If people had understood how patents would be granted when most of today's ideas were invented, and had taken out patents, the industry would be at a complete standstill today.
Evidently, Trey now wants the industry at a complete standstill.
...what with the required OCR and all that. Your search would work if you were searching for EXIF tags rather than pixels, which is what I expect MS to be doing, probably in a grossly inefficient and resource-intensive way.
My book-keepers installed a driver for a digital camera, and Windows ME stopped booting until they booted the machine from an install CD instead, hit F3 and used COMMAND to rename a system file. No reinstall, at least, but unquestionably outside the expertise of a typical computer user to fix.
we have to deal with a Linux shop because they always tell us nonsense things about our Windows XP servers
What? Like "it doesn't exist?" Bwuhahahahahaaaaa! (-:
You probably mean either MS-Windows 2003 Server, or MS-Windows XP Pro running somethign as a service. And you probably wonder why you have trouble understanding what the fellas from "the Linux shop" tell you.
Turkey. (-:
You really won't ruin x-mas if you try to plug a digital camera into a PC!
Most of the useful Linux packages can be found (albeit sometimes a bit crippled) for MS-Windows (or nicely packaged for the Mac). What you don't usually get is smooth integration, full-throttle performance or regular updates. OTOH, very few attackers expect to find, for example, EXIM on MS-Windows, so attempts to priv-escalate to root don't work all that well. (-:
The advantages of Linux which my cutomers see (all of my customer sites are either Linux or planning to be Linux soon) are that I can quickly and cheaply set it up and then more or less forget it for a couple of years until a fan or hard disk fails (and yes, I do have and have had systems with years of continuous uptime on them); it's flexible enough to readily do a lot of stuff which is difficult-to-impossible (and usually clumsy) under MS-Windows; you don't need to worry about licensing; and it's never had a CodeRed or MSBlast equivalent. Different customers have additional different reasons for liking it, but that's all pretty much common ground. Cheap, reliable and flexible.
I don't see how the GP can price MS-Windows and and Linux systems at about the same on the same hardware, since the sites I share with MS-Windows all (bar one, who is ultra-careful about everything he does and visits the 40-seat site maybe every fortnight) have the MS techs constantly visiting to fix stuff up that should never break.
In the Windows world, you can distribute a product in a single self-installing executable for many versions of Windows.
I think that's called "Stable API - at least where it matters most". If I click on an RPM in most of my web browsers or file managers (but usually Konqueror) it will offer to install it, and usually succeed. It's not yet reasonable to expect that to work across platforms (click on a.i586.rpm from a PPC machine and have it work).
MS-Windows packages generally wrap everything that they're likely to need, which can make for DLL hell, and can result in you running multiple copies of a facility, some of which are easy to upgrade and some of which are not; Linux package managers are much more granular, typically relying on the PM to find any needed dependencies and so getting them all from one place. This gives the administrator much more control over the app, particular quality control WRT dependent libraries, but until the Linux PMs all learned to resolve their own dependencies it had meant that the admin had to chase down other RPMs to bridge any gaps. Hooray for the march of progress. I'd expect the next generation of Linux package managers to be even more ambitious about hunting for dependencies, and to be intelligent enough to be able to find dependencies for and build from a.src.rpm if no suitable binary is to hand (or if you're an optimisation freak). Gentoo, of course, already does most of this.
Right now 99% of applications and services have proprietary config files. Not proprietary in the "closed" sense, but in the "different from each other one" sense.
Generally true. There is, however, a growing degree of consistency in where apps look for those files; in the case of a service it's usually/etc/nameofservice.conf or/etc/nameofservice/section.conf, in the case of an application it usually goes/usr/share/appname/appname.conf; ~/.appname.conf (or ~/.appname/section.conf). Systems like KDE have extended this to a tree of ~/.kde/share/apps/conf/appnamerc for KDE-aware services, and have internally done a lot of the homogenisation which you seek.
If that rebuild-from-source allowed you to do something difficult-or-impossible-or-expensive in MS-Windows land, it would be well worth it.
Recent case-in-point, if you want email virus scanning for 200 users, a.src.rpm fetch and "rpm -bb" of clamav would be a top investment even if you had to sit and watch it instead of going about your other business when you look at the cost of (for example) Sophos or Norton for 200 seats.
Another one, hunting down dependencies for AMaViS is worthwhile (my preferred distro, Mandrake Linux, has them all and it's just "urpmi amavisd" but this customer's site was running Red Hat) if you can discard mail containing semi-broken ZIP files (as sent by mutating MS-Windows viruses) where the commercial packages' attitude seems to universally be "can't read it, therefore it's not really a ZIP, therefore it's safe". Having this crap cut out at the gateway meant that the customer's internal MS-Exchange server could then handle the remaining load. Usually.
If I seriously wanted to upgrade software that old, I'd cross-grade to Mandrake Linux 6.0, then upgrade by x.0's from there. You might get away with upgrading in one jump, but your odds are not good either way.
IRL, I'd copy off the config and any databases, do a fresh install of a recent distro, then feather the config changes back in by hand.
Once you get up to Mandrake 9.0, a distro upgrade is a matter of destroying your old URPMI sources, defining new ones, updating the bits needed for RPM and URPMI, updating everything else, updating your kernel and you're done. I've done this on live (non-critical) servers.
Debian and APT is also that good, only you're not going to have as much luck with that first upgrade step (packaging system is completely different, filesystem structure is significantly different).
An equivalent question is "What would you upgrade Windows 3.1 to?"
Just copying files is one thing, backup up databases without bringing them offline is another.
With a typical Unix MTA, there is no practical difference.
PostFix uses databases by default (for efficiency reasons) for some of its config (like mapping email addresses to mailboxes) but these change very rarely (typically months to days between changes) and almost never after hours when backups are typically done. Even so, you could use flat files for the job and it would work fine. Everything else is just a file. Queues, spools, mailboxes etc are all just flat files.
This also means that access control etc is universal and consistent. MS had to write separate access controls for MS-Exchange (have a look at the origins of that one day, you'll be amazed that it works at all), and again for the Registry. Each set is different and complex, each set required work to set up, and work to maintain, and offers a separate opportunity to screw things up. Exchange has never been KISSed. (-:
It's worth adding that if the Linux guy is twice as expensive as the MS-Windows guy, it's because they have less than half as much work to do. Linux admins, in the case of small shops, actually get to leave again, where an MS-Windows admin will almost always be back every week or two.
That's like noticing that Fred seems to be doing fine at shuttling stuff around for the company with his ute, so it's OK to seat him in a semi-trailer (without training), load it up with sixty or seventy tonnes of gear, and send him out the gate.
At the first tight corner, your mistake will become obvious from the sounds of street furniture being destroyed, and when Fred gets to the highway your mistake will be backgrounded with the wail of accumulating sirens as the emergency vehicles arrive to free the survivors and cart away the dead.
Getting computer services right can be a lot more complex than piloting a long, heavy, unresponsive vehicle - although getting PostFix right for a single domain is usually a matter of selecting the right hostname and doing the occasional update.
FOAF living in the US (we live in Oz) kept getting bats infiltrating their garage, and they also keep their meat freezer in the same garage. One day, wife (who loathes bats) asks husband to do something about them, "and while you're out there, can you grab a couple of steaks?" Husband responds, "to drive through their little hearts, dear?" (-:
I agree that MS-Windows zealots will continue to do their thing until some years after Microsoft implodes and running their software ceases to be an option.
Well, actually, no. As I point out above, PostFix works by default out of the box. For the Microsoft Channel Partner involved (anyone smell a rat, there? anyone...?:-) to have failed to get it working, they would have had to break it themselves.
For instance, in Argentina, they used Grimaldi, a shoe manufacturer as the example. When you dig into the story, you discover that the company that was supposed to carry out Grimaldi's migration to Linux is a Windows certified partner and a windows-only shop. The idiots could not get sendmail or postfix up-and-running and thus claim that it doesn't work.
Um, PostFix works by default. They'd have to stop it from working in order to not get a positive result.
Take a typical Mandrake Linux install. You select "PostFix" as one of the packages to install, from the servers section under "Mail and News" - after the installer has asked you for a hostname, and the obvious one is mail.grimaldi.com.ar.
The PostFix service is started by default if it's installed, and it figures out from the hostname that it's doing mail for grimaldi.com.ar. You'd have to turn it off by hand to stop it from working at this point. Morons!
If you go to www.grimaldi.com.ar and look at the source for that placeholder, I think it tells you all you need to know about the competence of Grimaldi's IT section. The page is one centred line of plain text on a plain background, but it is delivered through a frameset (and those are referring to a fixed IP address). Surprisingly, they do actually have a <noframes> clause, and it's more informative than the main page.
When Linux and OpenOffice change from a thorn to a threat, Microsoft will turn on Open Source with every trick in their book, clean and dirty.
They're scraping the bottom of the barrel already.
Microsoft have never had to fight a competitor like FOSS before, which metaphorically plays its poker with all of their cards face-up on the table and wins anyway. Having half of the cards in play faced up makes it very difficult to cheat in the traditional ways, and the elimination of half of the bluff radically alters the game.
The other terrifying aspect is that in order to play poker at this table, Microsoft needs to face not one or two other players, but countless thousands. Their pin-striped suit and expensive jewelry isn't impressing the other players as much as it ought to, and it's hard to keep track of all of those other hands despite their all being faced up.
There are a few "enemy concentrations" which they can focus on and kind of play against as a collective, like Samba (they've been seriously down about the mouth on Samba for the last year or two), Linux (the FOSS poster child) and OpenOffice ("they wrote what?!"), which between them attack most of the profit avenues represented by the two cash cows, MS-Windows and MS-Office.
However, MS had kind of saturated those two markets, and needs to break into new markets - and guess what? Wherever they turn, FOSS is already there. So Microsoft have turned to peripheral areas, hardware and patents and social/political stuff that's difficult for FOSS to directly compete on.
Hardware is not much of a problem for FOSS, Microsoft has conventional competitors there to keep them in line as long as they can't leverage a Palladium-like monopoly.
Patents are a problem, but I think the whole patent arena is sliding out from under the cover of being merely sublime, and even conservative lawmakers are going to have to take notice of that soon.
This leaves the more nebulous area of social/political manipulation. Traditionally, Microsoft has used its massive presence, advertising and schmoozing budgets to leverage its core markets, together with the implied access to a lot more back doors (and we're not talking BackOrifice here). If they find a more direct way to turn all of that power into a cash-flow, the world is in trouble.
Good old "oops, we forgot to renew the domain - and the other domain" Passport is gradually jamming a foot into that kind of door, in the same way as constantly waving MSIE and MSN in people's faces has saved those items from the total obscurity they deserve. Watch out for Microsoft jamming levers in elsewhere.
Nevertheless, the FOSS community in general refuses to take Microsoft as seriously as MS'd like, and they often react in unpredictable was and predictable ways at the same time, which must cause endless ulcers in Redmond because they can never be sure that the predictable responses from the FOSS crowd aren't masking the launch of unpredictable responses (such as handing out FOSS CDs at Microsoft conventions, a tactic which Microsoft seems to be reciprocating by flooding FOSS_friendly websites with their ads).
...with the exception of the "write the serial modem code" argument, which would be reasonable if the code hadn't already been written for Linux (including emulators), none of your arguments hold water.
You can write and run your software on an ordinary Linux desktop, without even the need for a simulator module for the most part. The same Linux which runs your desktop can also run your 'plane, and for a considerably smaller performance and resource hit than XP (although in real life you'd go to the trouble of having a highly modular kernel and not load very much for the instance on the 'plane).
The important point is not the changes to the menus, it's that this reflects underlying conceptual and structural differences in the programs. OOo was raised as a specific, planned exception to the trend established by so many other FOSS programs. Kopete, for example, looks nothing much like MSN Messenger, but it does the same job (and also AIM's job, ICQ's job and so on).
You also haven't spoken to the points I raised about KMail.
Use it, love it, and the only concession to Outlook is the toolbar thingy they added with Kontact.
OOo (well, StarOffice) was deliberately designed to look like MS-Office, not for the sake of slavish cloning, but to make it easy for MS-Office users to transition. And it worked. There are still significant structural differences, generally for the better. Forex, in ooWriter you have Format/Character, then Format/Paragraph, then Format/Page, one after another on the menu. In MS-Word, you have Format/Character, Format/Paragraph, and... File/Page-Setup? This also reflects an internal difference in how they view documents. Sections are less foundational to ooWriter than MS-Word, and frames work better in ooWriter (and it produces nicer HTML when used for that).
KOffice is a little differenter, but it also does some things righter, and it's having more tools added to it than OOo. It is definitely not an MS-Office clone, and it is also quite popular.
You also neglected to speak on FireFox and countless other FOSS tools which are unquestionably not clones.
Slashdot Mirror
Evidently, Trey now wants the industry at a complete standstill.
Innovating in the IM field would be naff-all use if there were nobody else to talk to. Nevertheless, I cite Jabber.
Now... get a life.
...what with the required OCR and all that. Your search would work if you were searching for EXIF tags rather than pixels, which is what I expect MS to be doing, probably in a grossly inefficient and resource-intensive way.
My book-keepers installed a driver for a digital camera, and Windows ME stopped booting until they booted the machine from an install CD instead, hit F3 and used COMMAND to rename a system file. No reinstall, at least, but unquestionably outside the expertise of a typical computer user to fix.
You probably mean either MS-Windows 2003 Server, or MS-Windows XP Pro running somethign as a service. And you probably wonder why you have trouble understanding what the fellas from "the Linux shop" tell you.
Turkey. (-:
Well, hey, something like that worked for Bill, why not for you? (-:
Wrong, wrong (just), wrong!. (-:
Most of the useful Linux packages can be found (albeit sometimes a bit crippled) for MS-Windows (or nicely packaged for the Mac). What you don't usually get is smooth integration, full-throttle performance or regular updates. OTOH, very few attackers expect to find, for example, EXIM on MS-Windows, so attempts to priv-escalate to root don't work all that well. (-:
The advantages of Linux which my cutomers see (all of my customer sites are either Linux or planning to be Linux soon) are that I can quickly and cheaply set it up and then more or less forget it for a couple of years until a fan or hard disk fails (and yes, I do have and have had systems with years of continuous uptime on them); it's flexible enough to readily do a lot of stuff which is difficult-to-impossible (and usually clumsy) under MS-Windows; you don't need to worry about licensing; and it's never had a CodeRed or MSBlast equivalent. Different customers have additional different reasons for liking it, but that's all pretty much common ground. Cheap, reliable and flexible.
I don't see how the GP can price MS-Windows and and Linux systems at about the same on the same hardware, since the sites I share with MS-Windows all (bar one, who is ultra-careful about everything he does and visits the 40-seat site maybe every fortnight) have the MS techs constantly visiting to fix stuff up that should never break.
MS-Windows packages generally wrap everything that they're likely to need, which can make for DLL hell, and can result in you running multiple copies of a facility, some of which are easy to upgrade and some of which are not; Linux package managers are much more granular, typically relying on the PM to find any needed dependencies and so getting them all from one place. This gives the administrator much more control over the app, particular quality control WRT dependent libraries, but until the Linux PMs all learned to resolve their own dependencies it had meant that the admin had to chase down other RPMs to bridge any gaps. Hooray for the march of progress. I'd expect the next generation of Linux package managers to be even more ambitious about hunting for dependencies, and to be intelligent enough to be able to find dependencies for and build from a
Generally true. There is, however, a growing degree of consistency in where apps look for those files; in the case of a service it's usually
If that rebuild-from-source allowed you to do something difficult-or-impossible-or-expensive in MS-Windows land, it would be well worth it.
.src.rpm fetch and "rpm -bb" of clamav would be a top investment even if you had to sit and watch it instead of going about your other business when you look at the cost of (for example) Sophos or Norton for 200 seats.
Recent case-in-point, if you want email virus scanning for 200 users, a
Another one, hunting down dependencies for AMaViS is worthwhile (my preferred distro, Mandrake Linux, has them all and it's just "urpmi amavisd" but this customer's site was running Red Hat) if you can discard mail containing semi-broken ZIP files (as sent by mutating MS-Windows viruses) where the commercial packages' attitude seems to universally be "can't read it, therefore it's not really a ZIP, therefore it's safe". Having this crap cut out at the gateway meant that the customer's internal MS-Exchange server could then handle the remaining load. Usually.
If I seriously wanted to upgrade software that old, I'd cross-grade to Mandrake Linux 6.0, then upgrade by x.0's from there. You might get away with upgrading in one jump, but your odds are not good either way.
IRL, I'd copy off the config and any databases, do a fresh install of a recent distro, then feather the config changes back in by hand.
Once you get up to Mandrake 9.0, a distro upgrade is a matter of destroying your old URPMI sources, defining new ones, updating the bits needed for RPM and URPMI, updating everything else, updating your kernel and you're done. I've done this on live (non-critical) servers.
Debian and APT is also that good, only you're not going to have as much luck with that first upgrade step (packaging system is completely different, filesystem structure is significantly different).
An equivalent question is "What would you upgrade Windows 3.1 to?"
PostFix uses databases by default (for efficiency reasons) for some of its config (like mapping email addresses to mailboxes) but these change very rarely (typically months to days between changes) and almost never after hours when backups are typically done. Even so, you could use flat files for the job and it would work fine. Everything else is just a file. Queues, spools, mailboxes etc are all just flat files.
This also means that access control etc is universal and consistent. MS had to write separate access controls for MS-Exchange (have a look at the origins of that one day, you'll be amazed that it works at all), and again for the Registry. Each set is different and complex, each set required work to set up, and work to maintain, and offers a separate opportunity to screw things up. Exchange has never been KISSed. (-:
It's worth adding that if the Linux guy is twice as expensive as the MS-Windows guy, it's because they have less than half as much work to do. Linux admins, in the case of small shops, actually get to leave again, where an MS-Windows admin will almost always be back every week or two.
That's like noticing that Fred seems to be doing fine at shuttling stuff around for the company with his ute, so it's OK to seat him in a semi-trailer (without training), load it up with sixty or seventy tonnes of gear, and send him out the gate.
At the first tight corner, your mistake will become obvious from the sounds of street furniture being destroyed, and when Fred gets to the highway your mistake will be backgrounded with the wail of accumulating sirens as the emergency vehicles arrive to free the survivors and cart away the dead.
Getting computer services right can be a lot more complex than piloting a long, heavy, unresponsive vehicle - although getting PostFix right for a single domain is usually a matter of selecting the right hostname and doing the occasional update.
FOAF living in the US (we live in Oz) kept getting bats infiltrating their garage, and they also keep their meat freezer in the same garage. One day, wife (who loathes bats) asks husband to do something about them, "and while you're out there, can you grab a couple of steaks?" Husband responds, "to drive through their little hearts, dear?" (-:
I agree that MS-Windows zealots will continue to do their thing until some years after Microsoft implodes and running their software ceases to be an option.
Take a typical Mandrake Linux install. You select "PostFix" as one of the packages to install, from the servers section under "Mail and News" - after the installer has asked you for a hostname, and the obvious one is mail.grimaldi.com.ar.
The PostFix service is started by default if it's installed, and it figures out from the hostname that it's doing mail for grimaldi.com.ar. You'd have to turn it off by hand to stop it from working at this point. Morons!
If you go to www.grimaldi.com.ar and look at the source for that placeholder, I think it tells you all you need to know about the competence of Grimaldi's IT section. The page is one centred line of plain text on a plain background, but it is delivered through a frameset (and those are referring to a fixed IP address). Surprisingly, they do actually have a <noframes> clause, and it's more informative than the main page.
...the GP really, really did walk into that one! (-:
Microsoft have never had to fight a competitor like FOSS before, which metaphorically plays its poker with all of their cards face-up on the table and wins anyway. Having half of the cards in play faced up makes it very difficult to cheat in the traditional ways, and the elimination of half of the bluff radically alters the game.
The other terrifying aspect is that in order to play poker at this table, Microsoft needs to face not one or two other players, but countless thousands. Their pin-striped suit and expensive jewelry isn't impressing the other players as much as it ought to, and it's hard to keep track of all of those other hands despite their all being faced up.
There are a few "enemy concentrations" which they can focus on and kind of play against as a collective, like Samba (they've been seriously down about the mouth on Samba for the last year or two), Linux (the FOSS poster child) and OpenOffice ("they wrote what?!"), which between them attack most of the profit avenues represented by the two cash cows, MS-Windows and MS-Office.
However, MS had kind of saturated those two markets, and needs to break into new markets - and guess what? Wherever they turn, FOSS is already there. So Microsoft have turned to peripheral areas, hardware and patents and social/political stuff that's difficult for FOSS to directly compete on.
Hardware is not much of a problem for FOSS, Microsoft has conventional competitors there to keep them in line as long as they can't leverage a Palladium-like monopoly.
Patents are a problem, but I think the whole patent arena is sliding out from under the cover of being merely sublime, and even conservative lawmakers are going to have to take notice of that soon.
This leaves the more nebulous area of social/political manipulation. Traditionally, Microsoft has used its massive presence, advertising and schmoozing budgets to leverage its core markets, together with the implied access to a lot more back doors (and we're not talking BackOrifice here). If they find a more direct way to turn all of that power into a cash-flow, the world is in trouble.
Good old "oops, we forgot to renew the domain - and the other domain" Passport is gradually jamming a foot into that kind of door, in the same way as constantly waving MSIE and MSN in people's faces has saved those items from the total obscurity they deserve. Watch out for Microsoft jamming levers in elsewhere.
Nevertheless, the FOSS community in general refuses to take Microsoft as seriously as MS'd like, and they often react in unpredictable was and predictable ways at the same time, which must cause endless ulcers in Redmond because they can never be sure that the predictable responses from the FOSS crowd aren't masking the launch of unpredictable responses (such as handing out FOSS CDs at Microsoft conventions, a tactic which Microsoft seems to be reciprocating by flooding FOSS_friendly websites with their ads).
</ramble>
..."my goodness, Bill really didn't like that Slate article recommending a switch to FireFox, did he?" (-:
I note that Slate was put up for sale recently, and am quietly wondering if that is a real reaction to editorial honesty.
...with the exception of the "write the serial modem code" argument, which would be reasonable if the code hadn't already been written for Linux (including emulators), none of your arguments hold water.
You can write and run your software on an ordinary Linux desktop, without even the need for a simulator module for the most part. The same Linux which runs your desktop can also run your 'plane, and for a considerably smaller performance and resource hit than XP (although in real life you'd go to the trouble of having a highly modular kernel and not load very much for the instance on the 'plane).
I have no idea why you'd bother with Atmel chips.
Linux runs natively on the Crusoe, another performance gain.
So you used MS-FS for testing algorithms? Then FlightGear would have given you even more control and oversight over what you were testing.
Reading between the lines, you didn't even look.
In this case, fiction is stranger than truth.
...sharks.
The important point is not the changes to the menus, it's that this reflects underlying conceptual and structural differences in the programs. OOo was raised as a specific, planned exception to the trend established by so many other FOSS programs. Kopete, for example, looks nothing much like MSN Messenger, but it does the same job (and also AIM's job, ICQ's job and so on).
You also haven't spoken to the points I raised about KMail.
Use it, love it, and the only concession to Outlook is the toolbar thingy they added with Kontact.
OOo (well, StarOffice) was deliberately designed to look like MS-Office, not for the sake of slavish cloning, but to make it easy for MS-Office users to transition. And it worked. There are still significant structural differences, generally for the better. Forex, in ooWriter you have Format/Character, then Format/Paragraph, then Format/Page, one after another on the menu. In MS-Word, you have Format/Character, Format/Paragraph, and... File/Page-Setup? This also reflects an internal difference in how they view documents. Sections are less foundational to ooWriter than MS-Word, and frames work better in ooWriter (and it produces nicer HTML when used for that).
KOffice is a little differenter, but it also does some things righter, and it's having more tools added to it than OOo. It is definitely not an MS-Office clone, and it is also quite popular.
You also neglected to speak on FireFox and countless other FOSS tools which are unquestionably not clones.
Does this mean I can email the entire song to you as a WAV file? (-:
Pretty sure I have an MP3 or Ogg of it kicking around, something like 37MB even then.
...that sounds like the right company. )-: