Wellington, New Zealand, did this around four years ago with CityLink, and it has worked really well for the city, fostering quite a lot of internet connectivity in the days before the telcos really had their act together.
My company have bootstrapped ourselves into the whole area of developing and serving dynamic content, and using this was the first step that we needed to take three years ago. Without it we wouldn't be where we are today.
I'm not so sure if it is quite as needed now, although on a different scale I guess it has a lot of value. I also do some work for a small-town ISP who provide connectivity at special rates to schools and so forth, subsidised by the commercial providers.
Unfortunately these documents don't seem to present a balanced picture. The bias stands out so much that they are effectively useless to me to advocate Linux in organisations who I think would have value in considering it.
Yet again, it is a case of advocacy blinding sense.:-(
I have switched to using PDF format to send documents to people. This works really well for letters, quotes and so forth, but is obviously less useful for collaborative work where I have still stuck with RTF.
PDF works well, giving excellent print characteristics, everyone ends up with some sort of PDF reader installed - I can produce it entirely with libre tools and still depend on it being readable by my clients. It even includes compression and encryption, and being a read-only format is sometimes an advantage.
Abiword too is coming along nicely. Now that it supports styles, the only thing missing for me to use it for all my (fairly limited) WP needs is to support tables.
A hilarious parody of MozillaQuest can be found at http://mozillaquestquest.com/ although really, does it need a parody?
MozillaQuest is usually so creative in his reporting that he might as well not bother. His claims bear no resemblance to any reality I participate in, and there is little point in rebutting him. If we all ignore him then perhaps he will go away? We can hope so.
My wife has been happily using Mozilla for around a year now, and she doesn't seem to notice any missing content. She's reasonably computer literate, but not a geek. That's under Windows 98, since she needs to run Pagemaker as well.
I use Mozilla under Linux on my laptop, and it certainly works a lot better for me than IE5.x, let alone IE4.x.
Konqueror also looks pretty good, and people tell me that Opera is good too.
While what you suggest might have been true even nine months ago, the browser situation on Linux has matured immensely this year.
Personally, I do take the idealistic view, and I don't have flash or java installed in my own browser since I have yet to find a site I need it for that I actually care about.
My wife does have both of those installed on her system, and Paddington Bear works just fine for my son. I'm not sure what else beyond Flash and Java is really needed for 'correct' viewing of all sites, but the numbers that require more than those two really are a vanishingly small percentage - certainly I have not heard of any.
Interestingly, most of the people in our company (we build dynamic content websites) actually develop to Mozilla (by choice), and then go back and fix things when IE is incompatible in some way. This is at least partly because of the lack of functionality that IE provides to designers - have you ever tried to get the source of a frame, for example?
I have encoded around 12G of music onto my laptop so that I can listen to it anytime. This isn't music for anyone else, and I don't download music from the internet, or swap with other people.
Consequently I don't give a toss whether anyone else can read my music or not. Ogg works well for me, and the player support seems just fine and dandy - I've been using XMMS, FreeAmp and ogg123 which all do what I expect.
Ripping the CDs seems no different for Ogg than it is for mp3, except maybe easier, since Debian has never included mp3 ripping software because of the patent problems.
Next year I will be building a small computer to go into the stereo system to play music. I'll get more serious about ripping my music collection in it's entirety at that point, but I can't see that it is going to get any easier for me to rip to MP3 and I can't see that it will ever get any harder to rip to Ogg.
For the next few years I expect to see all of the.mp3,.ogg and.wma (and maybe others) in the market place. The comparison of VHS vs BETAMAX simply doesn't apply because I can play all three of these on any hardware - they are software codecs after all. I personally can't currently play.wma, but that's just a choice I have made, not a majority view:-)
I'd also like to express my thanks to Monty for his effort and vision in making this post possible!
There are a host of irreplaceable men in the country's graveyards.
Or something like that.
Anyway, you're an idiot if your project depends on having a particular individual doing a particular job at a particular time. Anytime I see this happening to any project I'm running, I move the person off to another job and put in an understudy for the role.
If you used a suitably structured request / response string it would get cached by ISPs all over the internet and wouldn't bog the servers down so much.
Also, you could allow people to mirror the results, and have a hierarchy of sites to provide the md5sum. The application itself should be able to cache the md5sum result too, which would be a huge win unless AOL are changing it randomly.
Probably you've already gone through these sorts of options.
Hell, you could even have a DNS server return the md5sum in response to a particular address lookup, and that would be cached nicely by the whole damn internet:-)
Here in New Zealand many of us pay per megabyte for internet connections. Downloading an ISO image, presumably around 650MB would cost my company around $100NZD.
Yep, and $100NZD is around USD$45, so downloading software like that is no bargain for us.
I will always buy from a local supplier who does the work of downloading and mastering the images and will produce a CD usually in my (snail) mail box with a 24 hour turnaround.
We get a pretty good rate for our internet connection too - I know other organisations would be paying around 30% - 50% above what we pay.
Religion isn't a big issue in NZ. From here it appears that the US is much more religious than we are, I think.
I've seen the census statistics for the last hundred years on this question and it's actually really interesting reading, changing from 99% claiming some form of 'christian' through to the current day, when fewer than 50% do, and of those who do, around 80% are pretty vague about what that means.
My girlfriend always used to put "Aztec Sun-Worshiper" down on her census forms. I'm sure Jedi will slip through with a laugh or two from the collaters.
I'm afraid I'm a boring old atheist, so they won't bat an eyelid at that:-)
I use Oracle in my day job. It definitely has some good features, and a good few mis-features as well.
While you are probably correct that there aren't any open source alternatives at the high end, at the mid-range PostgreSQL is definitely an alternative, with (better than) row level locking, transactions, foreign keys and nested selects all there in current production versions.
Performance is critical in _any_ database, and you can't add features at the expense of performance and expect to keep your customers happy. You can add slowish implementations of new features, but you better not downgrade existing performance.
PostgreSQL 7.1 is now in beta and should be out by the end of February (maybe earlier). With support for column sizes up to 2GB(!), as well as many other refinements, I certainly don't think that they are sacrificing feature efforts to performance enhancement.
That's not to say I find it slow either. My most recent web development (replacing a prior ASP/MSSQL/IIS installation:-) at http://newsroom.co.nz/
manages to do around 30-odd queries for the front page and still pops it out in just under 0.1 of a second on a Duron 700 based system.
I looked at MySQL (and I help out a client who uses it on occasion), but the potential performance improvement isn't worth it to me to lose all those neat features.
It should only change when you install a new key. Installation of new sshd shouldn't be sufficient.
Most of our systems have the same ssh keys they had when we were running Debian 1.3, for example. The upgrade to OpenSSH which was in Debian 2.1 didn't change the keys. Motherboard and disk upgrades have not changed the keys.
The Debian packages manage to avoid this problem. I think it's probably just easier for them to say this than to write the installer stuff to work around it.
No doubt you can get RPMs somewhere which bypass this as well.
Certainly, the nightly builds, which I have been using for at least six months, and which have been my only browser for the last three, write user-related settings into your home directory, with the software itself just needing to be "on path" somewhere.
Similarly, the Windows version which I use on NT installs in a perfectly fine multi-user way, with my settings associated with my profile, and the executables in a shared location.
From other news reports I have seen, and from the proposed budgets (3 million pounds setup / 9 million pounds per year) this appears to be a proposal not to log the call (email/...) data but to log the fact that the call (e-mail/...) has been made.
That would seem to be a much more achievable goal.
The proposal appears to be one to make telcos and ISPs hold the logged data for seven years too, rather than have MI5/MI6/NCIS hold the data.
That also seems more likely: just tell everyone they have to hold records of this for seven years. Then if you find you need them you get a warrant and requisition them.
Of course I live about as far away from Britain as you can get, so my analysis could be completely wrong!
There is an MP3 site (I forget the URL exactly) and we had a journalist sniffing around our company at the time who was writing about MP3s.
We mentioned the Fraunhofer license fees and he apparently told the owner of this site who didn't know and just about wet his pants.
Personally I've switched to OGG Vorbis now that it is supported in XMMS and in WinAmp. Good to see that streaming in Ogg is now working, and IceCast supports that too.
The Debian Way might seem ponderous and conservative if that's what you're looking for, but there is also a bleeding edge side of Debian which many people follow.
I apt-get my system most days to upgrade it to the latest packages from woody. Sure this sometimes lives up to it's name of 'unstable' but surpisingly infrequently. Most of the time this is just a good way to keep up with the latest bleeding edge packages.
I haven't got the time to program for the open source movement, but at least I can help with the testing:-)
I have been using Debian for the last few years and I rarely see them mentioned in this sort of article.
This makes me wonder if the statistics that articles such as this quote are really worth bothering with! It seems to me that unless you take statistics from debian.org and all it's mirrors that you won't be getting a true picture of the marketplace.
Debian is a significant force, and one which increasingly deserves recognition as we see other distributions based around it's excellent package management system.
I've been using Storm recently and find it an excellent way to get Debian installed, with all the power of Debian's management system available to me afterwards.
The Walnut Creek downloads of ISO images is particularly bogus in this regard - why download an ISO image of Debian? Grabbing a couple of floppies and doing a net install is much easier...
Yeah, I always find it hard to type a lowercase b after an uppercase M for some reason.
Switching technology costs have fallen enormously over the last couple of years and I actually seriously doubt that this sort of shit will happen too often over hubbed networks.
Certainly LANs seem to have transitioned pretty much from 10Mb hubs through to 100Mb switches. I see the occasional site that stepped into either 10Mb switched (very rare) or 100Mb hubbed (less rare), but by and large that seems to have been leapfrogged.
Come Gb ethernet being commonly installed on office LANs I don't expect to see any hubbed topologies really.
When I was running a bulletin board about ten years ago I remember that used to do this in a very basic form too.
Type in too many exclamation marks!!! or keep the CAPS LOCK key on too much and the editor status line would come up and say helpful things like "Flame mode on" and stuff.
With sensible ISPs at the other end of that fibre offering their services along with a transparent proxy cache I seriously doubt that one users bandwidth will screw things for everyone else.
Remember, that 100MB is only to the ISP, from there who knows what the topology is?
We have a fibre link into our office and, although I can occasionally get incredible transfer rates across it, by and large the bandwidth limitations hit downstream of our ISP. When we set this up we actually visited the ISP and they showed us their bandwidth ceiling and usage patterns before we signed - incredibly open policy which we appreciated. Their upstream bandwidth certainly is not 100MB, yet.
When many [home] users are on 100MB then the ISP will have increased their bandwidth to meet the statistical demand already.
PostgreSQL will certainly handle databases this big. On Linux there is a 2GB file size limitation (being removed in 2.4, I believe) but PostgreSQL will split it's files at around 1G anyway to get around this.
There are other filesystem limitations that may have to be worked around with various Unixes, but managing to get a partition of 100GB or greater should be achievable. In Linux you would probably use logical volumes, but you could simply do it with links if you wanted.
Slashdot Mirror
Wellington, New Zealand, did this around four years ago with CityLink, and it has worked really well for the city, fostering quite a lot of internet connectivity in the days before the telcos really had their act together.
My company have bootstrapped ourselves into the whole area of developing and serving dynamic content, and using this was the first step that we needed to take three years ago. Without it we wouldn't be where we are today.
I'm not so sure if it is quite as needed now, although on a different scale I guess it has a lot of value. I also do some work for a small-town ISP who provide connectivity at special rates to schools and so forth, subsidised by the commercial providers.
Unfortunately these documents don't seem to present a balanced picture. The bias stands out so much that they are effectively useless to me to advocate Linux in organisations who I think would have value in considering it.
:-(
Yet again, it is a case of advocacy blinding sense.
Speaking of Bullsh*t, 'defecit' is actually supposed to be spelt 'defecate'.
:-)
Or maybe you meant that other word...
I have switched to using PDF format to send documents to people. This works really well for letters, quotes and so forth, but is obviously less useful for collaborative work where I have still stuck with RTF.
PDF works well, giving excellent print characteristics, everyone ends up with some sort of PDF reader installed - I can produce it entirely with libre tools and still depend on it being readable by my clients. It even includes compression and encryption, and being a read-only format is sometimes an advantage.
Abiword too is coming along nicely. Now that it supports styles, the only thing missing for me to use it for all my (fairly limited) WP needs is to support tables.
A hilarious parody of MozillaQuest can be found at http://mozillaquestquest.com/ although really, does it need a parody?
MozillaQuest is usually so creative in his reporting that he might as well not bother. His claims bear no resemblance to any reality I participate in, and there is little point in rebutting him. If we all ignore him then perhaps he will go away? We can hope so.
My wife has been happily using Mozilla for around a year now, and she doesn't seem to notice any missing content. She's reasonably computer literate, but not a geek. That's under Windows 98, since she needs to run Pagemaker as well.
I use Mozilla under Linux on my laptop, and it certainly works a lot better for me than IE5.x, let alone IE4.x.
Konqueror also looks pretty good, and people tell me that Opera is good too.
While what you suggest might have been true even nine months ago, the browser situation on Linux has matured immensely this year.
Personally, I do take the idealistic view, and I don't have flash or java installed in my own browser since I have yet to find a site I need it for that I actually care about.
My wife does have both of those installed on her system, and Paddington Bear works just fine for my son. I'm not sure what else beyond Flash and Java is really needed for 'correct' viewing of all sites, but the numbers that require more than those two really are a vanishingly small percentage - certainly I have not heard of any.
Interestingly, most of the people in our company (we build dynamic content websites) actually develop to Mozilla (by choice), and then go back and fix things when IE is incompatible in some way. This is at least partly because of the lack of functionality that IE provides to designers - have you ever tried to get the source of a frame, for example?
I have encoded around 12G of music onto my laptop so that I can listen to it anytime. This isn't music for anyone else, and I don't download music from the internet, or swap with other people.
Consequently I don't give a toss whether anyone else can read my music or not. Ogg works well for me, and the player support seems just fine and dandy - I've been using XMMS, FreeAmp and ogg123 which all do what I expect.
Ripping the CDs seems no different for Ogg than it is for mp3, except maybe easier, since Debian has never included mp3 ripping software because of the patent problems.
Next year I will be building a small computer to go into the stereo system to play music. I'll get more serious about ripping my music collection in it's entirety at that point, but I can't see that it is going to get any easier for me to rip to MP3 and I can't see that it will ever get any harder to rip to Ogg.
For the next few years I expect to see all of the .mp3, .ogg and .wma (and maybe others) in the market place. The comparison of VHS vs BETAMAX simply doesn't apply because I can play all three of these on any hardware - they are software codecs after all. I personally can't currently play .wma, but that's just a choice I have made, not a majority view :-)
I'd also like to express my thanks to Monty for his effort and vision in making this post possible!
Thanks Monty!
There are a host of irreplaceable men in the country's graveyards.
Or something like that.
Anyway, you're an idiot if your project depends on having a particular individual doing a particular job at a particular time. Anytime I see this happening to any project I'm running, I move the person off to another job and put in an understudy for the role.
Also, you could allow people to mirror the results, and have a hierarchy of sites to provide the md5sum. The application itself should be able to cache the md5sum result too, which would be a huge win unless AOL are changing it randomly.
Probably you've already gone through these sorts of options.
Hell, you could even have a DNS server return the md5sum in response to a particular address lookup, and that would be cached nicely by the whole damn internet :-)
Here in New Zealand many of us pay per megabyte for internet connections. Downloading an ISO image, presumably around 650MB would cost my company around $100NZD.
Yep, and $100NZD is around USD$45, so downloading software like that is no bargain for us.
I will always buy from a local supplier who does the work of downloading and mastering the images and will produce a CD usually in my (snail) mail box with a 24 hour turnaround.
We get a pretty good rate for our internet connection too - I know other organisations would be paying around 30% - 50% above what we pay.
I've seen the census statistics for the last hundred years on this question and it's actually really interesting reading, changing from 99% claiming some form of 'christian' through to the current day, when fewer than 50% do, and of those who do, around 80% are pretty vague about what that means.
My girlfriend always used to put "Aztec Sun-Worshiper" down on her census forms. I'm sure Jedi will slip through with a laugh or two from the collaters.
I'm afraid I'm a boring old atheist, so they won't bat an eyelid at that :-)
I use Oracle in my day job. It definitely has some good features, and a good few mis-features as well.
While you are probably correct that there aren't any open source alternatives at the high end, at the mid-range PostgreSQL is definitely an alternative, with (better than) row level locking, transactions, foreign keys and nested selects all there in current production versions.
Performance is critical in _any_ database, and you can't add features at the expense of performance and expect to keep your customers happy. You can add slowish implementations of new features, but you better not downgrade existing performance.
PostgreSQL 7.1 is now in beta and should be out by the end of February (maybe earlier). With support for column sizes up to 2GB(!), as well as many other refinements, I certainly don't think that they are sacrificing feature efforts to performance enhancement.
That's not to say I find it slow either. My most recent web development (replacing a prior ASP/MSSQL/IIS installation :-) at http://newsroom.co.nz/
manages to do around 30-odd queries for the front page and still pops it out in just under 0.1 of a second on a Duron 700 based system.
I looked at MySQL (and I help out a client who uses it on occasion), but the potential performance improvement isn't worth it to me to lose all those neat features.
Most of our systems have the same ssh keys they had when we were running Debian 1.3, for example. The upgrade to OpenSSH which was in Debian 2.1 didn't change the keys. Motherboard and disk upgrades have not changed the keys.
Only YOU should be changing the keys.
This research is interesting in that it extends that further, finding a way to control the effect other than through diet management.
No doubt you can get RPMs somewhere which bypass this as well.
Certainly, the nightly builds, which I have been using for at least six months, and which have been my only browser for the last three, write user-related settings into your home directory, with the software itself just needing to be "on path" somewhere.
Similarly, the Windows version which I use on NT installs in a perfectly fine multi-user way, with my settings associated with my profile, and the executables in a shared location.
Andrew.
That would seem to be a much more achievable goal.
The proposal appears to be one to make telcos and ISPs hold the logged data for seven years too, rather than have MI5/MI6/NCIS hold the data.
That also seems more likely: just tell everyone they have to hold records of this for seven years. Then if you find you need them you get a warrant and requisition them.
Of course I live about as far away from Britain as you can get, so my analysis could be completely wrong!
:-)
We mentioned the Fraunhofer license fees and he apparently told the owner of this site who didn't know and just about wet his pants.
Personally I've switched to OGG Vorbis now that it is supported in XMMS and in WinAmp. Good to see that streaming in Ogg is now working, and IceCast supports that too.
The Debian Way might seem ponderous and conservative if that's what you're looking for, but there is also a bleeding edge side of Debian which many people follow.
I apt-get my system most days to upgrade it to the latest packages from woody. Sure this sometimes lives up to it's name of 'unstable' but surpisingly infrequently. Most of the time this is just a good way to keep up with the latest bleeding edge packages.
I haven't got the time to program for the open source movement, but at least I can help with the testing :-)
This makes me wonder if the statistics that articles such as this quote are really worth bothering with! It seems to me that unless you take statistics from debian.org and all it's mirrors that you won't be getting a true picture of the marketplace.
Debian is a significant force, and one which increasingly deserves recognition as we see other distributions based around it's excellent package management system.
I've been using Storm recently and find it an excellent way to get Debian installed, with all the power of Debian's management system available to me afterwards.
The Walnut Creek downloads of ISO images is particularly bogus in this regard - why download an ISO image of Debian? Grabbing a couple of floppies and doing a net install is much easier...
Yeah, I always find it hard to type a lowercase b after an uppercase M for some reason.
Switching technology costs have fallen enormously over the last couple of years and I actually seriously doubt that this sort of shit will happen too often over hubbed networks.
Certainly LANs seem to have transitioned pretty much from 10Mb hubs through to 100Mb switches. I see the occasional site that stepped into either 10Mb switched (very rare) or 100Mb hubbed (less rare), but by and large that seems to have been leapfrogged.
Come Gb ethernet being commonly installed on office LANs I don't expect to see any hubbed topologies really.
When I was running a bulletin board about ten years ago I remember that used to do this in a very basic form too.
Type in too many exclamation marks!!! or keep the CAPS LOCK key on too much and the editor status line would come up and say helpful things like "Flame mode on" and stuff.
I guess they won't get a patent on it then, huh.
Remember, that 100MB is only to the ISP, from there who knows what the topology is?
We have a fibre link into our office and, although I can occasionally get incredible transfer rates across it, by and large the bandwidth limitations hit downstream of our ISP. When we set this up we actually visited the ISP and they showed us their bandwidth ceiling and usage patterns before we signed - incredibly open policy which we appreciated. Their upstream bandwidth certainly is not 100MB, yet.
When many [home] users are on 100MB then the ISP will have increased their bandwidth to meet the statistical demand already.
Admittedly they still need to use a special functio of bochs to achieve it, but things are moving quite quickly on this even so.
After Linux comes Windows, they suggest.
PostgreSQL will certainly handle databases this big. On Linux there is a 2GB file size limitation (being removed in 2.4, I believe) but PostgreSQL will split it's files at around 1G anyway to get around this.
There are other filesystem limitations that may have to be worked around with various Unixes, but managing to get a partition of 100GB or greater should be achievable. In Linux you would probably use logical volumes, but you could simply do it with links if you wanted.