I am starting to wonder if there aren't some heavy RedHat investors out there banking on the GPL crumbling under the scrutiny of the courts.
You're forgetting something. If that did happen, Redhat (and Linux in general) would suddenly be playing squarely under Microsoft's rules. We all know what happens to companies that do that. It's not an exaggeration to say that RH's entire business plan depends on the legal viability of the GPL, and indeed their S-1 filing prior to their IPO says as much. I greatly doubt that typical investors, if cognizent of licensing issues at all, have any such thoughts.
"What if they declare Sendmail and GCC to be some RHPL"?
Then those projects will fork from the last GPL version, likely with funding from SuSE/TurboLinux/VA Linux/other distros. This has already happened with SSH, and there's no reason it couldn't for other stuff too.
Redhat still has the best record of any non-Debian distro on open source. *Every* component of a Redhat distro that Redhat develops themselves is GPL. Not any funny weird licence, real live RMS-tested-mother-approved GPL. I'll take a record like that over "but they COULD become evil" nonsense any day.
"Why don't they just stick to providing services"
Because big companies (and I work at one so I know how they think) will feel better about paying Redhat for service contracts if they know RH is associated with at least some key developer(s) of those programs. Let's face it, if you experience a kernel problem and you have a service contract with Redhat, they can have top kernel gurus like Alan Cox and Ingo Molnar take a look at it for you. That's powerful.
Compare this with the Linux service contracts being offered by companies like SCO. Not only do they not have any Linux developers, their main business is a competing product! Given the choice between them and Redhat I know which company's service contract wouldn't make me fear losing my job.
Because MS will for the forseeable future be maintaining Win32-based OSes with 2 radically different architectures but shipping apps that run on both, that limits what "innovations" they can do in user space. NT-only server apps are another story, but AFAIK there's no great desire for people to run IIS on Linux;-)
First off, this article is *totally* irresponsible on Justin's part for not doing any research. The GraphOn product has *NOTHING TO DO* with Wine. They aren't even competitive. Wine lets you run and port Windows apps. GraphOn's thing is a VNC clone that lets you display Windows apps remotely, while running them on real Windows.
And secondly, for the love of Christ, *change the Wine topic icon*. The Wine project has had a real, official logo that's much nicer looking for 2 years now. See http://www.winehq.com/.
PS: about the browser wars: some other developers got MSIE 5 displaying images over the weekend, so now you can do real surfing in it:) Maybe that'll light a fire under the Opera guys;-)
I think their percentage of coverage is very reasonable. There's an increasing amount of Linux stuff, and plenty of cool geek-but-not-necessarily-Linux stuff you'd NEVER see in other mags like the recent LegOS article.
Given that Windows has 90+% market share I think it's fair they mention it at least once an issue;-)
Simply look through the Windows DLLs with a dumper-type tool (PEDUMP by Matt Pietrek works fine). Nearly all Windows system DLLs contain API entry points exported without a name attached, and of these nearly all of them are not documented by Microsoft documentation. And yes, Microsoft apps do call them.
Want more? Download the latest WINE source and check out dlls/comctl32/comctl32undoc.c, or nearly all of dlls/shell32/*.c:) These interfaces are the heart of the IE "integration", and other apps (Mozilla) could "integrate" equally well using them. Third parties could also write credible Explorer desktop replacements (I don't consider LiteStep "credible", sorry;).
I have to agree with you there - I love the Propaganda tiles (don't tell JFK, but I even use them on my M$ machines at work). I just hope the new maintainer doesn't screw things up - I've not seen any of his work yet.
I've been considering doing just that given the VQF plugin that uses WineLib. XMMS and Winamp's plugin interfaces are ridiculously similar - XMMS just renames all the functions so AOL won't sue them:)
Wine has always been 2 projects in one. The actual Windows API reimplementation (WineLib) and a binary loader (Wine, or "the emulator part" as some people refer to it) which loads windows EXE and DLL files and hooks them up with the API stuff in WineLib. The binary loader itself is basically small and bugless -.EXE files are typically a lot less sophisticated than Unix ELF binaries and shared libraries so there's not much to do there in the first place.
So, Corel has put a lot of work into the Windows reimplementation end of things, reorganizing Wine's COM support and header files so you can now compile MFC (which contains a fair amount of MS "insider coding", although luckily you get the source with MSVC++) under WineLib. They also paid Cygnus to do the anonymous struct and union patches which everyone now gets to enjoy in GCC 2.95.2. As a result of this work they now have a common codebase between Windows and Linux for their office applications.
The best part is they've been willing to do all this work to the specs of Wine's current "Linus", Alexandre Julliard. So wine's gotten a ton of good professional work on stuff the regular spare-time developers would have taken much longer on. We now have a solid infrastructure for Win32 threading, impressive OLE/COM support, a much more debugged user interface, and lots of common controls.
As a result it's almost more accurate to say Corel was under Wine's wing - they've had patches rejected due to conflicts with Alexandre's architectural vision, and promptly resubmitted them with everything fixed. I wish all companies involved in open source were that way:)
Anyway, the first stage of that involvement's coming to a close. Corel naturally won't tell us their actual progress with their apps, but they are known to be working on an installer now and their recent patches have been for progressively more obscure bugs. Incidentally, their work has made a lot of other applications work much better too. ModPlug Tracker, a popular Windows tracker-style music application, now works with nary a glitch on Wine. Less than 6 months ago it was unstable and full of graphical glitches.
And as far as claims that app support is getting worse, that's generally false. As with any large project there are frequently bugs that break certain apps, and sometimes apps work accidentally due to combinations of bugs and stop when that bug is fixed. Most people track the "official" releases instead of CVS so something breaks and they don't see that it's fixed again the next day in CVS and end up with the wrong impression:)
Please read documentation/bugreports that comes with the Wine source and post a proper bugreport for your stuff on the newsgroup. We can't help you if you say "it doesn't work".
Read documentation/bugreports that comes with the source. If you report a bug properly it might be able to be fixed. Reporting "it doesn't work" on Slashdot doesn't really help:)
First things first. There's a file called wine/documentation/no-windows that explains in gory detail how to run without windows. If you don't compile wine from source you're missing out on a lot of the documentation.
For the "everyone is using wine", yes that's true. I believe that anyone distributing a binary-only version of wine should be killed. People who can't even operate "./configure ; make depend ; make" should not be running wine in it's current condition.
Wine DOES have a focus. That focus is to get the applications the wine developers use to run. Period. (and that includes the 2 dozen odd Corel guys helping out presently - their focus is the Corel Office apps).
Therefore, if you want an app to run, you have 4 options:
1) Fix it yourself, if you are a programmer 2) Report it to the developers. If one of them has access to the app they might look at it. (most of them are students and can't afford commercial apps though - broken shareware/freeware apps generally get better response). 3) Post a request on CoSource. When all else fails developers do want money:) 4) Shut up and use vmware;-)
Working on wine is annoying precisely because everyone's a freaking expert and nobody actually has any clue (see also the "What I Never Execute" post above).
Your pal, -Ian, wine developer since 1998. (my opinions are not those of the other wine developers, Alexandre, or the Corel guys, so there).
I'm using one right now (to control KDE, take that Bill;-). It's easily the coolest mouse ever - very smooth movement, nothing to clean. Highly recommended for all geeks.
The reason the Macs sold has *always* been because of MacOS (and certain "killer apps" that only exist on it or are radically better than the halfassed Windows ports, like DeBabelizer).
This is why Apple could have been Microsoft had "Star Trek" (the ill-fated x86 port of MacOS) been released prior to Windows 3.0. The Mac hardware has indeed often been spiffy, but it's also often had fiascos (the Mac IIci's unified memory slowing it to a crawl vs. the otherwise similar IIcx, f'rinstance). And their major hardware innovations have generally been asthetic rather than technical (the iMac, the curvy-tower PowerMacs, the G3/G4 slide-open side panel). Even their apparent technical achievements have frequently actually been cribbed (the Quadra AVs? 030/040 plus a DSP? Sounds a lot like NeXTStations.)
As far as the clone thing, yes all their revenues come from selling hardware. But again, why does that hardware sell? Becuase it's the only place you can get MacOS. Mac hardware would be a nonfactor if MacOS were on more platforms.
-Ian, another occasionally bitter former IIgs user (listen to IIgs music on Win32, Linux, and FreeBSD with MTP/MTPlug at http://home.twcf.rr.com/ischmidt/warez.html)
...but images aren't loading yet for reasons we're still working on. Still, basic page layout is intact and yes, it still blows Crapscape 4.7 out of the water on page loading speed.
They are in our experience at work the #1 reason gcc builds slower than MSVC++ (both compilers running native on Win98, so OS speed advantages don't exist:)
I've heard they may be in GCC 3.0 - does anyone know for sure?
RH pays Alan (and Stephen Tweedie, and Ingo Molnar, and other Big Names from linux-kernel), so I would consider that (plus their GNOME work) to be "giving money back to developers".
I installed IE 5 on my Windows partition when it came out ("early adopter", and all that;) and promptly uninstalled it after about 3 days. Yes, it displays pages a little faster than IE 4 or Netscape, but it takes up gigantic amounts of RAM and has a ridiculous number of bugs. I like IE 4.01 (with the appropriate security patches) a lot better.
Also, IE 4 comes stock on Win98, which is what most consumer (read new Internet user) PCs have. These people aren't very prone to upgrading a browser, particularly when it involves a 30 meg download over a liable-to-blow-at-any-time WinModem.
And of course all those iMacs (same sorta users, better modem;) come with IE 4.5 standard - I don't know if the original poster differentiated Mac IE 4 from Win, but that's another datapoint.
I have some Cambridge Soundworks speakers on my PC that aren't very big at all (except the subwoofer "brick") and they sound *terrific*. I have no doubt that Harmon/Kardon could do the same with the iMac.
CTRL-ALT-DEL in Windows normally a) gives you a desktop back (if you're in a fullscreen game or something) and b) pops up the Task Manager, where you can kill errant programs. Of course, there's also option c) BSoD, which happens waaay too often.
Slashdot Mirror
I am starting to wonder if there aren't some heavy RedHat investors out there banking on the GPL crumbling under the scrutiny of the courts.
You're forgetting something. If that did happen, Redhat (and Linux in general) would suddenly be playing squarely under Microsoft's rules. We all know what happens to companies that do that. It's not an exaggeration to say that RH's entire business plan depends on the legal viability of the GPL, and indeed their S-1 filing prior to their IPO says as much. I greatly doubt that typical investors, if cognizent of licensing issues at all, have any such thoughts.
"What if they declare Sendmail and GCC to be some RHPL"?
Then those projects will fork from the last GPL version, likely with funding from SuSE/TurboLinux/VA Linux/other distros. This has already happened with SSH, and there's no reason it couldn't for other stuff too.
Redhat still has the best record of any non-Debian distro on open source. *Every* component of a Redhat distro that Redhat develops themselves is GPL. Not any funny weird licence, real live RMS-tested-mother-approved GPL. I'll take a record like that over "but they COULD become evil" nonsense any day.
"Why don't they just stick to providing services"
Because big companies (and I work at one so I know how they think) will feel better about paying Redhat for service contracts if they know RH is associated with at least some key developer(s) of those programs. Let's face it, if you experience a kernel problem and you have a service contract with Redhat, they can have top kernel gurus like Alan Cox and Ingo Molnar take a look at it for you. That's powerful.
Compare this with the Linux service contracts being offered by companies like SCO. Not only do they not have any Linux developers, their main business is a competing product! Given the choice between them and Redhat I know which company's service contract wouldn't make me fear losing my job.
Corel showed 6 of their apps compiled with WineLib at COMDEX: WordPerfect, Quattro Pro, Presentations,
Paradox, CorelDRAW, and Photo-Paint.
There are some screenshots of some of the apps here:
http://newmedia.corel.com/webcast/gallery.htm
Because MS will for the forseeable future be maintaining Win32-based OSes with 2 radically different architectures but shipping apps that run on both, that limits what "innovations" they can do in user space. NT-only server apps are another story, but AFAIK there's no great desire for people to run IIS on Linux ;-)
First off, this article is *totally* irresponsible on Justin's part for not doing any research. The GraphOn product has *NOTHING TO DO* with Wine. They aren't even competitive. Wine lets you run and port Windows apps. GraphOn's thing is a VNC clone that lets you display Windows apps remotely, while running them on real Windows.
:) Maybe that'll light a fire under the Opera guys ;-)
And secondly, for the love of Christ, *change the Wine topic icon*. The Wine project has had a real, official logo that's much nicer looking for 2 years now. See http://www.winehq.com/.
PS: about the browser wars: some other developers got MSIE 5 displaying images over the weekend, so now you can do real surfing in it
-Ian, in the Wine AUTHORS file and damn proud.
I think their percentage of coverage is very reasonable. There's an increasing amount of Linux stuff, and plenty of cool geek-but-not-necessarily-Linux stuff you'd NEVER see in other mags like the recent LegOS article.
;-)
Given that Windows has 90+% market share I think it's fair they mention it at least once an issue
Simply look through the Windows DLLs with a dumper-type tool (PEDUMP by Matt Pietrek works fine). Nearly all Windows system DLLs contain API entry points exported without a name attached, and of these nearly all of them are not documented by Microsoft documentation. And yes, Microsoft apps do call them.
:) These interfaces are the heart of the IE "integration", and other apps (Mozilla) could "integrate" equally well using them. Third parties could also write credible Explorer desktop replacements (I don't consider LiteStep "credible", sorry ;).
Want more? Download the latest WINE source and check out dlls/comctl32/comctl32undoc.c, or nearly all of dlls/shell32/*.c
Me too :) Gotta be ready to meet the Techno Talking Babes(tm) :-)
He's been posting about his "college tour", and Leader Kibo's been all over him as usual :)
I have to agree with you there - I love the Propaganda tiles (don't tell JFK, but I even use them on my M$ machines at work). I just hope the new maintainer doesn't screw things up - I've not seen any of his work yet.
I've been considering doing just that given the VQF plugin that uses WineLib. XMMS and Winamp's plugin interfaces are ridiculously similar - XMMS just renames all the functions so AOL won't sue them :)
Wine has always been 2 projects in one. The actual Windows API reimplementation (WineLib) and a binary loader (Wine, or "the emulator part" as some people refer to it) which loads windows EXE and DLL files and hooks them up with the API stuff in WineLib. The binary loader itself is basically small and bugless - .EXE files are typically a lot less sophisticated than Unix ELF binaries and shared libraries so there's not much to do there in the first place.
:)
:)
So, Corel has put a lot of work into the Windows reimplementation end of things, reorganizing Wine's COM support and header files so you can now compile MFC (which contains a fair amount of MS "insider coding", although luckily you get the source with MSVC++) under WineLib. They also paid Cygnus to do the anonymous struct and union patches which everyone now gets to enjoy in GCC 2.95.2. As a result of this work they now have a common codebase between Windows and Linux for their office applications.
The best part is they've been willing to do all this work to the specs of Wine's current "Linus", Alexandre Julliard. So wine's gotten a ton of good professional work on stuff the regular spare-time developers would have taken much longer on. We now have a solid infrastructure for Win32 threading, impressive OLE/COM support, a much more debugged user interface, and lots of common controls.
As a result it's almost more accurate to say Corel was under Wine's wing - they've had patches rejected due to conflicts with Alexandre's architectural vision, and promptly resubmitted them with everything fixed. I wish all companies involved in open source were that way
Anyway, the first stage of that involvement's coming to a close. Corel naturally won't tell us their actual progress with their apps, but they are known to be working on an installer now and their recent patches have been for progressively more obscure bugs. Incidentally, their work has made a lot of other applications work much better too. ModPlug Tracker, a popular Windows tracker-style music application, now works with nary a glitch on Wine. Less than 6 months ago it was unstable and full of graphical glitches.
And as far as claims that app support is getting worse, that's generally false. As with any large project there are frequently bugs that break certain apps, and sometimes apps work accidentally due to combinations of bugs and stop when that bug is fixed. Most people track the "official" releases instead of CVS so something breaks and they don't see that it's fixed again the next day in CVS and end up with the wrong impression
-Ian, wine-devel but not speaking for 'em.
Please read documentation/bugreports that comes with the Wine source and post a proper bugreport for your stuff on the newsgroup. We can't help you if you say "it doesn't work".
Read documentation/bugreports that comes with the source. If you report a bug properly it might be able to be fixed. Reporting "it doesn't work" on Slashdot doesn't really help :)
First things first. There's a file called wine/documentation/no-windows that explains in gory detail how to run without windows. If you don't compile wine from source you're missing out on a lot of the documentation.
:) ;-)
For the "everyone is using wine", yes that's true. I believe that anyone distributing a binary-only version of wine should be killed. People who can't even operate "./configure ; make depend ; make" should not be running wine in it's current condition.
Wine DOES have a focus. That focus is to get the applications the wine developers use to run. Period. (and that includes the 2 dozen odd Corel guys helping out presently - their focus is the Corel Office apps).
Therefore, if you want an app to run, you have 4 options:
1) Fix it yourself, if you are a programmer
2) Report it to the developers. If one of them has access to the app they might look at it. (most of them are students and can't afford commercial apps though - broken shareware/freeware apps generally get better response).
3) Post a request on CoSource. When all else fails developers do want money
4) Shut up and use vmware
Working on wine is annoying precisely because everyone's a freaking expert and nobody actually has any clue (see also the "What I Never Execute" post above).
Your pal,
-Ian, wine developer since 1998.
(my opinions are not those of the other wine developers, Alexandre, or the Corel guys, so there).
Saturation over 40? Urgh - how can you see your icons? :)
I'm using one right now (to control KDE, take that Bill ;-). It's easily the coolest mouse ever - very smooth movement, nothing to clean. Highly recommended for all geeks.
The reason the Macs sold has *always* been because of MacOS (and certain "killer apps" that only exist on it or are radically better than the halfassed Windows ports, like DeBabelizer).
This is why Apple could have been Microsoft had "Star Trek" (the ill-fated x86 port of MacOS) been released prior to Windows 3.0. The Mac hardware has indeed often been spiffy, but it's also often had fiascos (the Mac IIci's unified memory slowing it to a crawl vs. the otherwise similar IIcx, f'rinstance). And their major hardware innovations have generally been asthetic rather than technical (the iMac, the curvy-tower PowerMacs, the G3/G4 slide-open side panel). Even their apparent technical achievements have frequently actually been cribbed (the Quadra AVs? 030/040 plus a DSP? Sounds a lot like NeXTStations.)
As far as the clone thing, yes all their revenues come from selling hardware. But again, why does that hardware sell? Becuase it's the only place you can get MacOS. Mac hardware would be a nonfactor if MacOS were on more platforms.
-Ian, another occasionally bitter former IIgs user
(listen to IIgs music on Win32, Linux, and FreeBSD with MTP/MTPlug at http://home.twcf.rr.com/ischmidt/warez.html)
...but images aren't loading yet for reasons we're still working on. Still, basic page layout is intact and yes, it still blows Crapscape 4.7 out of the water on page loading speed.
Or would it be McNealy or Ellison? :)
They are in our experience at work the #1 reason gcc builds slower than MSVC++ (both compilers running native on Win98, so OS speed advantages don't exist :)
I've heard they may be in GCC 3.0 - does anyone know for sure?
RH pays Alan (and Stephen Tweedie, and Ingo Molnar, and other Big Names from linux-kernel), so I would consider that (plus their GNOME work) to be "giving money back to developers".
I installed IE 5 on my Windows partition when it came out ("early adopter", and all that ;) and promptly uninstalled it after about 3 days. Yes, it displays pages a little faster than IE 4 or Netscape, but it takes up gigantic amounts of RAM and has a ridiculous number of bugs. I like IE 4.01 (with the appropriate security patches) a lot better.
;) come with IE 4.5 standard - I don't know if the original poster differentiated Mac IE 4 from Win, but that's another datapoint.
Also, IE 4 comes stock on Win98, which is what most consumer (read new Internet user) PCs have. These people aren't very prone to upgrading a browser, particularly when it involves a 30 meg download over a liable-to-blow-at-any-time WinModem.
And of course all those iMacs (same sorta users, better modem
I have some Cambridge Soundworks speakers on my PC that aren't very big at all (except the subwoofer "brick") and they sound *terrific*. I have no doubt that Harmon/Kardon could do the same with the iMac.
CTRL-ALT-DEL in Windows normally a) gives you a desktop back (if you're in a fullscreen game or something) and b) pops up the Task Manager, where you can kill errant programs. Of course, there's also option c) BSoD, which happens waaay too often.