This is always an option, but VMware involves less space, power consumption, etc. This is particularly significant for laptops, where lugging around two of them is not really a good option.
It all depends on how many systems still need to run Windows - if you can migrate most desktops and servers to native Linux apps, the cost of any extra Windows licenses for VMware is quite small. In practice, you should be able to re-use the existing Windows license and installation on disk (i.e. run Windows within VMware from the original disk partition), so there is no extra license cost.
Cairo was much-hyped but never delivered. It was to be the successor to NT 3.x, and MS talked about it before Chicago (Win95) was launched, as early as l993. Cairo was meant to have lots of neat features such as dragging components from one window into another (a bit like drag and drop of text, but of a live object such as a news ticker), and a clever object oriented distributed file system.
Great demos, but like all the best vaporware, it never actually went from cloud form to solid... Instead, NT 4.0 was delivered, which had only a few new features other than the Win95 shell.
There are several ways of running Windows software on Linux - WINE is the low cost option which can be used for specific applications that work OK with it, while VMware plus Win9x/NT/2000 is a more robust solution that even has server variants so you can put these Windows apps on a VMware server (a la Citrix but with more flexibility and isolation). VMware uses a lot more hardware resource than WINE, but PCs are fairly cheap and even a PII/350 runs VMware OK.
99% of Cisco routers don't support read/write SNMP, so this is not as big a hole as you suggest - you can't just upload a new config typically. Of course, admins should change the SNMP passwords, and set ACLs on SNMP so that only a few hosts can do SNMP queries in any case, and ideally use SNMPv3 authentication/encryption.
Cisco have offered SecurID hardware token support (actually by supporting TACACS+ and RADIUS, which support SecurID) for a long time, but not everyone uses it, particularly in service providers. Cisco, Juniper and Riverstone all support SSH (Cisco supports it only in some IOS versions).
Like Telnet and most HTTP, SNMPv1 and v2 have passwords (community strings) in the clear, but that's why most people don't allow read/write functions from SNMP, only read-only. SNMPv3 fixes this, but it's still not that widely used.
Backdoors are (IMO) less frequent in routers, since most of these are out on the Internet, where any such backdoors would inevitably be discovered quite rapidly. I've seen vendors claim that they have no such backdoors, which tends to support this. ATM switches may be another matter since telcos often manage them via an out-of-band network, which probvides some security by disabling management from other network links.
Anyone who leaves the passwords set to defaults deserves what they get, but it's true to say that quite a lot of networks don't change the passwords frequently (if at all). Those that use TACACS+ or RADIUS authentication servers are in much better shape, since they can change passwords from a single point, and particularly if they use SecurID, which prevents a re-usable password from being used. The best solution is to use SSH, with the caveat that this has been known to have its own security holes - so you must be prepared to update your router OS images quickly if necessary.
Multiple layers of defence are a good idea - e.g. choose strong passwords, proper password encryption, and enable SSH, and then put on ACLs so that SSH is only permitted from a limited set of addresses.
Re:From inside an asbestos bunker...
on
GNU Emacs 21
·
· Score: 2
As the other post said, Emacs evolved independently of Unix during its early years - it comes from the Lisp hacker world, where Lisp functions are the means of integrating functions, by contrast to Unix's use of pipes.
For an Emacs timeline, see http://www.jwz.org/doc/emacs-timeline.html
vim is the only version of vi to use, IMO, even if only for the ability to cursor-up from the ':' prompt to edit a complicated 's///' command.
The \L lowercasing feature was actually in Unix vi as of System V.something, but it was undocumented - fortunately I was working for a Unix porting house so we had the Unix source to discover this, but a key benefit of open source is that undocumented commands don't stay that way for long!
Re:Why Linux isn't on the desktop or never will be
on
Why Linux is About to Lose
·
· Score: 4, Insightful
The Red Hat techie who erased her files was irresponsible and stupid - you don't win people over by switching them without consultation, and particularly not by erasing people's data. However, it's unlikely that this would ever happen outside a Linux-only culture, so it's hardly applicable to the rest of the world. One idiot does not make a trend...
Linux evangelism needs a lot of work on subtle and effective techniques (as opposed to flaming), but this is not really a good example.
The biggest stimulus to Linux on the desktop is Microsoft's recent squeezing of its installed based for more revenue through changes to its licensing model - there are several local government and police organisations in the UK that are going to save millions of pounds through switching to Linux.
How many minutes of battery life do you think you'd get with all that stuff...? If you want this much flexibility, get a Visor, which can plug in most of these things, or an iPaq with a CF sleeve. The point of the Treo is to have a compact device that works well as both a phone and a mini web/email/IM device. It would be nice to have colour, but GPRS is quite a power hungry technology, so it's probably best at the moment to stay monochrome and have a longer battery life using GPRS.
Once they manage to do the 'push email' feature a la BlackBerry, of course, everyone has the interesting prospect of spam or other unexpected email actually draining the batteries...
GPRS is a lot more widespread in Europe - it's been out between 3 and 12 months depending on country and operator, with reasonable coverage although still a bit beta-quality. My main issue with dropping Graffiti is that it's still very useful for some things where you already know Graffiti - however, if they include a stylus with the keyboard version, that would probably be enough for 'mouse' type GUI interactions rather than writing (although you then have the pain of switching from keyboard to stylus).
Yes, I read the article and I also read the FAQ in detail, as well as all the other pages on the Treo...
Note that my post said 'No Graffiti *with the keyboard version*' - the point is that the *keyboard version* of this product does not have *any* built-in Graffiti support.
This is an interesting device, but I see a few problems:
- no GPRS at launch time - this is amazing given the early 2002 launch, surely this device is meant for GPRS? There are already several Pocket PC devices with GPRS that can be bought today (in Europe anyway), but I'm still waiting just for an *announcement* of a Palm OS device with built in GPRS support. The device is meant to be software-upgradeable, but it's unclear when that upgrade will be available.
- can't use Graffiti with keyboard version - there is no built in Graffiti *at all* with this version, not even a pop-up Graffiti window. You have to download a third party tool to do this. What are Handspring thinking?
And Handspring are still using Palm OS 3.5 for no very good reason, and still not using flash for the OS (unlike virtually every mobile phone). The openness, software and integration looks nice, but until they fix these issues I'll just have to carry on waiting.
Try finding a local RSI support group - they will know the best local doctors. Some good sites are www.tifaq.com and www.rsi-uk.org.uk (UK based but has some good info). If you don't have a local support group, try email lists, but people are often more forthcoming in person or on the phone.
The most important part is usually finding a physical therapist (physiotherapist) who understands RSI - ask them about adverse neural tension (ANT) and adverse mechanical tension (AMT), and how many people with RSI they've treated. Also, investigate complementary therapies such as Alexander technique, tai chi and chi kung.
If you can get ADSL for £25 in the UK, you must have a very special deal... Everyone else is paying £40 per month, $60 approx, and that's after recent price cuts. This is one reason why the UK has far fewer broadband users than the US, Germany and many other countries (in South Korea, the *majority* of Internet users are on broadband).
Cable broadband does cost about £25 per month, and there's a recently announced a lower-speed £15/$22 per month cable connection service, 128 Kbps but still always on and flat rate - a lot better than ISDN.
Roughly 80 per cent of the Swedish population has a mobile phone (not sure if that includes children, but in the UK many teenagers have mobile phones and I'm sure Sweden is even more mobile mad). Source: http://www.line56.com/articles/default.asp?NewsID= 2924
In almost every developed country other than the US, mobile phones are used by the majority of the population. Outside the US, the calling party pays extra for calls to a mobile (rather than the called person), which means that people are more likely to give out their mobile numbers, generating more traffic and making mobiles more popular. That's one reason why the US lags greatly in the use of mobiles, along with the lack of unified standards (CDMA, GSM and TDMA, and still lots of analogue).
Even when you have all-optical switches, they will initially be circuit switches, with nailed-up optical paths betweeen edge devices, putting more stress on the edge routers because they have more circuits to switch between. People are researching packet switching and burst switching but that's even further off.
Have a look at fixed wireless technology - either LMDS/MMDS type technology (licensed, via a provider) or 802.11b wireless LANs (no provider needed, roll your own...). There are some people who have wired very remote communities using point-to-point wireless (to reach the place) then 802.11b to connect individual homes.
The latest LMDS technology gives great coverage and lower costs, but 802.11b (Wi-Fi) is a better bet for DIY. If you get subs from enough people it could easily become viable. Have a look at www.opentelco.net for examples of this.
Yes, a button would be similar to autoload (as I mentioned), but at least a certain proportion of people would realise 'clicking a button means something will happen, maybe something bad' - most people don't expect viewing a document to cause something other than the doc appearing on screen.
Autoload macros are the real issue - I have never understood why Microsoft didn't simply disable these completely. I can't see a valid use for autoload macros that couldn't be substituted by a button within the document that says 'click here to start'. Since 99% of Word, Excel and PPT docs would never have such buttons, it would be much more obvious to the user that something odd was going on. This wouldn't remove the problem but it would make it much harder for such viruses to propagate.
Thanks for posting that - presumably these were minimum size ICMP packets? They are fairly poor, as you say - TCP could probably survive this as long as the sliding window ramped up a bit, but it would not be very fast. The network may need to tune the radio link encoding or use more slots, perhaps - e.g. by default, voice data is spread across several frames so that loss of any one frame does not lose all voice data, I believe.
Yes, this is known as SIM locking in the UK, and does happen. Some operators sell you SIM locked phones and charge a fee for unlocking, others do not. In practice, many people dump the phone and service after a year and get a new one - this is often cheaper than getting an upgrade phone. However, for people who do stay with one operator, it is generally possible to transfer the SIM - it's just a matter of checking you are not SIM locked, or paying the fee if you are.
QoS is being implemented in IP networks, and is out there already, using plain old IPv4. IPv6 has precisely ONE feature related to QoS, and it requires use of RSVP (the Flow Label) - since RSVP has never really been deployed, this is irrelevant tomorrow.
Having said that, GPRS QoS is pretty much a waste of time at present, because there are so few slots available. Perhaps if you can guarantee high priority customers access to the radio link, that is saleable, but it's not exactly QoS as most people think of it. In the longer term GPRS QoS may take off, as operators roll out denser cells and phones providing more bandwidth, and it becomes worth assigning QoS based on application, user, etc.
The problem with WAP is incompatibility/bugginess, low usability and lack of useful content - speed is barely faster using GPRS, which simply makes it more convenient by being always-on. Every time I try a new WAP site, it doesn't work on my phone (a Nokia 7110, supposedly a standard one), due to WAP gateway, WAP browser or site problems. I would actually use WAP a bit if it worked, despite the usability - some of the content is becoming useful, e.g. a site that tells you driving directions from any point in the UK.
Slashdot Mirror
This is always an option, but VMware involves less space, power consumption, etc. This is particularly significant for laptops, where lugging around two of them is not really a good option.
It all depends on how many systems still need to run Windows - if you can migrate most desktops and servers to native Linux apps, the cost of any extra Windows licenses for VMware is quite small. In practice, you should be able to re-use the existing Windows license and installation on disk (i.e. run Windows within VMware from the original disk partition), so there is no extra license cost.
Cairo was much-hyped but never delivered. It was to be the successor to NT 3.x, and MS talked about it before Chicago (Win95) was launched, as early as l993. Cairo was meant to have lots of neat features such as dragging components from one window into another (a bit like drag and drop of text, but of a live object such as a news ticker), and a clever object oriented distributed file system.
Great demos, but like all the best vaporware, it never actually went from cloud form to solid... Instead, NT 4.0 was delivered, which had only a few new features other than the Win95 shell.
There are several ways of running Windows software on Linux - WINE is the low cost option which can be used for specific applications that work OK with it, while VMware plus Win9x/NT/2000 is a more robust solution that even has server variants so you can put these Windows apps on a VMware server (a la Citrix but with more flexibility and isolation). VMware uses a lot more hardware resource than WINE, but PCs are fairly cheap and even a PII/350 runs VMware OK.
99% of Cisco routers don't support read/write SNMP, so this is not as big a hole as you suggest - you can't just upload a new config typically. Of course, admins should change the SNMP passwords, and set ACLs on SNMP so that only a few hosts can do SNMP queries in any case, and ideally use SNMPv3 authentication/encryption.
Cisco have offered SecurID hardware token support (actually by supporting TACACS+ and RADIUS, which support SecurID) for a long time, but not everyone uses it, particularly in service providers. Cisco, Juniper and Riverstone all support SSH (Cisco supports it only in some IOS versions).
Like Telnet and most HTTP, SNMPv1 and v2 have passwords (community strings) in the clear, but that's why most people don't allow read/write functions from SNMP, only read-only. SNMPv3 fixes this, but it's still not that widely used.
Backdoors are (IMO) less frequent in routers, since most of these are out on the Internet, where any such backdoors would inevitably be discovered quite rapidly. I've seen vendors claim that they have no such backdoors, which tends to support this. ATM switches may be another matter since telcos often manage them via an out-of-band network, which probvides some security by disabling management from other network links.
Anyone who leaves the passwords set to defaults deserves what they get, but it's true to say that quite a lot of networks don't change the passwords frequently (if at all). Those that use TACACS+ or RADIUS authentication servers are in much better shape, since they can change passwords from a single point, and particularly if they use SecurID, which prevents a re-usable password from being used. The best solution is to use SSH, with the caveat that this has been known to have its own security holes - so you must be prepared to update your router OS images quickly if necessary.
Multiple layers of defence are a good idea - e.g. choose strong passwords, proper password encryption, and enable SSH, and then put on ACLs so that SSH is only permitted from a limited set of addresses.
As the other post said, Emacs evolved independently of Unix during its early years - it comes from the Lisp hacker world, where Lisp functions are the means of integrating functions, by contrast to Unix's use of pipes.
For an Emacs timeline, see http://www.jwz.org/doc/emacs-timeline.html
vim is the only version of vi to use, IMO, even if only for the ability to cursor-up from the ':' prompt to edit a complicated 's///' command.
The \L lowercasing feature was actually in Unix vi as of System V.something, but it was undocumented - fortunately I was working for a Unix porting house so we had the Unix source to discover this, but a key benefit of open source is that undocumented commands don't stay that way for long!
The Red Hat techie who erased her files was irresponsible and stupid - you don't win people over by switching them without consultation, and particularly not by erasing people's data. However, it's unlikely that this would ever happen outside a Linux-only culture, so it's hardly applicable to the rest of the world. One idiot does not make a trend...
Linux evangelism needs a lot of work on subtle and effective techniques (as opposed to flaming), but this is not really a good example.
The biggest stimulus to Linux on the desktop is Microsoft's recent squeezing of its installed based for more revenue through changes to its licensing model - there are several local government and police organisations in the UK that are going to save millions of pounds through switching to Linux.
How many minutes of battery life do you think you'd get with all that stuff...? If you want this much flexibility, get a Visor, which can plug in most of these things, or an iPaq with a CF sleeve. The point of the Treo is to have a compact device that works well as both a phone and a mini web/email/IM device. It would be nice to have colour, but GPRS is quite a power hungry technology, so it's probably best at the moment to stay monochrome and have a longer battery life using GPRS.
Once they manage to do the 'push email' feature a la BlackBerry, of course, everyone has the interesting prospect of spam or other unexpected email actually draining the batteries...
GPRS is a lot more widespread in Europe - it's been out between 3 and 12 months depending on country and operator, with reasonable coverage although still a bit beta-quality. My main issue with dropping Graffiti is that it's still very useful for some things where you already know Graffiti - however, if they include a stylus with the keyboard version, that would probably be enough for 'mouse' type GUI interactions rather than writing (although you then have the pain of switching from keyboard to stylus).
Yes, I read the article and I also read the FAQ in detail, as well as all the other pages on the Treo...
Note that my post said 'No Graffiti *with the keyboard version*' - the point is that the *keyboard version* of this product does not have *any* built-in Graffiti support.
They are supplying this through various mobile operators - at least in Europe, you typically sign up for a year's service to get the discount.
This is an interesting device, but I see a few problems:
- no GPRS at launch time - this is amazing given the early 2002 launch, surely this device is meant for GPRS? There are already several Pocket PC devices with GPRS that can be bought today (in Europe anyway), but I'm still waiting just for an *announcement* of a Palm OS device with built in GPRS support. The device is meant to be software-upgradeable, but it's unclear when that upgrade will be available.
- can't use Graffiti with keyboard version - there is no built in Graffiti *at all* with this version, not even a pop-up Graffiti window. You have to download a third party tool to do this. What are Handspring thinking?
And Handspring are still using Palm OS 3.5 for no very good reason, and still not using flash for the OS (unlike virtually every mobile phone). The openness, software and integration looks nice, but until they fix these issues I'll just have to carry on waiting.
Try finding a local RSI support group - they will know the best local doctors. Some good sites are www.tifaq.com and www.rsi-uk.org.uk (UK based but has some good info). If you don't have a local support group, try email lists, but people are often more forthcoming in person or on the phone.
The most important part is usually finding a physical therapist (physiotherapist) who understands RSI - ask them about adverse neural tension (ANT) and adverse mechanical tension (AMT), and how many people with RSI they've treated. Also, investigate complementary therapies such as Alexander technique, tai chi and chi kung.
If you can get ADSL for £25 in the UK, you must have a very special deal... Everyone else is paying £40 per month, $60 approx, and that's after recent price cuts. This is one reason why the UK has far fewer broadband users than the US, Germany and many other countries (in South Korea, the *majority* of Internet users are on broadband).
Cable broadband does cost about £25 per month, and there's a recently announced a lower-speed £15/$22 per month cable connection service, 128 Kbps but still always on and flat rate - a lot better than ISDN.
Roughly 80 per cent of the Swedish population has a mobile phone (not sure if that includes children, but in the UK many teenagers have mobile phones and I'm sure Sweden is even more mobile mad). Source: http://www.line56.com/articles/default.asp?NewsID= 2924
In almost every developed country other than the US, mobile phones are used by the majority of the population. Outside the US, the calling party pays extra for calls to a mobile (rather than the called person), which means that people are more likely to give out their mobile numbers, generating more traffic and making mobiles more popular. That's one reason why the US lags greatly in the use of mobiles, along with the lack of unified standards (CDMA, GSM and TDMA, and still lots of analogue).
Somebody mod this up!
Even when you have all-optical switches, they will initially be circuit switches, with nailed-up optical paths betweeen edge devices, putting more stress on the edge routers because they have more circuits to switch between. People are researching packet switching and burst switching but that's even further off.
Have a look at fixed wireless technology - either LMDS/MMDS type technology (licensed, via a provider) or 802.11b wireless LANs (no provider needed, roll your own...). There are some people who have wired very remote communities using point-to-point wireless (to reach the place) then 802.11b to connect individual homes.
The latest LMDS technology gives great coverage and lower costs, but 802.11b (Wi-Fi) is a better bet for DIY. If you get subs from enough people it could easily become viable. Have a look at www.opentelco.net for examples of this.
Yes, a button would be similar to autoload (as I mentioned), but at least a certain proportion of people would realise 'clicking a button means something will happen, maybe something bad' - most people don't expect viewing a document to cause something other than the doc appearing on screen.
Autoload macros are the real issue - I have never understood why Microsoft didn't simply disable these completely. I can't see a valid use for autoload macros that couldn't be substituted by a button within the document that says 'click here to start'. Since 99% of Word, Excel and PPT docs would never have such buttons, it would be much more obvious to the user that something odd was going on. This wouldn't remove the problem but it would make it much harder for such viruses to propagate.
Thanks for posting that - presumably these were minimum size ICMP packets? They are fairly poor, as you say - TCP could probably survive this as long as the sliding window ramped up a bit, but it would not be very fast. The network may need to tune the radio link encoding or use more slots, perhaps - e.g. by default, voice data is spread across several frames so that loss of any one frame does not lose all voice data, I believe.
Yes, this is known as SIM locking in the UK, and does happen. Some operators sell you SIM locked phones and charge a fee for unlocking, others do not. In practice, many people dump the phone and service after a year and get a new one - this is often cheaper than getting an upgrade phone. However, for people who do stay with one operator, it is generally possible to transfer the SIM - it's just a matter of checking you are not SIM locked, or paying the fee if you are.
QoS is being implemented in IP networks, and is out there already, using plain old IPv4. IPv6 has precisely ONE feature related to QoS, and it requires use of RSVP (the Flow Label) - since RSVP has never really been deployed, this is irrelevant tomorrow.
Having said that, GPRS QoS is pretty much a waste of time at present, because there are so few slots available. Perhaps if you can guarantee high priority customers access to the radio link, that is saleable, but it's not exactly QoS as most people think of it. In the longer term GPRS QoS may take off, as operators roll out denser cells and phones providing more bandwidth, and it becomes worth assigning QoS based on application, user, etc.
The problem with WAP is incompatibility/bugginess, low usability and lack of useful content - speed is barely faster using GPRS, which simply makes it more convenient by being always-on. Every time I try a new WAP site, it doesn't work on my phone (a Nokia 7110, supposedly a standard one), due to WAP gateway, WAP browser or site problems. I would actually use WAP a bit if it worked, despite the usability - some of the content is becoming useful, e.g. a site that tells you driving directions from any point in the UK.