Looks like an early christmas this year. I've fixed the reasons why it wouldn't work in a subdirectory now, so it will be able to operate within a subdirectory of a virtual host from the next release.
I've put a "don't require a whole virtual server" thing on the road map now, and will see if I can have that done before christmas.
The "caldav.php" in the URL can be pretty easily got rid of with a URL rewriting line, so I might add a note on how to do that to the installation docs, or somewhere in the Wiki as well.
I'd love to see a package that acts as both a CalDAV server, but also gives you the ability to view and edit the calendars via a nice looking web-interface as well
Cosmo does, of course. I don't intend to do that with RSCDS, although the current code (in Git) will let you see basic information it does not let you do any add / edit of the calendar entries, and presents events in a list, rather than as dates on a calendar.
If you do have any suggestions for improvements in RSCDS, please tell me about them though:-)
Yeah, I know. It uses register globals in a safe way, but I need to bite the bullet one day and just disable it and go through my libraries and find where it's used. Probably I'm at that point now and should Just Do It:-)
The problem is basically that the libraries I use have been written over the last five years or so and there are some places where that is expected. For all of the code written in the last couple of years (including the actual RSCDS code) it isn't required.
Happy to accept patches, of course:-)
Andrew McMillan.
Clearly this is contradictory. If you "can't wait" until it hist Debian stable then you are looking for a release which is less stable than Debian stable.
The reality is, of course much simpler. Odds are that given it's optional "runs on top of Xorg" nature it will be available in Debian testing within 3 months and will consequently be released next time the 16000 or so Debian packages are declared stable enough for a release.
Setup a site to advertise a product, then restrict people from using it....
Similar to the main linked site, GeShi is the product of a very smart young coder (who I happen to employ) and who does not have the funds to pay for bandwidth. He doesn't advertise on the site, and GeShi is GPL and something he did for fun in his spare time.
I'll see if we can get it onto some better bandwidth / equipment for him so he can cope with Slashdot in the future, but it's not going to happen during this event.
Well, I think we're trying to talk about things that FOSS solutions can't do here...
Out of office replies are easy, and generally much better than the brain-dead activities of Exchange. You can say "only send one out of office reply to any given individual every seven days" for example. Or "don't send out of office replies to automatic e-mails". Exchange can't do that (or if it can, someone please tell me - I would love a clue-stick to beat some admins with...:-)
I doubt if anyone would seriously suggest that the sorts of server-side filtering you can do with Procmail are lacking in functionality. And if they are, there's always Perl... The problems here are maybe in the UI, but that's fairly easily fixable and something I don't know about probably already has.
Read receipts are likewise easy. Every MTA I have come across (including Sendmail, Exim and Postfix - the three most populare FOSS examples) has had the ability to do delivery receipts, except every configuration I have ever seen has had it turned off... Likewise, Thunderbird and Evolution and so forth will also do read receipts. Likewise also I have never seen anyone turn it on (and why would you?)
Delegation is also pretty straightforward - it is really only a permission model and there are several variations on permission models that could be installed, depending on the sort of needs of an organisation.
Message recall is something that doesn't exist once it's left the company mail server, and I do find it occasionally amusing to see people attempting to recall e-mail they've thoughtlessly banged off to a mailing list. Even internally you'll be lucky to get away with it on a mail server that's notifying all recipients when a new mail arrives. People who think they can take back those words are making a mistake...
The current outstanding problem with FOSS "e-mail" applications is none of these, and is simply about sharing and scheduling calendaring functionality with an off-line client. It's a known problem and there are many attempts at a solution out there which are stabilising. If I knew which ones were going to be a success I would be very happy, but there are at least five efforts that are seriously close to being ready for real time and I'm confident something will get there in the coming 12-18 months.
Calendaring is only one issue with migration to FOSS desktops though. In my experience it's the corner cases in the organisation that make it hard, and I tend to recommend goals of "get 80% switched" for organisations that ask my help in migrating". It works fine that way, and there's a lot less stress for everyone concerned. The the other 20% just end up migrating naturally as new applications get developed (internally or externally) to meet their needs.
My son has just turned 8, and he loves Lego. We buy new stuff regularly (and the odd tub of basic pieces) and he makes them up, pulls them apart and they all end up in that enormous soup of Lego blocks. He makes some amazing stuff from that primordial soup, none of it scripted, but he frequently does use some of those parts that came with the Spiderman sets, or the Harry Potter sets or the X-Pods or the Orient Expedition sets or the Star Wars sets.
In fact I don't see that things have changed a whole heap, except that with a big pile of Lego now you can make a damn site more interesting things than I could thirty years ago when I used to babysit for some kids who had Lego (I had Meccano as a kid myself). I used to build houses (well, with bricks, windows and roofs what else are you going to make?) and the kids I was babysitting for used to play with them for the next few weeks. When their parents would tell them they were going out and they would have a babysitter they would destroy everything, in the hope that I would have some more fun with their lego and they could have some more fun playing with a new set of designs.
Sure, so I never read the books and it was just purely creative play. My son's read all the instructions, but that just doesn't challenge him and he moves on.
Lego, on the other hand, has substantially more variety than it did when I was a kid, and that means that what can be created is exponentially more varied.
Great article though. It would be nice to see Lego producing some of the sorts of kits that are suggested at the end, and perhaps that is what the X-Pods do, and some of the other things that encourage the kids to build and rebuild in different configurations.
Ubuntu might be popular within its own community, but the distro won't go mainstream until its image matures past high school sophomore.
To me, the name "Dapper Drake" conjures up images of some big daddy duck, decked out in his finest feathers and preened to the max. Drawn by one of those totally masterful more-detail-than-real-life children's book artists like Jane Hissey.
I've seen some blog review of Ubuntu that endlessly puzzles over the naming conventions, and I think it's just hilarious. For fuck's sake: IT IS A MEMORABLE NAME! Marketing and branding is partly about exposure, but without being memorable that exposure doesn't travel a tenth the distance.
As for it's going mainstream: I am going to have fewer problems rolling out Ubuntu (Warty, Hoary, Breezy, Dapper) to clients than I have had rolling out Debian (Bo, Hamm, Slink, Woody, Sarge, Etch). Not that they hugely care about the release name, and if they do, I'm sure I'll be able to spot their humourlessness a mile away and refer to it as "Ubuntu 7.04, sir".
In fact I think that their version numbering is also masterful. To think this is only the third Ubuntu release coming up, but it's already 6.10 - what an awesome dodge. No mucking around trying to "justify" the version numbers on the basis of swank new functionality! And perhaps even more importantly: no more mucking around trying to justify swank new functionality on the basis of incrementing the version number. Excellent.
Mark has done an awesome thing for the software community by following this model that he has chosen. It may well be chump change for him to have done so, but many wealthier people are considerably less philanthropic. He is so very passionate and involved in this project, and so astutely sticking to his vision that I don't doubt that Canonical will end up paying it's way, eventually, but it will truly be a side-effect of that vision and passion.
I recommend you take a look at the video of him talking at Debconf (or of his informal talk there) which someone linked above.
And in two or three years, when Ubuntu 8.04 is coming out, perhaps Mark will be starting to look around for something else to do. I really wonder what it will be.
Dude. I don't know what you've been smoking, but the Ubuntu installer is the
Debian installer.
At this point, anyway. I believe there are hopes to maybe have a GUI installer
for the next version, or maybe the one after that.
Debian can't do this as easily because of the complexity of doing it across all supported architectures. And in fact that was one of the complicating factors for the installer that Debian introduced for Sarge (i.e. the one that is used in Ubuntu), but that work has paid off, and now there is a good modular installer that is a solid foundation for the future.
Wouldn't Rsync be better for what you want? Why do you need to be able to choose different versions to fetch?
If the files contains parts that are constant along with parts that vary then rsync will in many cases only transfer the partial file. With Subversion that won't apply for binary files, but rsync will still recognise partial matches even on those.
The packaging methods for each distribution aren't that incredibly hard, but they do make it incredibly easier to install the software.
If you do this, you should really only need to build RPM and DEB (but keep TGZ around for people this doesn't work for, as it is even more trivial).
Where I work we build Debian packages of all sorts of things, simply because it is so easy. We even build Debian packages to "install" a new user on a computer. It ain't hard to maintain after you've done it the first time - if it was, then a lot of Debian developers would find it difficult to do as volunteers.
The article makes a great point. The RIAA gets to oversee music. The MPAA gets to oversee movies. The ESRB is impotent and the goverment must oversee games.
I must confess it surprises me that these ratings are not independent in the US. They certainly are in New Zealand, where I live, and this does not appear to cause problems with unreasonable ratings.
I would be surprised if the US is in the majority on this point, and there certainly would appear, prima-facie, to be a good case for suspicion of partiality.
Yeah, except that desktop systems aren't asked to suspend very often.
Suspend is broken in the ATI fglrx drivers, and has been for the eighteen months that I have owned my current laptop. For this reason I use the 2D drivers in the standard X.Org release, although I am hopeful that the r300 project is showing some real traction now.
We simply tell our clients (who are all windows users anyway) to use IE.
Thus limiting your future clients to people who haven't already chosen a different browser.
I also consult to a company who has decided for various reasons to mandate the use of Firefox on all company desktops. They did have one application system that people used which was not cross-browser compatible, and they elected to discontinue their commerce with that supplier.
Note that the commerce that got discontinued had nothing to do with the people who wrote that application. The application in question was an extranet to allow the supplier to be closer to their clients, but in this case it was having exactly the opposite effect.
That was an extranet application, and clearly making that browser-specific is stupidity on a vaster scale than merely making an intranet browser-specific. But then some of the best intranet applications will end up being accessed with other browsers too, whether simply because they last longer than versions, and have to support IE5 -> IE7, or because they broaden their reach and get rolled out for extranet or internet users...
I guess you need to live in hope that your own applications aren't useful enough for that sort of fate!
We discussed having torrents, but the host for the videos supposedly has multiple Gb to the internet and is better able to handle streaming (if the client software supports it) than torrents.
I'd be interested to know what the bandwidth usage is, but I doubt that the bandwidth is challenged unless a lot of people are downloading 150MB videos by accident. I'd expect more cherry-picking than mirroring (it's a total of 7G or so).
Feel free to set up a torrent though - this is open source, scratch-your-own-itch territory, after all...:-)
Here in New Zealand I use one of these for remote access, and sometimes it's great and sometimes it ain't.
Networks in New Zealand come in three flavours:
CDMA 1x
GSM
CDMA 1x EVDO
The best of these is definitely the 1x EVDO network. If I get on that one I get ping times down around the 100mS mark, and surfing is good, ssh is good, everything is good. Well good, as in "gee, I can surf the intarweb thingy from the back of a taxi", but it doesn't compare to WLAN or wired speeds really.
The other two are significantly less good. I haven't used the GSM much because my own card is CDMA, but one of my co-workers has a GSM card, and it seems about the same as basic CDMA, perhaps a bit slower raw bandwidth, but maybe a bit better on latency. The latency on the CDMA often sucks. In poorer coverage areas 600mS is normal, and I have seen it all over the map, from 400mS to 2000mS. That sort of latency can then lead to DNS lookup failures, and so surfing suffers. SSH is really laggy, and so forth. Raw downloads are OK-ish.
In my case I'm using these cards under Linux, where they work, but no real support from the supplier for them in my environment. It's OK, though, so I wrote my own page about the Aircard 580 on Linux to help people get it going.
The only setbacks seen with the fglrx drivers would be that of the mentioned lack of XRandR support as well as a lack of XCompMgr support (for drop shadows/transparency). However, such minor setbacks on 'beauty' shouldn't be a big decision when choosing which drivers to use.
Of course corporate environments have to essentially build their own Windows distribution for a consistent desktop across the enterprise.
The tools to do this for Linux distributions aren't as immature as you think. I have designed a custom Linux installation for a smaller corporate desktop release (~140 desktops) and I doubt if it was really much more work than a Windows rollout would have been.
Once we had the release out there though I think that the tools for managing upgrades and provisioning case-by-case variations within that environment were a lot easier.
Of course that could just be a reflection of my greater experience with linux, or perhaps that's the way an unbiased observer would see it too.
If the difference is arguable though it probably really does come down to personal preferences.
It is very interesting the assumptions that they state have been made to bias this report in Microsoft's favour.
He said given the fact that the company deals in open source products, four aspects had been factored in to tip the scales towards Microsoft:
The model was not modified to to reflect research by the Robert Frances Group which showed that Linux needed 82 percent fewer staff resources.
The costs of malware - viruses, spyware, worms, keyloggers, adware - were not taken into account. Zymaris said every research point found had suggested that this cost was essentially and predominantly a Windows platform cost, resulting in billions lost by business every year.
Costs which arose when systems need to be pre-emptively rebooted or crashed, resulting in unscheduled downtime, were not taken into account. "All our research indicates that Linux rarely if ever suffers such problems and open source platforms on the whole are extremely robust," Zymaris said.
"Finally, because Microsoft has claimed that introducing Linux into an environment will lead to increased reliance on external consultants, we have tripled the amount budgeted for such requirements on the Linux models," he said.
Indeed. Looking at the stats for Stuff.co.nz - one of New Zealand's largest news sites - I see Firefox currently at around 8-9% and the total for all of Mozilla at around 13-14%. That's on traffic of around 7-8 million hits per day.
Not a geek site this one - Linux usage is around 1%.
Well I don't know about Gentoo (although I doubt it's any different), but my company provides "a phone number that can get a live trained tech to fix the problem" for Debian users.
This is a service that we offer to our clients, and it works well. After all, we have full access to the source code, so why should we not be able to fix the problem? There are companies all over the world who happily provide this kind of service - what we do is not unusual.
So-called "enterprise-level" support in the Microsoft world still also has a lot to do with your effectiveness at searching google. I haven't personally come across an MS geek who has expected to automatically have problems solved by (e.g.) calling MS.
The point, for Venezuela, I guess is that the support can be a local call. The experts can be speaking the same language, with the same dialect, and still be experts. And the bottom line is that they will be paid at local rates, and the money won't be getting shipped out of the country.
Slashdot Mirror
Looks like an early christmas this year. I've fixed the reasons why it wouldn't work in a subdirectory now, so it will be able to operate within a subdirectory of a virtual host from the next release.
I expect that will be in the next couple of days.
Regards,
Andrew McMillan.
Hi,
I've put a "don't require a whole virtual server" thing on the road map now, and will see if I can have that done before christmas.
The "caldav.php" in the URL can be pretty easily got rid of with a URL rewriting line, so I might add a note on how to do that to the installation docs, or somewhere in the Wiki as well.
Regards
Andrew McMillan.
Hmmmm...
Today I disabled register_globals to see what broke when I did so. To my surprise it appears that nothing does :-)
So I guess I need to update my documentation on that point!
Cheers,
Andrew McMillan.
Cosmo does, of course. I don't intend to do that with RSCDS, although the current code (in Git) will let you see basic information it does not let you do any add / edit of the calendar entries, and presents events in a list, rather than as dates on a calendar.
If you do have any suggestions for improvements in RSCDS, please tell me about them though :-)
Thanks,
Andrew McMillan.
Yeah, I know. It uses register globals in a safe way, but I need to bite the bullet one day and just disable it and go through my libraries and find where it's used. Probably I'm at that point now and should Just Do It :-)
The problem is basically that the libraries I use have been written over the last five years or so and there are some places where that is expected. For all of the code written in the last couple of years (including the actual RSCDS code) it isn't required.
Happy to accept patches, of course :-)
Andrew McMillan.
Clearly this is contradictory. If you "can't wait" until it hist Debian stable then you are looking for a release which is less stable than Debian stable.
The reality is, of course much simpler. Odds are that given it's optional "runs on top of Xorg" nature it will be available in Debian testing within 3 months and will consequently be released next time the 16000 or so Debian packages are declared stable enough for a release.
</TrollFodder>
Similar to the main linked site, GeShi is the product of a very smart young coder (who I happen to employ) and who does not have the funds to pay for bandwidth. He doesn't advertise on the site, and GeShi is GPL and something he did for fun in his spare time.
I'll see if we can get it onto some better bandwidth / equipment for him so he can cope with Slashdot in the future, but it's not going to happen during this event.
Well, I think we're trying to talk about things that FOSS solutions can't do here...
Out of office replies are easy, and generally much better than the brain-dead activities of Exchange. You can say "only send one out of office reply to any given individual every seven days" for example. Or "don't send out of office replies to automatic e-mails". Exchange can't do that (or if it can, someone please tell me - I would love a clue-stick to beat some admins with...:-)
I doubt if anyone would seriously suggest that the sorts of server-side filtering you can do with Procmail are lacking in functionality. And if they are, there's always Perl... The problems here are maybe in the UI, but that's fairly easily fixable and something I don't know about probably already has.
Read receipts are likewise easy. Every MTA I have come across (including Sendmail, Exim and Postfix - the three most populare FOSS examples) has had the ability to do delivery receipts, except every configuration I have ever seen has had it turned off... Likewise, Thunderbird and Evolution and so forth will also do read receipts. Likewise also I have never seen anyone turn it on (and why would you?)
Delegation is also pretty straightforward - it is really only a permission model and there are several variations on permission models that could be installed, depending on the sort of needs of an organisation.
Message recall is something that doesn't exist once it's left the company mail server, and I do find it occasionally amusing to see people attempting to recall e-mail they've thoughtlessly banged off to a mailing list. Even internally you'll be lucky to get away with it on a mail server that's notifying all recipients when a new mail arrives. People who think they can take back those words are making a mistake...
The current outstanding problem with FOSS "e-mail" applications is none of these, and is simply about sharing and scheduling calendaring functionality with an off-line client. It's a known problem and there are many attempts at a solution out there which are stabilising. If I knew which ones were going to be a success I would be very happy, but there are at least five efforts that are seriously close to being ready for real time and I'm confident something will get there in the coming 12-18 months.
Calendaring is only one issue with migration to FOSS desktops though. In my experience it's the corner cases in the organisation that make it hard, and I tend to recommend goals of "get 80% switched" for organisations that ask my help in migrating". It works fine that way, and there's a lot less stress for everyone concerned. The the other 20% just end up migrating naturally as new applications get developed (internally or externally) to meet their needs.
My son has just turned 8, and he loves Lego. We buy new stuff regularly (and the odd tub of basic pieces) and he makes them up, pulls them apart and they all end up in that enormous soup of Lego blocks. He makes some amazing stuff from that primordial soup, none of it scripted, but he frequently does use some of those parts that came with the Spiderman sets, or the Harry Potter sets or the X-Pods or the Orient Expedition sets or the Star Wars sets.
In fact I don't see that things have changed a whole heap, except that with a big pile of Lego now you can make a damn site more interesting things than I could thirty years ago when I used to babysit for some kids who had Lego (I had Meccano as a kid myself). I used to build houses (well, with bricks, windows and roofs what else are you going to make?) and the kids I was babysitting for used to play with them for the next few weeks. When their parents would tell them they were going out and they would have a babysitter they would destroy everything, in the hope that I would have some more fun with their lego and they could have some more fun playing with a new set of designs.
Sure, so I never read the books and it was just purely creative play. My son's read all the instructions, but that just doesn't challenge him and he moves on.
Lego, on the other hand, has substantially more variety than it did when I was a kid, and that means that what can be created is exponentially more varied.
Great article though. It would be nice to see Lego producing some of the sorts of kits that are suggested at the end, and perhaps that is what the X-Pods do, and some of the other things that encourage the kids to build and rebuild in different configurations.
To me, the name "Dapper Drake" conjures up images of some big daddy duck, decked out in his finest feathers and preened to the max. Drawn by one of those totally masterful more-detail-than-real-life children's book artists like Jane Hissey.
I've seen some blog review of Ubuntu that endlessly puzzles over the naming conventions, and I think it's just hilarious. For fuck's sake: IT IS A MEMORABLE NAME! Marketing and branding is partly about exposure, but without being memorable that exposure doesn't travel a tenth the distance.
As for it's going mainstream: I am going to have fewer problems rolling out Ubuntu (Warty, Hoary, Breezy, Dapper) to clients than I have had rolling out Debian (Bo, Hamm, Slink, Woody, Sarge, Etch). Not that they hugely care about the release name, and if they do, I'm sure I'll be able to spot their humourlessness a mile away and refer to it as "Ubuntu 7.04, sir".
In fact I think that their version numbering is also masterful. To think this is only the third Ubuntu release coming up, but it's already 6.10 - what an awesome dodge. No mucking around trying to "justify" the version numbers on the basis of swank new functionality! And perhaps even more importantly: no more mucking around trying to justify swank new functionality on the basis of incrementing the version number. Excellent.
Mark has done an awesome thing for the software community by following this model that he has chosen. It may well be chump change for him to have done so, but many wealthier people are considerably less philanthropic. He is so very passionate and involved in this project, and so astutely sticking to his vision that I don't doubt that Canonical will end up paying it's way, eventually, but it will truly be a side-effect of that vision and passion.
I recommend you take a look at the video of him talking at Debconf (or of his informal talk there) which someone linked above.
And in two or three years, when Ubuntu 8.04 is coming out, perhaps Mark will be starting to look around for something else to do. I really wonder what it will be.
Dude. I don't know what you've been smoking, but the Ubuntu installer is the Debian installer.
At this point, anyway. I believe there are hopes to maybe have a GUI installer for the next version, or maybe the one after that.
Debian can't do this as easily because of the complexity of doing it across all supported architectures. And in fact that was one of the complicating factors for the installer that Debian introduced for Sarge (i.e. the one that is used in Ubuntu), but that work has paid off, and now there is a good modular installer that is a solid foundation for the future.
Wouldn't Rsync be better for what you want? Why do you need to be able to choose different versions to fetch?
If the files contains parts that are constant along with parts that vary then rsync will in many cases only transfer the partial file. With Subversion that won't apply for binary files, but rsync will still recognise partial matches even on those.
The packaging methods for each distribution aren't that incredibly hard, but they do make it incredibly easier to install the software.
If you do this, you should really only need to build RPM and DEB (but keep TGZ around for people this doesn't work for, as it is even more trivial).
Where I work we build Debian packages of all sorts of things, simply because it is so easy. We even build Debian packages to "install" a new user on a computer. It ain't hard to maintain after you've done it the first time - if it was, then a lot of Debian developers would find it difficult to do as volunteers.
I must confess it surprises me that these ratings are not independent in the US. They certainly are in New Zealand, where I live, and this does not appear to cause problems with unreasonable ratings.
I would be surprised if the US is in the majority on this point, and there certainly would appear, prima-facie, to be a good case for suspicion of partiality.
Yeah, except that desktop systems aren't asked to suspend very often.
Suspend is broken in the ATI fglrx drivers, and has been for the eighteen months that I have owned my current laptop. For this reason I use the 2D drivers in the standard X.Org release, although I am hopeful that the r300 project is showing some real traction now.
Thus limiting your future clients to people who haven't already chosen a different browser.
I also consult to a company who has decided for various reasons to mandate the use of Firefox on all company desktops. They did have one application system that people used which was not cross-browser compatible, and they elected to discontinue their commerce with that supplier.
Note that the commerce that got discontinued had nothing to do with the people who wrote that application. The application in question was an extranet to allow the supplier to be closer to their clients, but in this case it was having exactly the opposite effect.
That was an extranet application, and clearly making that browser-specific is stupidity on a vaster scale than merely making an intranet browser-specific. But then some of the best intranet applications will end up being accessed with other browsers too, whether simply because they last longer than versions, and have to support IE5 -> IE7, or because they broaden their reach and get rolled out for extranet or internet users...
I guess you need to live in hope that your own applications aren't useful enough for that sort of fate!
We discussed having torrents, but the host for the videos supposedly has multiple Gb to the internet and is better able to handle streaming (if the client software supports it) than torrents.
I'd be interested to know what the bandwidth usage is, but I doubt that the bandwidth is challenged unless a lot of people are downloading 150MB videos by accident. I'd expect more cherry-picking than mirroring (it's a total of 7G or so).
Feel free to set up a torrent though - this is open source, scratch-your-own-itch territory, after all... :-)
Here in New Zealand I use one of these for remote access, and sometimes it's great and sometimes it ain't.
Networks in New Zealand come in three flavours:
The best of these is definitely the 1x EVDO network. If I get on that one I get ping times down around the 100mS mark, and surfing is good, ssh is good, everything is good. Well good, as in "gee, I can surf the intarweb thingy from the back of a taxi", but it doesn't compare to WLAN or wired speeds really.
The other two are significantly less good. I haven't used the GSM much because my own card is CDMA, but one of my co-workers has a GSM card, and it seems about the same as basic CDMA, perhaps a bit slower raw bandwidth, but maybe a bit better on latency. The latency on the CDMA often sucks. In poorer coverage areas 600mS is normal, and I have seen it all over the map, from 400mS to 2000mS. That sort of latency can then lead to DNS lookup failures, and so surfing suffers. SSH is really laggy, and so forth. Raw downloads are OK-ish.
In my case I'm using these cards under Linux, where they work, but no real support from the supplier for them in my environment. It's OK, though, so I wrote my own page about the Aircard 580 on Linux to help people get it going.
Yeah, and ain't it interesting that this year's DebConf is in Helsinki, Finland, and that it's at least partly sponsored by Nokia...
I'm off to Helsinki for Debconf5 in five weeks myself, and I'm hoping to see if I can lay my grubby little mitts on one of these devicen :-)
And also the ATI inability to survive a suspend / resume cycle.... That's the one that kills the fglrx drivers every release for me, and I've had this laptop for 18 months now with that issue.
Hopefully it'll be working before I hand it on to someone else :-)
The tools to do this for Linux distributions aren't as immature as you think. I have designed a custom Linux installation for a smaller corporate desktop release (~140 desktops) and I doubt if it was really much more work than a Windows rollout would have been.
Once we had the release out there though I think that the tools for managing upgrades and provisioning case-by-case variations within that environment were a lot easier.
Of course that could just be a reflection of my greater experience with linux, or perhaps that's the way an unbiased observer would see it too.
If the difference is arguable though it probably really does come down to personal preferences.
It is very interesting the assumptions that they state have been made to bias this report in Microsoft's favour.
Wow!
For the number of times "ii" occurs in english you could save yourself a character, right there.
Now, I'm off for a bit of skiing...
Indeed. Looking at the stats for Stuff.co.nz - one of New Zealand's largest news sites - I see Firefox currently at around 8-9% and the total for all of Mozilla at around 13-14%. That's on traffic of around 7-8 million hits per day.
Not a geek site this one - Linux usage is around 1%.
Well I don't know about Gentoo (although I doubt it's any different), but my company provides "a phone number that can get a live trained tech to fix the problem" for Debian users.
This is a service that we offer to our clients, and it works well. After all, we have full access to the source code, so why should we not be able to fix the problem? There are companies all over the world who happily provide this kind of service - what we do is not unusual.
So-called "enterprise-level" support in the Microsoft world still also has a lot to do with your effectiveness at searching google. I haven't personally come across an MS geek who has expected to automatically have problems solved by (e.g.) calling MS.
The point, for Venezuela, I guess is that the support can be a local call. The experts can be speaking the same language, with the same dialect, and still be experts. And the bottom line is that they will be paid at local rates, and the money won't be getting shipped out of the country.