Plugins should be fine, but XPCOM components might pose problems. Firefox is statically linked to a lot of things it uses whereas Mozilla sticks them in separate DLLs.
The latter approach is more suitable for a modular suite of apps (and for development & debugging) , but it means that DLLs built for one may or may not work for the other if they pull in external symbols that are not there in the static build.
It should pose a problem unless you're doing something hardcore with Gecko but it is something to be aware of.
It's only a bargain if your bandwidth is free, if the picture quality is watchable, if you can cease the service at any time, if you can watch a movie at a time of your choosing (from a cache), if you happen to use MS Windows, if the movies are recent and if the movies are actually worth watching.
If the answer to any of these things is "not bloody likely", it seems a rather pointless thing to me. Why subscribe to kill your bandwidth watching crappy movies in crappy quality?
Hover over a buddy in AIM and you'll see what capabilities the other person has available to them, e.g. chat, video etc. I assume that AIM would be smart enough to grey out options that were not applicable for the person you had selected.
LOL - I will correct post right away - I got the press releases all mixed up. The SCO press release says 4000 stores (anywhere), the Novell one says 1200 stores (in Germany) but talks of running SUSE Enterprise Linux. So my points are based on a total misreading of what was in front of my eyes. My bad.
Anyway sooner or later anything running SCO is doomed.
Well it does says 4000 stores - I assume that most stores have at most 2 server PCs in a back office, maybe even only one holding the accounts, talking to the terminals and talking to head office.
And the really ancient thing they're using here is SCO Unix which as far back as I can remember has never been 'modern' at any point in time. But as a *nix, porting from it to Linux should be a relatively straightforward proposition (as Autozone already know much to SCO's consternation and bogus lawsuits). And it may be that the vendors who wrote the databases, stock control / POS / accounting / payrole software have already done the work since Linux is a tier 1 platform.
Who knows, perhaps it was the vendors who motivated McDonalds to quit. Maybe they said to management what everyone already else knows - SCO is a piece of shit, the company is going down the plug hole, it won't run on hardware X and it'll cost $$$$ to support it even if it did. No doubt McD's IT department pushed the same story from the inside.
It's another kick in the balls for SCO which is always good for a laugh. It's too bad the markets are closed today because their stock is doing a good impression of the Titanic.
But to be fair, the stock for Novell tanked something like 20% in the last fortnight. I know because I'm a shareholder. And I'm a shareholder in IBM though they are somewhat more resilient to market shakes and are actually up on what I paid for them.
Anyway, don't take this as advice (since I'm a newbie investor & obviously not an advisor of any sort) but the reason I say I am still a shareholder (not was - am) is because I have confidence in NOVL & IBM. I could have dumped the shares at a loss, but I believe that both have a business and products to emerge on the other side through the legal tussles whereas SCO won't. The only plan SCO seems to have is sue, fud, sue, threaten, bluster and whine. That would be fine if there were a genuine case, but nothing from their own lips or otherwise says there is.
It would have been great with hindsight to have ridden the coattails of the SCO FUD and make a small fortune trading against the hype - the highs and the lows of SCO are enviable with hindsight. But by this point the future of the company is already carved in stone - RIP. Anyone who hangs onto their shares at this stage is either extraordinarily bullish or fantastically stupid.
Not every window manager arranges your files and apps like a desktop, putting things where you left them, not giving you nasty surprises, working in a way that makes sense to anyone who doesn't give a damn about window managers. It's taken GNOME / KDE long enough to get to that point, and they're all the better for it.
While I appreciate that GNOME / KDE aren't the lightest WMs, they're about the only ones that are proper desktops as far as mere mortals are concerned.
I'm sure you could get similar functionality by cobbling a WM, a terminal app, some kind of file browser, etc. I've put up with that kind of crap on Unix for 15 years when lesser machines such as the Mac, Amiga, Atari ST etc. had it way back then. I'm glad that the mainstream has finally left that kind of mentality behind. It doesn't stop anyone using GNUStep or E or CDE or whatever, but unless you are seriously strapped for memory, or the box lives most of its life unattended there is little point.
Personally I just enjoy having a proper desktop because I despise screwing around in some config file to add a lousy icon or to change the screen resolution, or having to run mix and match apps to be able to browse files, networks, printers etc. when they are all inconsistent with each other and the WM. Give me GNOME any day.
I post the odd talkback on LT. Aside from the lengthy delay between posting and approval, I have no issues with the website.
If Microsoft want to waste their money preaching to the unconvertible then let them. Whether it is Slashdot or Linux Today they might as well be pissing their money into the wind for all the good it will do them. At least this way, a Linux oriented enterprise benefits from Microsoft's desperation.
Plus, if some customer on a Segway is brutally assaulted / smashes into a pedestrian / plows into a post you can look forward to an expensive lawsuit in addition to the loss of the segway he / she was riding on at the time.
Nautlius works just fine in 2.6. The spatial functionality combined with fast file typing (by using a file extension) means that Linux works and feels like a proper desktop operating system for once. Previous versions were okay, but much too slow.
Fortunately it is still possible to run an explorer style window too, but overall GNOME 2.6 gets a big thumbs up from me - spatial, minimalist, uncluttered.
While it would be nice to see tools such as blogging etc. but I'd rather see services to support such tools. For example, Evolution has a nice day view that gives weather, rss news and other stuff. Something like that would be very useful for the desktop, where the content is supplied through an API that any app can plug into.
Perhaps but then again you *need* it. I don't think twice about sticking a Fedora disk in and installing it. With Gentoo I'd be afraid to even start installing unless I had another computer open beside the target machine so I could tentatively proceed one step at a time.
I wonder if Gentoo actually needs to be so hard to install. After all, it could probably boot into a graphical or menu driven installer just like other dists, but with the difference that it additionally builds some packages instead of copying the binaries.
Too contrary to type two words and get the results that you demand me to manually type out instead? Red Hat, Suse, Mandrake provide their own lists are you'd discover on the first 20 results, as well as several independent lists, as well as lists dedicated to particular areas such as audio, firewire, modems etc.
In summary, if you had a point (and it's hard to see that you did) you've lost it.
Whatever other qualities Solaris might boast, support for as wide a range of hardware as Linux is most certainly not one of them.
Why should I do your work for you? Just type a simple search into Google, such as "Linux hardware" or "Linux HCL" and you'll find the sites for yourself.
The thing is... a properly configured Linux can be installed on pretty much whatever hardware you have lying around. Whereas Solaris x86 can't.
So that fact alone pisses on Sun's parade in rather a big way.
Who is going to upgrade from JDS 2 to 3 if half the upgrades won't even work properly because the supported kit is slashed? Are countries like China or large companies expected to restrict their hardware to items in Sun's draconian Hardware Compatibility List?
Besides which, who cares that it's running a Solaris kernel? This is a desktop system, not big iron. It would make more business sense to put a Solaris compatibility layer on top of Linux and benefit from the development momentum that it has and Solaris doesn't.
But I don't think Sun is thinking straight these days. Just like AOL before them, it just takes a big chunk of cash to be waved before their eyes and suddenly they're Microsoft's bitch. It would not surprise me if JDS started shipping with.NET libraries in some future incarnation.
It's pretty clear that MS are playing to their strengths here. They've always had solid support for their products (no matter how buggy they might be to begin with). And extending the support just cements that message in the minds of customers and 3rd party vendors.
This can be contrasted sharply with Linux dists like SuSE or RH. Good luck trying to find a commercial Linux that features some level of free support five years on. A year seems to be your lot in life without paying somebody more money. RH9 may even take the prize for the fastest End Of Lifed commercial OS ever. It must have certainly come as a surprise to those who bought it Near The End that their new OS was practically obsolete. Perhaps OSs should carry an expiry date sticker.
Naturally, technically competant people can Google for support after the date. But this does nothing to help inexperienced users keep their machines up to date and safe from the latest exploits. Neither does it help enterprises who *must* pay for 24/7 support and for whom the support bill is part of the TCO.
Even vendors are faced with a dilema when supporting an OS with a short life span. Do they support end of life'd OSs with all the issues that entails, or do they only ever support the latest and greatest and confuse the hell out of their customers? It's hard enough already to ship a driver or a game for Linux and the rapidly moving target makes it nigh impossible to do in a satisfactory manner.
I wouldn't say it is an MS problem at all. Perhaps XP does have a problem with certain configurations, but the plain and simple fact is that XP has been out for 2 years now. And FC1 (and other dists) work fine, so if there is a problem it lies with the Fedora QA folks and their own quality control.
Personally I'm gagging to try FC2 (I cut the DVD today), but I don't see the point of deflecting blame when it lies squarely with Fedora's own process.
The learning curve for MFC is high too. If it weren't for the wizards, you would have to write the swathes of code just to do something simple. WTL is no worse, and the consensus from anyone who has used MFC in anger is what a baroque piece of junk it really is. It's like a dogshit covered icing and marzipan - superficially tempting but take a deep bite and see how much you like it.
Really. MFC works okay if you want stay on the path, but get off the beaten track a bit - say to implement something in OLE like an OLE message filter - and it becomes a nightmare. Half the methods are not virtual compelling you to cut and paste whole classes to change a few lines. I've literally copied huge chunks of MFC to change a function that I could not override.
And MS in their wisdom have tried to merge MFC & ATL into atlmfc with duplicate classes galore. Now they're trashing ATL by deprecating some of the tried and trusted conventions such as object maps in favour of meta info that ties ATL tightly to VC++7.x. Still, ATL is fast and produces very tight code compared to MFC.
WTL works much like ATL, and neither is particularly hard to learn assuming you've seen any STL, MFC, wxWindows etc. in your life.
Interestingly, no one has pointed out that WTL depends on ATL. The CPL'd WTL depends on the proprietary ATL. So its use is rather limited. It would be great to see MS open up MFC, ATL and WTL since it would make porting apps to Wine a lot easier. Perhaps that's why it is unlikely to ever happen.
We do in my place of work, and even the likes of Mozilla was stuck with nmake all the way up to 1.0 on Win32. I doubt there is a tool that could port 100% of the changes but it could do bulk fixes to macros, and conditional directives and mark other parts with TODOs
.dsp /.vcproj files are probably more straightforward since they are machine generated although pre and post build steps would need work. Tools like cl, link and rc can be emulated with wrappers. Source code can be processed to conditionally wrap #pragmas or supply equivalent functionality.
Naturally nothing is going to do everything, but a good tool could remove hours if not days from the process. Likewise with a decent subset of MFC or ATL that ran natively - something that took a fair stab at implementing the most common classes - document / view, windows messages, GDI and the ATL COM classes could make a dramatic impact.
Of course there is no silver bullet, but a good set of tools makes the task merely daunting rather than impossible.
Why? Because unless you are prepared to write better (and unencumbered) documentation for each API, no one is going to bother to read it. After all, why should they when DevStudio offers all the help they need just by pressing F1?
And even if they did read it, it still wouldn't do much good since most software is written against ATL or MFC, or in a language that doesn't even hit Win32 directly. Not to mention that WINE would still have to support the 'broken' Win32 API anyway in order to run the tens of thousands of apps that already work out there.
In short this is a waste of time. A better approach would be to produce some porting tools to help people move their Win32 apps and build them natively using Linux using the WINE lib. That means scripts to turn nmake files into gmake files, wrappers for cl.exe, open source versions of libraries such as MFC & ATL. These libs alone are the biggest impediment to moving source code over at this time. It would non trivial to replicate to be sure, but necessary for a lot of code to move across. Even half implementations that did the more common classes (CString, CWnd, CWindow, CComObject, CDialog, CComPtr, CComBSTR etc.) would be very useful for a lot of projects.
... Atlantis is a myth. The only person who mentions the place was Plato and he may have made it up to illustrate an argument. It doesn't mean it actually existed or that if it did that Plato wasn't passing on highly distorted stories he had heard somewhere. After all, the Mediterranean is wracked by seismic activity so perhaps there was a civilization that was snuffed out by a particularly nasty event.
But just like Noah's Ark, Atlantis is a magnet for cranks and pseudo scientists. Forget painstaking archeological research - how many books have been and gone that supposedly pinpoint its exact location and toss in a few references to aliens and UFOs? We're in Graham Hancock territory here. I thought it was meant to be off Cuba only last year! Or perhaps not. We'll know more about this latest endeavour if / when a proper scientific paper appears to back it up. We'll know more when these supposed features are actually studied properly.
I reckon given a thousand years the kook brigade will be still looking for Atlantis, that is when they are not busy looking for Minas Tirith, Xanadu and Hogwarts.
Slashdot Mirror
The latter approach is more suitable for a modular suite of apps (and for development & debugging) , but it means that DLLs built for one may or may not work for the other if they pull in external symbols that are not there in the static build.
It should pose a problem unless you're doing something hardcore with Gecko but it is something to be aware of.
If the answer to any of these things is "not bloody likely", it seems a rather pointless thing to me. Why subscribe to kill your bandwidth watching crappy movies in crappy quality?
Hover over a buddy in AIM and you'll see what capabilities the other person has available to them, e.g. chat, video etc. I assume that AIM would be smart enough to grey out options that were not applicable for the person you had selected.
Anyway sooner or later anything running SCO is doomed.
And the really ancient thing they're using here is SCO Unix which as far back as I can remember has never been 'modern' at any point in time. But as a *nix, porting from it to Linux should be a relatively straightforward proposition (as Autozone already know much to SCO's consternation and bogus lawsuits). And it may be that the vendors who wrote the databases, stock control / POS / accounting / payrole software have already done the work since Linux is a tier 1 platform.
Who knows, perhaps it was the vendors who motivated McDonalds to quit. Maybe they said to management what everyone already else knows - SCO is a piece of shit, the company is going down the plug hole, it won't run on hardware X and it'll cost $$$$ to support it even if it did. No doubt McD's IT department pushed the same story from the inside.
It's another kick in the balls for SCO which is always good for a laugh. It's too bad the markets are closed today because their stock is doing a good impression of the Titanic.
So long SCO, we hardly knew ye.
Anyway, don't take this as advice (since I'm a newbie investor & obviously not an advisor of any sort) but the reason I say I am still a shareholder (not was - am) is because I have confidence in NOVL & IBM. I could have dumped the shares at a loss, but I believe that both have a business and products to emerge on the other side through the legal tussles whereas SCO won't. The only plan SCO seems to have is sue, fud, sue, threaten, bluster and whine. That would be fine if there were a genuine case, but nothing from their own lips or otherwise says there is.
It would have been great with hindsight to have ridden the coattails of the SCO FUD and make a small fortune trading against the hype - the highs and the lows of SCO are enviable with hindsight. But by this point the future of the company is already carved in stone - RIP. Anyone who hangs onto their shares at this stage is either extraordinarily bullish or fantastically stupid.
While I appreciate that GNOME / KDE aren't the lightest WMs, they're about the only ones that are proper desktops as far as mere mortals are concerned.
I'm sure you could get similar functionality by cobbling a WM, a terminal app, some kind of file browser, etc. I've put up with that kind of crap on Unix for 15 years when lesser machines such as the Mac, Amiga, Atari ST etc. had it way back then. I'm glad that the mainstream has finally left that kind of mentality behind. It doesn't stop anyone using GNUStep or E or CDE or whatever, but unless you are seriously strapped for memory, or the box lives most of its life unattended there is little point.
Personally I just enjoy having a proper desktop because I despise screwing around in some config file to add a lousy icon or to change the screen resolution, or having to run mix and match apps to be able to browse files, networks, printers etc. when they are all inconsistent with each other and the WM. Give me GNOME any day.
If Microsoft want to waste their money preaching to the unconvertible then let them. Whether it is Slashdot or Linux Today they might as well be pissing their money into the wind for all the good it will do them. At least this way, a Linux oriented enterprise benefits from Microsoft's desperation.
Plus, if some customer on a Segway is brutally assaulted / smashes into a pedestrian / plows into a post you can look forward to an expensive lawsuit in addition to the loss of the segway he / she was riding on at the time.
Fortunately it is still possible to run an explorer style window too, but overall GNOME 2.6 gets a big thumbs up from me - spatial, minimalist, uncluttered.
While it would be nice to see tools such as blogging etc. but I'd rather see services to support such tools. For example, Evolution has a nice day view that gives weather, rss news and other stuff. Something like that would be very useful for the desktop, where the content is supplied through an API that any app can plug into.
Perhaps but then again you *need* it. I don't think twice about sticking a Fedora disk in and installing it. With Gentoo I'd be afraid to even start installing unless I had another computer open beside the target machine so I could tentatively proceed one step at a time.
I wonder if Gentoo actually needs to be so hard to install. After all, it could probably boot into a graphical or menu driven installer just like other dists, but with the difference that it additionally builds some packages instead of copying the binaries.
In summary, if you had a point (and it's hard to see that you did) you've lost it.
Whatever other qualities Solaris might boast, support for as wide a range of hardware as Linux is most certainly not one of them.
Why should I do your work for you? Just type a simple search into Google, such as "Linux hardware" or "Linux HCL" and you'll find the sites for yourself.
I suggest you review Sun's own hardware compatibility list before crying foul.
So that fact alone pisses on Sun's parade in rather a big way.
Who is going to upgrade from JDS 2 to 3 if half the upgrades won't even work properly because the supported kit is slashed? Are countries like China or large companies expected to restrict their hardware to items in Sun's draconian Hardware Compatibility List?
Besides which, who cares that it's running a Solaris kernel? This is a desktop system, not big iron. It would make more business sense to put a Solaris compatibility layer on top of Linux and benefit from the development momentum that it has and Solaris doesn't.
But I don't think Sun is thinking straight these days. Just like AOL before them, it just takes a big chunk of cash to be waved before their eyes and suddenly they're Microsoft's bitch. It would not surprise me if JDS started shipping with .NET libraries in some future incarnation.
This can be contrasted sharply with Linux dists like SuSE or RH. Good luck trying to find a commercial Linux that features some level of free support five years on. A year seems to be your lot in life without paying somebody more money. RH9 may even take the prize for the fastest End Of Lifed commercial OS ever. It must have certainly come as a surprise to those who bought it Near The End that their new OS was practically obsolete. Perhaps OSs should carry an expiry date sticker.
Naturally, technically competant people can Google for support after the date. But this does nothing to help inexperienced users keep their machines up to date and safe from the latest exploits. Neither does it help enterprises who *must* pay for 24/7 support and for whom the support bill is part of the TCO.
Even vendors are faced with a dilema when supporting an OS with a short life span. Do they support end of life'd OSs with all the issues that entails, or do they only ever support the latest and greatest and confuse the hell out of their customers? It's hard enough already to ship a driver or a game for Linux and the rapidly moving target makes it nigh impossible to do in a satisfactory manner.
Have you seen the system requirements for games like Far Cry?
Personally I'm gagging to try FC2 (I cut the DVD today), but I don't see the point of deflecting blame when it lies squarely with Fedora's own process.
How do they keep their prices so low and still make a profit?
Legolas surfed down the trunk of an Oliphaunt in ROTK too. Cool scene though, partly because it was so completely over the top.
Change code from:
IF kaboomSignal THEN
blowup
ENDIF
To:
IF NOT dontKaboomSignal THEN
blowup
ENDIF
Really. MFC works okay if you want stay on the path, but get off the beaten track a bit - say to implement something in OLE like an OLE message filter - and it becomes a nightmare. Half the methods are not virtual compelling you to cut and paste whole classes to change a few lines. I've literally copied huge chunks of MFC to change a function that I could not override.
And MS in their wisdom have tried to merge MFC & ATL into atlmfc with duplicate classes galore. Now they're trashing ATL by deprecating some of the tried and trusted conventions such as object maps in favour of meta info that ties ATL tightly to VC++7.x. Still, ATL is fast and produces very tight code compared to MFC.
WTL works much like ATL, and neither is particularly hard to learn assuming you've seen any STL, MFC, wxWindows etc. in your life.
Interestingly, no one has pointed out that WTL depends on ATL. The CPL'd WTL depends on the proprietary ATL. So its use is rather limited. It would be great to see MS open up MFC, ATL and WTL since it would make porting apps to Wine a lot easier. Perhaps that's why it is unlikely to ever happen.
Naturally nothing is going to do everything, but a good tool could remove hours if not days from the process. Likewise with a decent subset of MFC or ATL that ran natively - something that took a fair stab at implementing the most common classes - document / view, windows messages, GDI and the ATL COM classes could make a dramatic impact.
Of course there is no silver bullet, but a good set of tools makes the task merely daunting rather than impossible.
Why? Because unless you are prepared to write better (and unencumbered) documentation for each API, no one is going to bother to read it. After all, why should they when DevStudio offers all the help they need just by pressing F1?
And even if they did read it, it still wouldn't do much good since most software is written against ATL or MFC, or in a language that doesn't even hit Win32 directly. Not to mention that WINE would still have to support the 'broken' Win32 API anyway in order to run the tens of thousands of apps that already work out there.
In short this is a waste of time. A better approach would be to produce some porting tools to help people move their Win32 apps and build them natively using Linux using the WINE lib. That means scripts to turn nmake files into gmake files, wrappers for cl.exe, open source versions of libraries such as MFC & ATL. These libs alone are the biggest impediment to moving source code over at this time. It would non trivial to replicate to be sure, but necessary for a lot of code to move across. Even half implementations that did the more common classes (CString, CWnd, CWindow, CComObject, CDialog, CComPtr, CComBSTR etc.) would be very useful for a lot of projects.
But just like Noah's Ark, Atlantis is a magnet for cranks and pseudo scientists. Forget painstaking archeological research - how many books have been and gone that supposedly pinpoint its exact location and toss in a few references to aliens and UFOs? We're in Graham Hancock territory here. I thought it was meant to be off Cuba only last year! Or perhaps not. We'll know more about this latest endeavour if / when a proper scientific paper appears to back it up. We'll know more when these supposed features are actually studied properly.
I reckon given a thousand years the kook brigade will be still looking for Atlantis, that is when they are not busy looking for Minas Tirith, Xanadu and Hogwarts.