much of that is from plugin? I can't imagine turbine->electric->battery->motors is an efficient drive train.
Turbo-electric (ie: turbine->electric->motors) are quite efficent, and commonly used in large equipment, like boats and trains (see http://en.wikipedia.org/wiki/Turbo-electric)
The difficulties here will be 1. how efficent the battery is, and how much the battery is used verses running in direct turbo-electric mode. 2. How well the turbine has been scaled down. Turbines get harder to make efficent the smaller they are - efficency is quite dependent on things like the ratio of the gap at the edge of the blades to the blade area. Small turbines need a lot more precision manufacturing to make properly efficent. A good single-cycle gas turbine such as this one: http://www.geoilandgas.com/businesses/ge_oilandgas/en/literature/en/downloads/LM6000.pdf can get 42% efficency, but small models often languish at 25% or so. [NB: combined cycle can get you as high as 60% but I will be *very* surprised if they've crammed that into a car... though they did say 2 turbines....]
I would like to know more about this single table relational database concept you just implied exists, where should I start reading?;)
I would have thought almost any database theory book would do, after all the "relation" in a relational database is between the primary key of a tuple and the elements (think mathematical relationship - eg: x=y*2 where x is integers 1 to 10 and write that as a table with columns for x and y and you'll get the picture)
However outside of theory most things that are called relational databases do have lots of tables and SQL engines and so on....
I thought that the most interesting thing about the article was the info about the launch method (for those who missed the link it's here.) I had not realised that this (logical) idea had actually been capitalised on! I had wondered for a while if a jet powered launch module with wings might be more cost effective (cost + reuseability) than using rockets to throw stuff into space, and here is someone who's simply using old jet aircraft to do the job! They've even been doing it for 10 years, just to make me feel silly for having not noticed it before.
Now all we need is for them to be able to take larger cargos, and we can begin to work our way up to putting humans in space at reasonable prices. That might require specially designed launchers, but I suspect that you could still use standard jet engines... The advantage with my not having noticed it for 10 years is that their patents must be expiring soon:-)
I'd go for the keeping secrets theory myself. They did say that this was immersed in sulfuric acid. This is a strong acid, and would supply large numbers of H+ ions....
I'm not sure I understand what they're doing either, but it looks more chemical and "battery" (cell to the accuracy-obcessed) like than physicsey and capacitor like.
The press release claims that the energy density is the same as lead-acid batteries. We can presume that is for weight/unit volume. No clear indication of the weight of this unit though... I would presume much lighter (I associate lead with heavy) and though this would be it's biggest advantage in many cases (esp. cars), don't forget that Lithium/Ion and even Nickel Metal Hydride have higher energy density (erg/vol) than lead acid...
I think that the speed of recharge might be most useful, as e-cars would not require an overnight charge. 5 Minutes at a servo (with a thick enough cable to carry the amps) wouldn't be so bad.
You're right. Note also that direct sequence 'chirps' and Frequency hopping *shouldn't* interfere, as the FH signal should only be able to kill one bit of the chirp at a time... leaving the other 11(?) or so to get through intact.
This article gives very little detail on whats going on here. They don't even say if the author even tried adjusting the frequency settings of the interfering components. 802.11 has 11 (in the US) -non Frequency hop mode frequency settings (most of which overlap) and selecting another one might have helped a bit!
Some background for the curious: 802.11 sends 'chirps' - the same bit is sent simultaniously on a range of frequencies, with some bits reversed. this is done to prevent interference, or blocking on one band from interfering with the transmition.
The channels look somthing like this (in ASCII anyway) (if this looks wrong, paste it into an xterm, or notepad, or somthing with a fixed-width font)
Note that the vertical axis doesn't represent anything, it's just used to stop everything going on top of everything else. The dots are there because slashdot slashes spaces, but leaves dots! Nor is this diagram accurate, or to scale or anything, it's just ment to give you the gist.....
Here each ---x--- is a range of frequencies over which the bits of the chirp are spread.
There only one set of 3 channels which don't overlap, so if you need more than 3 802.11 networks in the same place, you're our of luck.
If you run your network in frequency hopping mode, you only transmit one bit on one frequency at a time (chirps send about 12 bits), but change frequencies often, across the whole range (no channels) This means that interference on one range will only kill some of your data. You obviously than need to retransmit failed sends (by the time a retransmit happens you will have switched to a different frequency.
The quality of the hardware you use can also make a big difference. The better equipment uses two aerials, spaced apart, to prevent reflected signals and some other kinds of interference from silencing the signal. The idea is that if a signal and it's reflection interfere to create a minimum (no signal) at one point, there will be signal just a short distance sway.
Most devices just ship set to a channel, and it's nearly always the same one - surprise surprise - 1! I don't know about the phones, but they would probably be similar.
I guess no detail, or background research is about what we expect from ZDNet.....:-(
One big advantage of digital for those of us who don't live in the same country as Hollywood, would be the potential for simultanious world-wide releases of movies. The production cost of rolls of film is such that the studios wait until US cinemas have finished with the reels before releasing them overseas. This is the cause of the hated DVD regional encoding (to stop the home release version being mail ordered from the US by countries still viewing the celluloid version!) I would very much have liked that situation to end.
OTOH I want the best cinema experience possible, and if physical film offers that, then that's what I want.
Interestingly, the sale of APS (The new still film in the smaller cartrage, with the small data store for picture information) cameras vastly outstrips the sale of digital ones. Analogue has won here, despite the enormous digital camera hype (remember that?) so there's no reason why it shouldn't do so for motion film too.
Companies often say that things that are *obviously* initals don't stand for anything. This is done because you can't trademark initals (you trademark the full name.) So they say they have no meaning (even when they do) as a workaround. This may well be a case in point.
The story goes that they originally chose characters that sounded exactly like coca-cola, but later found that they meant "Bite the wax tadpole", so they changed it for characters sounding less like the western name, but meaning "Happiness in the mouth"
AOL would be nuts to use a technique like this, for one simple reason - SECURITY. Anyone who'se cycles are used to used to do work for the service provider has the ability to tamper with the process. They could return junk data, or altered data deliberately to mess with the system. As they weren't AOLs cycles to use anyway, they would have little recourse, and (if it were put to the uses the patent suggests) public content would be affected. A whole new way to crack AOL!!
You could just wait for NDS for Linux to come out - it *is* in development, and as they have done the work for a Solaris version, shouldn't take too long now...
You could use one of the other machine imaging packages which has App distribution on top. Imagecast (which I can't recommend, BTW) and RapidDeploy spring to mind. These mostly require windows server boxes to do their work though.
You could use Microsoft's own sysdiff utility for distributing the apps contents to the workstation, and just use a profile provided common desktop to create the icons.
Ultimately, I don't think anything comes near ZEN for all round ease of use and flexabilty to mass administer 95 and NTWS. It is Novell's crown jewels at the moment, and really is the thing that keeps sites using it from going to other systems. This is not all bad, as it at least stops managers from taking the organisation to NT! In Linux's quest to become a complete OS we need a desktop mass-administrator - for Linux desktops as well as Windows. How about an RPM installer on Windows. repackage Windows apps in RPMs, then have a launcher to check for the packages installation before starting the app.....
I have used ZEN extensively (for many clients on varied sites), and not seen any flakyness in the core functionality, that is, the Application Launcher.
Some of the other components have been problematic at times (don't get me started on the problems with the remote control software) But never Appication launching. Policy distribution, and workstation management bits have been good too.
It appears that the use of IE5 may be prohibited under this law! --quote--- 6.1. A registered user must, on each occasion, input allocated PIN or password together with date of birth to gain access to any website subject to the system.
6.1.1. The system must not allow for automated input of login information, for example, by saving on a cookie file stored on the user's PC the allocated PIN or password and date of birth. ---endquote----
6.1.1 appears to disallow IE5 because it allows password storage, and the system *must not allow for* password storage!
It is interesting that they don't mention WINE as competition. Normally ZDNet not mentioning the Open Source product wouldn't surprise me, but this *is* a Linux article.
Wine (by my guesstimation) is looking at a similar time period to be stable enough to port sellable applications with. Corel must think so too, or Corel Office on Linux would be too far off to be worth doing this way (IMHO, of course))
To head off wasted posts quickly, please remember that WINE is *BOTH* a Win32 API implementation, AND a Windows emulator (The latter being a binary loader and interface to the former, of course)
I agree with most of what you're saying. PC managment isn't perfect by a long shot, and could be a lot better.
Remote managment was added to PC systems because of the threat that NCs posed. I suspect that any improved NC will have it's features copied into the PCs, and the NC itself will disappear.
Replacing 1/3 of an organisation's PCs would (in most cases) make things harder, as there would now be TWO different systems to maintain. Maybe if you were large and single site, or you had to support the back-end platform the NCs relyed on anyway....
I *Could* see a Linux distribution that was designed to be remote managable working though. Current systems (telnet, X-sessions) work fine for remote managment of single boxes, but remote mass-managment has a way to go yet. It would counter many of the proprietry hardware and network speed issues in my original post. There is still the issue of application availability, user re-training
And before anyone suggests Citrix for Windows Apps, it's just as hard to tie down as individual PCs, and much less reliable (one user's blue-screen is everyone's problem!) Good for dial-up access though.
Sounds like you really need a workstation administration system like ZENWorks (Novell) or SMS (Micros~1). disclaimer: I can't speak for SMS but I presume it provides a reasonably similar feature set to ZEN.
With ZEN (2.0), I can package up new applications, or service packs, or even O/S upgrades, and deploy them automatically, even whilst the workstation is unattended. With NT workstations it can perform the installs as administrator so that you can lock down normal users. It can automaically repair some forms of "damaged" apps (such as when win9x users delete DLLs. It also allows you to apply policies (both the MS ones, and others) to lock down the PCs, and applies them much more reliably than MS policy files (they are stored in NDS) Coupled with Ghost which lets you deploy a standard image quickly, or a small stock of swap-out PCs, you can have any sort of damage fixed very quickly.
If you run a Novell Network, go download a copy of ZENWorks (the starter pack is free, and has *most* of the features.) It's most useful if you deploy a new SOE (Standard Operating Environment - OS and Apps) where you can lock everything down, but can reduce headaches even in existing environments. Standard desktop managment doesn't even run any server side components! (not counting NDS)
I have serviced and worked on NC based networks too, and have found that, in general, they are as much, if not more, work than a well-designed, locked down PC based network.
Note: There seem to be two things called NCs -A- The "Reincarnated X-Terminal" Type Such as the NCD explora (alright, they *can* run some local apps, but they mostly don't) -B- The "Reincarnated diskless workstation" Type Such as the Sun Javastation
I couldn't be *certain* which were being talked about here, so I have tried to address both. Please bear in mind that some of the below points don't apply in both cases however. My thoughts are as follows 1. Price - Non-standardness Raises the price - The price gets kept down only by removing essential functionality. - PC prices get reduced to match NC price gains. - Type As require hefty servers which are expensive and depreciate quickly. 2. Administration - NCs usually introduce a new, and different way to administer the Applications on them. - By the release of the NCs, PC software had cought up with any new ideas in it. I suspect this will be repeated in the next round. 3. Speed - Transferring all your programs and data across the wire always winds up slower than reading them off a local disk. Much better to pre-transfer them to local storage, and load from there. - Type As are terrible for any multimedia application. And too many standard applications have multimediaish bits in them. 3. Maintainence - PCs are easy to replace. Changing brands may be difficult if the environment is really tied down, but being locked into some propritery technology is worse. (this applies more to type Bs) - Networked PCs can work in some sort of limited "stand-alone" mode if the network dies.
The concept of NCs has been useful, however. It has forced software vendors into creating products that make WinTel PCs less of a nightmare to administer. I use Novell's ZENWorks a lot, and it can be used to provide "NC-like" administrability (have I just made a new word?:-) ) with PCs
Another Idea PCs might usefully steal now is to make a machine at a lower price by only having iMac type expandability. Most corporations and govt. depts don't care about expandability for most of their PCs.
So, I guess although I think NCs will die, their useful components will be merged into PCs.
Yeah, but all of the Linux patches all fix more than one thing also.
I understood that, but I was just trying to point out that counting patches was neither a measure of the number of bugs in the original product, nor of the software's inherent quality (or lack of such)
I understand this but my point was, it has no FORMAL test methodology. The methodology is "throw it out to the beta testers", rather than first testing it.
Ok, I see what you meant. I thought you were implying that just because test products were freely available, they weren't in test.
I however, don't think that a formal pre-release test would necessarally improve things, or even make sense for many projects. Most free projects these days are completely open development model, and there is no stage iduring which it is not "thrown out to the testers". Some form of formal testing as a part of the unstable phase might be useful in some cases though (especally before the transition to a "stable" version).
Firstly, there are not 5 patches to Windows NT, there are 5 Service Packs. These are a combination of a large number of smaller fixes rolled togeather. There are a large number of "Hot fixes" also, which usually just fix one, or a number of smaller bugs, and are later rolled into the next service pack.
Secondly, Free software certianly does have a testing phsae. This is what all the "unstable" release versions are all about. With free software however you are free to disregard the warnings if you wish.
Slashdot Mirror
It's like writing a C compiler in Bourne shell. The point is less about the name than about the complexity and absurdity.
Isn't it more like writing a C compiler in C-shell - at least name-wise :-)
You need to understand that Slot machines, unlike roulette wheels, are not purely random. They contain logic that pre-decides where the wheels will spin to and discard outlier values that would take the machine outside programmed min and max payout rates (as well as do other things designed to hook and encourage play). These payout rates are frequently set by law. Thus if a machine is reaching the low point there actually is a higher chance of a win as the machine discards more losing combinations.
This one does work on iPhone. Slow as a wet week, of course, but very impressive. Crashed first time, but then ran fine. Perhaps if I had an iPad 2 with it's super processor I might start vi...
See Oztivo where they've hacked the series 1 frimware to get it's updates over the Internet from a community run guide service. I'm in NZ and using the New Zealand variant on an English Series 1 TiVo (the kind we're discussing here) to good effect. TiVo have resisted people doing this in countries where they're selling the guide service - perhaps now is the time for the community to have a go?
I'm a little suspicious of the emission claims though. How much of that is from plugin? I can't imagine turbine->electric->battery->motors is an efficient drive train.
Turbo-electric (ie: turbine->electric->motors) are quite efficent, and commonly used in large equipment, like boats and trains (see http://en.wikipedia.org/wiki/Turbo-electric)
The difficulties here will be
1. how efficent the battery is, and how much the battery is used verses running in direct turbo-electric mode.
2. How well the turbine has been scaled down. Turbines get harder to make efficent the smaller they are - efficency is quite dependent on things like the ratio of the gap at the edge of the blades to the blade area. Small turbines need a lot more precision manufacturing to make properly efficent. A good single-cycle gas turbine such as this one: http://www.geoilandgas.com/businesses/ge_oilandgas/en/literature/en/downloads/LM6000.pdf can get 42% efficency, but small models often languish at 25% or so. [NB: combined cycle can get you as high as 60% but I will be *very* surprised if they've crammed that into a car... though they did say 2 turbines....]
All the same, I still want one!
I would like to know more about this single table relational database concept you just implied exists, where should I start reading? ;)
I would have thought almost any database theory book would do, after all the "relation" in a relational database is between the primary key of a tuple and the elements (think mathematical relationship - eg: x=y*2 where x is integers 1 to 10 and write that as a table with columns for x and y and you'll get the picture)
However outside of theory most things that are called relational databases do have lots of tables and SQL engines and so on....
I thought that the most interesting thing about the article was the info about the launch method (for those who missed the link it's here .) I had not realised that this (logical) idea had actually been capitalised on! I had wondered for a while if a jet powered launch module with wings might be more cost effective (cost + reuseability) than using rockets to throw stuff into space, and here is someone who's simply using old jet aircraft to do the job! They've even been doing it for 10 years, just to make me feel silly for having not noticed it before.
Now all we need is for them to be able to take larger cargos, and we can begin to work our way up to putting humans in space at reasonable prices. That might require specially designed launchers, but I suspect that you could still use standard jet engines... The advantage with my not having noticed it for 10 years is that their patents must be expiring soon :-)
I'd go for the keeping secrets theory myself. They did say that this was immersed in sulfuric acid. This is a strong acid, and would supply large numbers of H+ ions....
I'm not sure I understand what they're doing either, but it looks more chemical and "battery" (cell to the accuracy-obcessed) like than physicsey and capacitor like.
The press release claims that the energy density is the same as lead-acid batteries. We can presume that is for weight/unit volume. No clear indication of the weight of this unit though... I would presume much lighter (I associate lead with heavy) and though this would be it's biggest advantage in many cases (esp. cars), don't forget that Lithium/Ion and even Nickel Metal Hydride have higher energy density (erg/vol) than lead acid...
I think that the speed of recharge might be most useful, as e-cars would not require an overnight charge. 5 Minutes at a servo (with a thick enough cable to carry the amps) wouldn't be so bad.
Yes, I realised after posting that I had not actually used the words "direct sequence" in the description. Doh!
I did, however mention frequency hopping, and describe it briefly.
You're right. Note also that direct sequence 'chirps' and Frequency hopping *shouldn't* interfere, as the FH signal should only be able to kill one bit of the chirp at a time... leaving the other 11(?) or so to get through intact.
This article gives very little detail on whats going on here. They don't even say if the author even tried adjusting the frequency settings of the interfering components. 802.11 has 11 (in the US) -non Frequency hop mode frequency settings (most of which overlap) and selecting another one might have helped a bit!
:-(
Some background for the curious: 802.11 sends 'chirps' - the same bit is sent simultaniously on a range of frequencies, with some bits reversed. this is done to prevent interference, or blocking on one band from interfering with the transmition.
The channels look somthing like this (in ASCII anyway) (if this looks wrong, paste it into an xterm, or notepad, or somthing with a fixed-width font)
|
|............---7---
|..........---6---
|........---5---
|......---4---........---11---
|....---3---........---10---
|..--2---.........---9---
|---1---........---8---
+---------------------------------------------
frequency ----->
Note that the vertical axis doesn't represent anything, it's just used to stop everything going on top of everything else. The dots are there because slashdot slashes spaces, but leaves dots! Nor is this diagram accurate, or to scale or anything, it's just ment to give you the gist.....
Here each ---x--- is a range of frequencies over which the bits of the chirp are spread.
There only one set of 3 channels which don't overlap, so if you need more than 3 802.11 networks in the same place, you're our of luck.
If you run your network in frequency hopping mode, you only transmit one bit on one frequency at a time (chirps send about 12 bits), but change frequencies often, across the whole range (no channels) This means that interference on one range will only kill some of your data. You obviously than need to retransmit failed sends (by the time a retransmit happens you will have switched to a different frequency.
The quality of the hardware you use can also make a big difference. The better equipment uses two aerials, spaced apart, to prevent reflected signals and some other kinds of interference from silencing the signal. The idea is that if a signal and it's reflection interfere to create a minimum (no signal) at one point, there will be signal just a short distance sway.
Most devices just ship set to a channel, and it's nearly always the same one - surprise surprise - 1! I don't know about the phones, but they would probably be similar.
I guess no detail, or background research is about what we expect from ZDNet.....
One big advantage of digital for those of us who don't live in the same country as Hollywood, would be the potential for simultanious world-wide releases of movies. The production cost of rolls of film is such that the studios wait until US cinemas have finished with the reels before releasing them overseas. This is the cause of the hated DVD regional encoding (to stop the home release version being mail ordered from the US by countries still viewing the celluloid version!) I would very much have liked that situation to end.
OTOH I want the best cinema experience possible, and if physical film offers that, then that's what I want.
Interestingly, the sale of APS (The new still film in the smaller cartrage, with the small data store for picture information) cameras vastly outstrips the sale of digital ones. Analogue has won here, despite the enormous digital camera hype (remember that?) so there's no reason why it shouldn't do so for motion film too.
Companies often say that things that are *obviously* initals don't stand for anything. This is done because you can't trademark initals (you trademark the full name.) So they say they have no meaning (even when they do) as a workaround. This may well be a case in point.
The story goes that they originally chose characters that sounded exactly like coca-cola, but later found that they meant "Bite the wax tadpole", so they changed it for characters sounding less like the western name, but meaning "Happiness in the mouth"
AOL would be nuts to use a technique like this, for one simple reason - SECURITY. Anyone who'se cycles are used to used to do work for the service provider has the ability to tamper with the process. They could return junk data, or altered data deliberately to mess with the system. As they weren't AOLs cycles to use anyway, they would have little recourse, and (if it were put to the uses the patent suggests) public content would be affected. A whole new way to crack AOL!!
You could just wait for NDS for Linux to come out - it *is* in development, and as they have done the work for a Solaris version, shouldn't take too long now...
You could use one of the other machine imaging packages which has App distribution on top. Imagecast (which I can't recommend, BTW) and RapidDeploy spring to mind. These mostly require windows server boxes to do their work though.
You could use Microsoft's own sysdiff utility for distributing the apps contents to the workstation, and just use a profile provided common desktop to create the icons.
Ultimately, I don't think anything comes near ZEN for all round ease of use and flexabilty to mass administer 95 and NTWS. It is Novell's crown jewels at the moment, and really is the thing that keeps sites using it from going to other systems. This is not all bad, as it at least stops managers from taking the organisation to NT! In Linux's quest to become a complete OS we need a desktop mass-administrator - for Linux desktops as well as Windows. How about an RPM installer on Windows. repackage Windows apps in RPMs, then have a launcher to check for the packages installation before starting the app.....
ZEN is notoriously flaky, too.
I have used ZEN extensively (for many clients on varied sites), and not seen any flakyness in the core functionality, that is, the Application Launcher.
Some of the other components have been problematic at times (don't get me started on the problems with the remote control software) But never Appication launching. Policy distribution, and workstation management bits have been good too.
It appears that the use of IE5 may be prohibited under this law!
--quote---
6.1. A registered user must, on each occasion, input allocated PIN or password together with date of birth to gain access to any website subject to the system.
6.1.1. The system must not allow for automated input of login information, for example, by saving on a cookie file stored on the user's PC the allocated PIN or password and date of birth.
---endquote----
6.1.1 appears to disallow IE5 because it allows password storage, and the system *must not allow for* password storage!
Another hole?
It is interesting that they don't mention WINE as competition. Normally ZDNet not mentioning the Open Source product wouldn't surprise me, but this *is* a Linux article.
Wine (by my guesstimation) is looking at a similar time period to be stable enough to port sellable applications with. Corel must think so too, or Corel Office on Linux would be too far off to be worth doing this way (IMHO, of course))
To head off wasted posts quickly, please remember that WINE is *BOTH* a Win32 API implementation, AND a Windows emulator (The latter being a binary loader and interface to the former, of course)
I agree with most of what you're saying. PC managment isn't perfect by a long shot, and could be a lot better.
Remote managment was added to PC systems because of the threat that NCs posed. I suspect that any improved NC will have it's features copied into the PCs, and the NC itself will disappear.
Replacing 1/3 of an organisation's PCs would (in most cases) make things harder, as there would now be TWO different systems to maintain. Maybe if you were large and single site, or you had to support the back-end platform the NCs relyed on anyway....
I *Could* see a Linux distribution that was designed to be remote managable working though. Current systems (telnet, X-sessions) work fine for remote managment of single boxes, but remote mass-managment has a way to go yet. It would counter many of the proprietry hardware and network speed issues in my original post. There is still the issue of application availability, user re-training
And before anyone suggests Citrix for Windows Apps, it's just as hard to tie down as individual PCs, and much less reliable (one user's blue-screen is everyone's problem!) Good for dial-up access though.
Sounds like you really need a workstation administration system like ZENWorks (Novell) or SMS (Micros~1).
disclaimer: I can't speak for SMS but I presume it provides a reasonably similar feature set to ZEN.
With ZEN (2.0), I can package up new applications, or service packs, or even O/S upgrades, and deploy them automatically, even whilst the workstation is unattended. With NT workstations it can perform the installs as administrator so that you can lock down normal users. It can automaically repair some forms of "damaged" apps (such as when win9x users delete DLLs. It also allows you to apply policies (both the MS ones, and others) to lock down the PCs, and applies them much more reliably than MS policy files (they are stored in NDS) Coupled with Ghost which lets you deploy a standard image quickly, or a small stock of swap-out PCs, you can have any sort of damage fixed very quickly.
If you run a Novell Network, go download a copy of ZENWorks (the starter pack is free, and has *most* of the features.) It's most useful if you deploy a new SOE (Standard Operating Environment - OS and Apps) where you can lock everything down, but can reduce headaches even in existing environments. Standard desktop managment doesn't even run any server side components! (not counting NDS)
I have serviced and worked on NC based networks too, and have found that, in general, they are as much, if not more, work than a well-designed, locked down PC based network.
Note: There seem to be two things called NCs
:-) ) with PCs
-A- The "Reincarnated X-Terminal" Type
Such as the NCD explora (alright, they *can* run some local apps, but they mostly don't)
-B- The "Reincarnated diskless workstation" Type
Such as the Sun Javastation
I couldn't be *certain* which were being talked about here, so I have tried to address both. Please bear in mind that some of the below points don't apply in both cases however.
My thoughts are as follows
1. Price
- Non-standardness Raises the price
- The price gets kept down only by removing essential functionality.
- PC prices get reduced to match NC price gains.
- Type As require hefty servers which are expensive and depreciate quickly.
2. Administration
- NCs usually introduce a new, and different way to administer the Applications on them.
- By the release of the NCs, PC software had cought up with any new ideas in it. I suspect this will be repeated in the next round.
3. Speed
- Transferring all your programs and data across the wire always winds up slower than reading them off a local disk. Much better to pre-transfer them to local storage, and load from there.
- Type As are terrible for any multimedia application. And too many standard applications have multimediaish bits in them.
3. Maintainence
- PCs are easy to replace. Changing brands may be difficult if the environment is really tied down, but being locked into some propritery technology is worse. (this applies more to type Bs)
- Networked PCs can work in some sort of limited "stand-alone" mode if the network dies.
The concept of NCs has been useful, however. It has forced software vendors into creating products that make WinTel PCs less of a nightmare to administer. I use Novell's ZENWorks a lot, and it can be used to provide "NC-like" administrability (have I just made a new word?
Another Idea PCs might usefully steal now is to make a machine at a lower price by only having iMac type expandability. Most corporations and govt. depts don't care about expandability for most of their PCs.
So, I guess although I think NCs will die, their useful components will be merged into PCs.
Yeah, but all of the Linux patches all fix more than one thing also.
I understood that, but I was just trying to point out that counting patches was neither a measure of the number of bugs in the original product, nor of the software's inherent quality (or lack of such)
I understand this but my point was, it has no FORMAL test methodology. The methodology is "throw it out to the beta testers", rather than first testing it.
Ok, I see what you meant. I thought you were implying that just because test products were freely available, they weren't in test.
I however, don't think that a formal pre-release test would necessarally improve things, or even make sense for many projects. Most free projects these days are completely open development model, and there is no stage iduring which it is not "thrown out to the testers". Some form of formal testing as a part of the unstable phase might be useful in some cases though (especally before the transition to a "stable" version).
Firstly, there are not 5 patches to Windows NT, there are 5 Service Packs. These are a combination of a large number of smaller fixes rolled togeather. There are a large number of "Hot fixes" also, which usually just fix one, or a number of smaller bugs, and are later rolled into the next service pack.
Secondly, Free software certianly does have a testing phsae. This is what all the "unstable" release versions are all about. With free software however you are free to disregard the warnings if you wish.