It was, but there were communications issues. So now they're going to not only try to implement the entirely of S.W.F in C# (that's an entire widget toolkit!) but also a fake WndProc/Handle property for apps that rely on the underlying Win32ness of it.
For *really* hard core apps (read: most.NET apps) that actually mix native and.NET code, they were talking about having a special "winelib" mode where it's somehow used, but I've not seen any credible technical details on this.
Basically it looks like in order to run.NET code that is mixed with native Win32 code the Wine team will have to do their own S.W.F implementation one day. Currently there are no killer.NET apps that I know of (by killer I mean public/retail software sufficiently unique that it's a "blocks my migration to Linux" app, rather than custom corp apps) so it's not a big deal. But I expect they'll come eventually.
Well, clearly Mozilla is not an alternative as Mozilla is not fully compatible with many web pages out there.
This is a common problem, especially in corporate scenarios. Mozilla/Firefox are great but for many users who browse non-geeky sites (my brother is a musician and has this problem) there are still way too many sites with IE only code in them.
So I can't say I blame them for wanting to run IE.
Anyway, I'd object to a Mac on other grounds: the whole reason the IT industry is in the mess it is, is because a proprietary platform took hold and then took over. The logical conclusion to pushing Macs is for Apple to be the next Microsoft, which would be a very silly thing indeed.
Yes, this concerns me a bit - I really, really doubt they will be able to just magically "port" all their Windows software. For starters, I'd guess a lot of it isn't owned by them. So they have to pressure the vendors. That may work when the vendor is a little family outfit in Munich itself but not when the vendor is a bigass consulting company who can live without the city of Munich bugging them for a ton of ports.
Even when they're little vendors in Germany, I seem to recall there being severe problems with most of them flat out refusing to do ports. Why? They didn't know how. They were your typical Windows VC++/Delphi developers, had probably never even used Linux let alone written software for it, and so on. Doing a Linux port means serious retraining, a lot of time etc, which would probably have been economically unjustifiable unless they charged unfeasable amounts.
So basically I think if they press on with using VMware (which ain't cheap!) it'll end up being a PR disaster for the wider Linux community. Microsoft and their allies will just point and say "Look, they went Linux for political reasons and ended up with Windows in a box on 80% of their desktops anyway: what's the point? Might as well stick with us". Shills like Thurout have pretty much already said that, in fact. It'll only get worse.
The right way to do migrations, which is fortunately what some other big organizations are doing, is to improve Wine so it runs their programs acceptably. Not only have you done the migration using Free software, you sidestepped problems with apps that can't/won't be ported, plus you also laid the groundwork for these companies to port their apps in future - it's a lot easier to port an app piece by piece using Winelib than it is to junk and rewrite the entire thing (which is often what a port implies).
Um, the guy said that "IE was a minimum", that's IE for Windows not for MacOS - they are totally different and IE for Mac can't render all the pages IE for Windows can.
Let's assume the poster is serious about their requirements. If you're just going to ignore hard customer specifications like "must run IE" then feel free to post but it will be ignored by the people actually doing the work (ie, them).
If you read the whole post it seems the only two systems that'd do what (s)he needs are:
1) Windows Terminal Services
2) Some form of Linux/LTSP with a copy of Wine/CrossOver Office Server edition to run IE and MS Office. CodeWeavers sell a version of Wine specifically kitted out to run thin clients, because it's massively more efficient than Citrix/RDP-type solutions.
It rules out anything based on "pure" open source because of the IE/MSOffice requirement, probably anything Linux based is ruled out on the grounds of staff availability and familiarity, and I'm afraid Mac is definitely ruled out. Sorry.
I read that page, it sounds exactly like Microsofts implementation to me (if you ignore the bizarro Qt emulation layer).
iexplore.exe is a tiny program which doesn't do much, basically just activating some COM components. Even the IE main loop is in a supporting DLL. So you can delete that just fine, (obviously) as long as you set your default browser to something else. The IE core will still be there, but then the same is true of Safari. I don't see what would break?
By that definition then you can delete c:\program files\internet explorer with no ill consequences: forget the fact that 99% of IE is in DLLs in c:\windows\system - look ma, the icon has gone!
No, when a web browser has OS level components, to delete that browser entirely you have to delete those components as well. WebCore is written by Apple, controlled by Apple, and shipped as part of Safari. It's a part of Safari as much as Trident/MLANG/URLMonikers is a part of IE.
With that attitude you end up with the EULA mess you've got on Windows where people actually copy and paste each others EULAs because they feel they should have one.
No no, if we want this sort of thing to stop, better to nip it in the bud before it becomes a culture.
1) They have a fully functional GUI on top of an open source OS
A kernel is not an OS - not just the MacOS X GUI is proprietary but also the API implementations, which vastly outweigh the size of Darwin. Litmus test: can you run MacOS X GUI apps (and nearly all are) on a plain old Darwin build? No? Thought not.
2) their open source OS is still building on BOTH Power PC and Intel platforms.
So what? "Building" does not equal "installs easily", "has drivers", "optimized for" etc etc
3) a version of Microsoft Office (like it or not, this is a huge advantage that the Mac has over other Open Source OS's)
Not really. Office for Mac is good, but it's not the software that people have already paid for, trained on, scripted and written arcane spreadsheet/OLE hybrid apps for.
Why rebuy all your copies of Office at huge expensive (Office for Mac costs more than MS Office does) when you can run the real thing on an open source OS via Windows emulation?
I can completely delete safari from my computer. ..can you say the same about IE? Of course not, your entire system won't work if you do.
No you can't, not unless you also delete WebCore at which point the help center will break (along with a few other things, I suspect).
That said, having a web renderer (which is basically what IE is) integrated with an OS is not necessarily a bad thing.
The rest of your points are fairly vague : wow, both Microsoft and Apple issue OS updates. Big deal. I'd say that SP2 definitely has as much new functionality as an Apple update, even if it focusses on security rather than "lifestyle changes": things like the new security center are not just "settings changes" or "bugfixes".
Anyway, it's not like Apple can throw rocks: it's a glass house. How long did that URL handler exploit go unpatched again?
Seems like they've done a good job, judging from the screenshots they have nice, straightforward UI for a lot of things. In particular, their DEP UI (I'm assuming DEP is a direct equivalent to execshield on Linux) impressed me: when execshield came out there was no such control system for it. Now, the only program that I'm aware of which it broke is Wine which is a special case, but I would not be surprised if there were other programs out there which it would have broken too. It's this sort of "GUI for everything" approach the Linux community still needs to catch up on.
The new security center looked nice too, I can't imagine many people misunderstanding it.
By definition for 95%+ of users if supporting multiple architectures means something as fundamental as a graphical installer is lacking, then extreme multi-arch support is not a good thing.
Yes yes, very amusing:) Obviously Mac IE is not actually "Internet Explorer", it's a different program that happens to have the same name. Much like "Office for Mac" is not what people mean when they say "Microsoft Office".
Suffice it to say that a lot of sites don't work with IE for the Mac. This is especially true of internal corporate sites.
As far as I'm aware the only OS that can run the real IE6 that isn't Windows, is Linux.
No, they don't, in fact their own development teams have a habit of writing utility libraries and wrapper DLLs to abstract themselves from the lower levels of Windows. Unfortunately those DLLs themselves tend to morph into semi-public API: the SHLWAPI DLL is one example of this. The "Shell Lightweight API" was written by the IE team when they had to run on lots of different versions of Windows, and is sorta half documented these days. Layer MFC + ATL + SHLWAPI on top of all the other random DLLs Windows has, and you get something barely comprehensible.
Don't even get me started on DCOM, which is involved in services as basic as the clipboard (that's why WinXP reboots if the RPCSS service dies). MSDN can't even describe some of the DCOM interfaces without making mistakes - they actually miss out methods and stuff. Only God (or Don Box) fully understands how it works, but it does...
Darwin doesn't include 99% of the APIs that Mac apps use, ie Carbon and Cocoa. Because of that it's merely a slow implementation of UNIX, not something that is especially interesting for running Mac apps.
I don't know but suspend works great on my HP OmniBook so Linux (2.6 at any rate) definitely supports it nicely. I think for PowerBooks some fiddling is still required on a few distros, maybe YDL has this nailed.
Suffice it to say that suspend is really cool. I don't think I've switched my laptop off for ages, I just press the half-moon button and within a few seconds it's "off". A few more seconds and it's back on. The only gripe I have with it is that when I resume the screensaver has come on, as xscreensaver is time based. That should be easily fixed though.
Well, Linux is free software (as in GNU free) which is important to some. Desktop Linux has some things that MacOS X lacks or can only gain with extra hacks. For instance, I'm very comfortable with virtual desktops in Gnome 2.6 and miss them terribly when I'm on a Windows or Mac machine. There are other reasons too - the shareware culture is still prevalent in the Mac world: you can often find yourself being asked to pay ~$30 for a little utility that would be free under Linux, if needed at all.
Other reasons include apps like Evolution fitting in on Linux, but not on MacOS (and Evo 2 is really, really sweet, far better than Apple Mail in my estimation), improved semantic/source compatibility with the x86 Linux world, want to learn it for future job markets etc etc.
Given the sheer size of the API, the breadth of its features and the fact that nearly all "real".NET apps I've seen P/Invoke and use handles heavily, I think that's a severe over-estimate. I'd think it was more realistic to say we'd have a "good enough to run complex apps" with a bit of hacking in a few years.
Dropping Wine wasn't a smart idea either IMHO, the most common kind of.NET app on Windows is not likely to be a pure.NET app but rather an extended version of code that already exists. Certainly, native code will be reused. So I don't think there's any way to make a useful duplicate of the Microsoft stack without Windows emulation being involved at some point.
GTK themes can have their colours changed as the BlueCurve sets in FC2 show, however their is no GUI for setting whatever colours you like. There could be, it's just that nobody wrote one yet. Why don't you do it?
You realise that most of the changes in 2.8 are about fixing bugs and polishing, right?
The new MIME system is "fixing" the old one by totally replacing it, no other approach would work. The new system by the way is a lot easier to use for both users and developers, and is a freedesktop standard shared with KDE:)
The rest of the desktop is not receiving any major new features really, just lots of bugfixing too small to go in these sort of "what's new" pages and various cleanups. Actually Gnome seems to have slowed down in this release as a lot of the Red Hat and Novell hackers are tied up with non-Gnome work as they round out the rest of the Linux desktop (so, hardware integration, management tools, backwards compatibility work etc).
Slashdot Mirror
For *really* hard core apps (read: most .NET apps) that actually mix native and .NET code, they were talking about having a special "winelib" mode where it's somehow used, but I've not seen any credible technical details on this.
Basically it looks like in order to run .NET code that is mixed with native Win32 code the Wine team will have to do their own S.W.F implementation one day. Currently there are no killer .NET apps that I know of (by killer I mean public/retail software sufficiently unique that it's a "blocks my migration to Linux" app, rather than custom corp apps) so it's not a big deal. But I expect they'll come eventually.
I think the number of "bedroom inventors" these days is sufficiently small that this is an acceptable loss for reigning in rampant patent abuse.
This is a common problem, especially in corporate scenarios. Mozilla/Firefox are great but for many users who browse non-geeky sites (my brother is a musician and has this problem) there are still way too many sites with IE only code in them.
So I can't say I blame them for wanting to run IE.
Anyway, I'd object to a Mac on other grounds: the whole reason the IT industry is in the mess it is, is because a proprietary platform took hold and then took over. The logical conclusion to pushing Macs is for Apple to be the next Microsoft, which would be a very silly thing indeed.
Even when they're little vendors in Germany, I seem to recall there being severe problems with most of them flat out refusing to do ports. Why? They didn't know how. They were your typical Windows VC++/Delphi developers, had probably never even used Linux let alone written software for it, and so on. Doing a Linux port means serious retraining, a lot of time etc, which would probably have been economically unjustifiable unless they charged unfeasable amounts.
So basically I think if they press on with using VMware (which ain't cheap!) it'll end up being a PR disaster for the wider Linux community. Microsoft and their allies will just point and say "Look, they went Linux for political reasons and ended up with Windows in a box on 80% of their desktops anyway: what's the point? Might as well stick with us". Shills like Thurout have pretty much already said that, in fact. It'll only get worse.
The right way to do migrations, which is fortunately what some other big organizations are doing, is to improve Wine so it runs their programs acceptably. Not only have you done the migration using Free software, you sidestepped problems with apps that can't/won't be ported, plus you also laid the groundwork for these companies to port their apps in future - it's a lot easier to port an app piece by piece using Winelib than it is to junk and rewrite the entire thing (which is often what a port implies).
Let's assume the poster is serious about their requirements. If you're just going to ignore hard customer specifications like "must run IE" then feel free to post but it will be ignored by the people actually doing the work (ie, them).
If you read the whole post it seems the only two systems that'd do what (s)he needs are:
1) Windows Terminal Services
2) Some form of Linux/LTSP with a copy of Wine/CrossOver Office Server edition to run IE and MS Office. CodeWeavers sell a version of Wine specifically kitted out to run thin clients, because it's massively more efficient than Citrix/RDP-type solutions.
It rules out anything based on "pure" open source because of the IE/MSOffice requirement, probably anything Linux based is ruled out on the grounds of staff availability and familiarity, and I'm afraid Mac is definitely ruled out. Sorry.
iexplore.exe is a tiny program which doesn't do much, basically just activating some COM components. Even the IE main loop is in a supporting DLL. So you can delete that just fine, (obviously) as long as you set your default browser to something else. The IE core will still be there, but then the same is true of Safari. I don't see what would break?
No, when a web browser has OS level components, to delete that browser entirely you have to delete those components as well. WebCore is written by Apple, controlled by Apple, and shipped as part of Safari. It's a part of Safari as much as Trident/MLANG/URLMonikers is a part of IE.
No no, if we want this sort of thing to stop, better to nip it in the bud before it becomes a culture.
A kernel is not an OS - not just the MacOS X GUI is proprietary but also the API implementations, which vastly outweigh the size of Darwin. Litmus test: can you run MacOS X GUI apps (and nearly all are) on a plain old Darwin build? No? Thought not.
2) their open source OS is still building on BOTH Power PC and Intel platforms.
So what? "Building" does not equal "installs easily", "has drivers", "optimized for" etc etc
3) a version of Microsoft Office (like it or not, this is a huge advantage that the Mac has over other Open Source OS's)
Not really. Office for Mac is good, but it's not the software that people have already paid for, trained on, scripted and written arcane spreadsheet/OLE hybrid apps for.
Why rebuy all your copies of Office at huge expensive (Office for Mac costs more than MS Office does) when you can run the real thing on an open source OS via Windows emulation?
No you can't, not unless you also delete WebCore at which point the help center will break (along with a few other things, I suspect).
That said, having a web renderer (which is basically what IE is) integrated with an OS is not necessarily a bad thing.
The rest of your points are fairly vague : wow, both Microsoft and Apple issue OS updates. Big deal. I'd say that SP2 definitely has as much new functionality as an Apple update, even if it focusses on security rather than "lifestyle changes": things like the new security center are not just "settings changes" or "bugfixes".
Anyway, it's not like Apple can throw rocks: it's a glass house. How long did that URL handler exploit go unpatched again?
What a surprise .... too bad GUIs for this sort of thing are not prioritised.
The new security center looked nice too, I can't imagine many people misunderstanding it.
By definition for 95%+ of users if supporting multiple architectures means something as fundamental as a graphical installer is lacking, then extreme multi-arch support is not a good thing.
Suffice it to say that a lot of sites don't work with IE for the Mac. This is especially true of internal corporate sites.
As far as I'm aware the only OS that can run the real IE6 that isn't Windows, is Linux.
Amusingly that's the first hit on Google for internet explorer installer. Microsoft Downloads comes second.
You mean like this?
Don't even get me started on DCOM, which is involved in services as basic as the clipboard (that's why WinXP reboots if the RPCSS service dies). MSDN can't even describe some of the DCOM interfaces without making mistakes - they actually miss out methods and stuff. Only God (or Don Box) fully understands how it works, but it does ...
Darwin doesn't include 99% of the APIs that Mac apps use, ie Carbon and Cocoa. Because of that it's merely a slow implementation of UNIX, not something that is especially interesting for running Mac apps.
Suffice it to say that suspend is really cool. I don't think I've switched my laptop off for ages, I just press the half-moon button and within a few seconds it's "off". A few more seconds and it's back on. The only gripe I have with it is that when I resume the screensaver has come on, as xscreensaver is time based. That should be easily fixed though.
Thanks but the only time I use a Mac is when I'm round at a friends who has one. I'm one of those "open platforms are important" guys I'm afraid ;)
Other reasons include apps like Evolution fitting in on Linux, but not on MacOS (and Evo 2 is really, really sweet, far better than Apple Mail in my estimation), improved semantic/source compatibility with the x86 Linux world, want to learn it for future job markets etc etc.
There are lots of reasons.
Yes you could reimplement libpng in a safe language that allowed for C export like D or maybe (with some hackery) Java. Nobody has though.
Given the sheer size of the API, the breadth of its features and the fact that nearly all "real" .NET apps I've seen P/Invoke and use handles heavily, I think that's a severe over-estimate. I'd think it was more realistic to say we'd have a "good enough to run complex apps" with a bit of hacking in a few years.
Dropping Wine wasn't a smart idea either IMHO, the most common kind of .NET app on Windows is not likely to be a pure .NET app but rather an extended version of code that already exists. Certainly, native code will be reused. So I don't think there's any way to make a useful duplicate of the Microsoft stack without Windows emulation being involved at some point.
It's also currently totally undiscoverable. At the very least it should appear under "Computer" or something.
GTK themes can have their colours changed as the BlueCurve sets in FC2 show, however their is no GUI for setting whatever colours you like. There could be, it's just that nobody wrote one yet. Why don't you do it?
The new MIME system is "fixing" the old one by totally replacing it, no other approach would work. The new system by the way is a lot easier to use for both users and developers, and is a freedesktop standard shared with KDE :)
The rest of the desktop is not receiving any major new features really, just lots of bugfixing too small to go in these sort of "what's new" pages and various cleanups. Actually Gnome seems to have slowed down in this release as a lot of the Red Hat and Novell hackers are tied up with non-Gnome work as they round out the rest of the Linux desktop (so, hardware integration, management tools, backwards compatibility work etc).