Microsoft: It's TIME at LAST. Yes - .NET is going OPEN and X-PLATFORM Microsoft, aiming to broaden the appeal of its .Net software development platform, has aggressively committed to delivering a new version of the toolset that is not only language agnostic but also cross-platform and entirely open source. The software giant announced the move at its Connect(); virtual developer event, which is …
You don't have to get involved if you don't want to.
You could sit here with with the other trolls and complain about proprietary software, conveniently ignoring all those annoying facts. I rather hope you do, actually. That means there'll be less competition for cross-platform .NET developers.
Personally, I'm rather looking forward to using C#5 on linux without having to step back a version or two to the last stable mono stack.
Personally, I'm rather looking forward to using C#5 on linux without having to step back a version or two to the last stable mono stack.
Sure, you do that. I'm a bit older, so I am a tad more cautious - I really don't want to help MS in another round of embrace, extend and extinguish so pardon me while I keep well clear of this until the lawyers have finished. And then I'll wait until the security people have taken it apart. After that, OK, we may allow it on a non-MS platform for tests.
But we'll probably just get on with the alternatives we developed a while back, because that's stuff we already trust.
> embrace, extend and extinguish
Oh fuck off with your FUD unless you can find a way to Embrace Extend Extinguish the Apache 2 license.
Here's a hint - lots of people would make a lot of money if you can so it's probably been tried.
And by the way - I very much doubt that you're older than I am unless this is elderly paranoia we're witnessing. Have you gone racist as well?
The real question is .Net going to still be relevant by the time its been fully developed and is truly production ready on non MS platforms.? After the Silverlight debacle Microsoft has in past given mixed signals about .Net's long term future. In some ways this may turn out just like it did with Symbian.
Well, it's been around for 13 years so far. It's probably not going to get EOLd anytime soon.
There's always been a fear that they will, purely from their track record.
But now that it's under Apache 2.0, even if MS do EOL it, it can continue just like every other framework we enjoy.
In fact, would it be at all bad if they did EOL it? - Imagine a "community" driven version, rather than one that goes only in the direction that MS decides.
> There's always been a fear that they will, purely from their track record.
What, a track record of supporting operating systems longer than any other vendor and development frameworks for - if we count the switch from Sun to Oracle as a new owner, which it is - longer than any other vendor?
About the only framework MS ever axed without legal pressure (JScript) is Silverlight and as I recall, all the massed ranks of El Reg commentards laughed at Silverlight because it was stupid and M$ and clearly shitty M$ stuff and useless.
Does my memory fail me?
"What, a track record of supporting operating systems longer than any other vendor"
...
"Does my memory fail me?"
I suspect that your memory has quietly forgotten Vista, and you haven't ever heard of folks like IBM and HP. In IBM's case they're still supporting an OS that was shipping before Bill flunked college. As it happens even newbies like Slackware, SUSE & Redhat support OSes that predate Microsoft's NT 3.51. :)
> In IBM's case they're still supporting an OS that was shipping before Bill flunked college.
That goes two ways. IBM don't support any codebase older than 9 years without special (expensive) arrangement. I know this because the company I work for has just upgraded rather than enter into such an arrangement. Now, the name "DB2" may be older than 9 years - the version is not.
SlackWare, Suse and RedHat do not support "OSes that predate Microsoft's NT 3.51" - they support more recent iterations.
If I were judging MS the way you judge IBM et al, I could say "Microsoft have been supporting this single OS (Windows) since September 1981" but that would be disingenuous at best so I won't. I would be obliged if you'd show others the same courtesy.
"SlackWare, Suse and RedHat do not support "OSes that predate Microsoft's NT 3.51" - they support more recent iterations."
Err, an iteration of an OS is just a new version of the same code base. You were talking about "operating systems", not releases of OSes after all.
"If I were judging MS the way you judge IBM et al, I could say "Microsoft have been supporting this single OS (Windows) since September 1981" but that would be disingenuous at best so I won't. I would be obliged if you'd show others the same courtesy."
MS didn't ship Windows 1.0 until 1985, so no, you couldn't make that claim on the basis of product name. If you wanted to base it on code-base then you would have to limit your claim to NT and it's descendents released back in '93 (I don't know of anyone who ran NT 3.1).
If we take the "every release is a new OS" approach then MS don't look any better than Redhat. Their current oldest supported release is Vista, released in 2006, extended support ending 2017, by contrast Redhat released RHEL4 in 2005 and it's extended support will also end in 2017.
So "a track record of supporting operating systems longer than any other vendor " is not true. MS does support their releases for a long time, just not quite as long as (some) other vendors.
There's always been a fear that they will, purely from their track record.
I'm not sure there has. There's always been an accusation of fear, from Java developers mostly, but no actual fear. I've been hearing the same rumour professionally since sometime around version 2.0 and yet here we are, still more spend, more R&D, and a greater embedding in Microsofts own products.
Cross platforming this and making (some of) the toolset free will massively reset the market. Most companies that use Java use it because it's free not because it's the best framework.... it just costs nothing to use running on linux. Free is a powerful motivator for a start up.
Personally, I hope neither Java nor .Net become dominant, as competition between them will produce the best results for us all.
Oh fuck off with your FUD unless you can find a way to Embrace Extend Extinguish the Apache 2 license.
Here's a hint - lots of people would make a lot of money if you can so it's probably been tried.
And by the way - I very much doubt that you're older than I am unless this is elderly paranoia we're witnessing. Have you gone racist as well?
I think you're a tad deluded to find racism anywhere in that post, but I guess you have to distract people from what happened when Microsoft touched Kerberos. If I recall correctly that was also under an Open style license. As for age - it appears my memory still seems to work..
What happened when Microsoft touched Kerberos was that they built a completely standards based version. Everyone got fussed about the way they used the PAC data field to store group memberships in the token, forgetting (deliberately, mostly) that that field is there for implementers to use as they will. MS did endless interoperability testing with other implementations, and as it happened found problems with all of them where they weren't compliant with the standard.
Kerberos was/is under an open style license - and MS honoured that. I doubt you will remember the truth of this, because you probably don't want to.
You do it with versions:
The aim is to get devs back to windows programming. .net version X is Apache and once everyone is using it, you release the next version as proprietary. You can stay on version X but windows 2016 doesn't support it any more. If you're doing a one-off product, you might be ok, but I'd always be wary of a library that a company tosses over the fence. Libraries have no intrinsic value so you want to be sure the FLOSS community is interested in supporting it before you use it.
Paranoid? Perhaps, but gifts from Greeks come to mind.
While this is extremely good news - .NET is a very well thought out development platform - releasing the "entire .NET stack" isn't quite the ultimate step to portability it might appear. Big chunks of .NET are wrappers around the Windows API, which is largely why Mono is "mostly" complete rather than a complete drop-in replacement.
Nevertheless, it's nice to see .NET getting some attention again after Microsoft appeared to be losing interest.
I used to be a pretty big critic of Miguel. I'm pretty sure at one point I accused him of secretly working for Microsoft to trap GNU/Linux and that Mono was a cunning ploy to tie GNU/Linux to an environment that Microsoft controlled. I was a pretty fierce critic of Mono for that reason.
I may have been a little hard on him given today's news. Oops.
Could be very useful, the problem may come if MS decide that this doesnt work for them any more. Ok, they can't 'un-open source' but they can close off any potential development routes that might conflict for them. That may hurt someone who has gone wholesale into .net because of this.
Another option opens up, which is a good thing.
What will you do if you have developed something, run in on Linux machines... and one day Microsoft decided not to support Linux any longer.
You're stuck with whatever version there was and hope there would be enough community to keep supporting/developing it for Linux, or you bite the bullet and switch to Windows (like God intended)
> What will you do if you have developed something, run in on Linux machines... and one day Microsoft decided not to support Linux any longer.
What happens if Steve Ballmer's angry ghost comes along and rapes your penguin?
Let's not deal in paranoia, shall we?
> What will you do if you have developed something, run in on Linux machines... and one day Microsoft decided not to support Linux any longer.
What happens if Steve Ballmer's angry ghost comes along and rapes your penguin?
Let's not deal in paranoia, shall we?
Microsoft has been gaming the software world pretty much since Gates relabelled 86-DOS, so that's +30 years worth of track record they have to undo. I will not trust them with anything more complex than a pocket calculator, and only if I have my own to validate their results.
The DNA of Microsoft does not contain innovation or collaboration genes, and there is so much evidence of their "friendliness" being nothing more than the smile of a shark that it would be seriously stupid to just jump into bed with anything they have dreamt up without covering your back pretty well. That's not paranoid, that's called learning from history. You know, the stuff you're doomed to repeat if you haven't learned it?
>>"What will you do if you have developed something, run in on Linux machines... and one day Microsoft decided not to support Linux any longer."
Well that's true of any language or framework, Open Source or otherwise - you have to hope there are enough others out there like you that it wont wither away and die. A couple of things with .NET are that (a) it's already pretty mature and widely-used so that's good. And following from that, even if MS abandon support for it on GNU/Linux, they'll continue to support .NET itself so the job of maintaining it on GNU/Linux is a lot simpler than if it were abandoned altogether. You're always going to have a sizeable community next door and that has knock-on effects - you essentially just have to re-implement any changes rather than come up with them yourself. And it massively enlarges the pool of available programmers too.
In fact, I'm surprised at this decision because of that last point. What this does is enable a lot of Windows programmers to [more] easily migrate to writing software on GNU/Linux.
dogged - "Yeah because using the Apache license means MS retain TOTAL CONTROL"
An Apache licence means that they can return to a closed source model again if they want to without having to purge any community contributions. They may not do that, but then again we don't know what some future CEO might have a brain wave over and it's happened before with other software. You just have to trust them not to do that.
They couldn't do that with GPL licenced software without purging all contributed third party code, which would be rather difficult after a while. That's why projects like the Linux kernel use it - it guaranties that everyone who contributes to it gets treated equally.
There is method in Microsoft's madness, and they are keeping their options open in case they ever change their minds about open source. Yes, you'll still have the old versions, but those would get out of sync with new releases within a few years.
*sigh* I forgot to use the 'Joke' icon and it seems there are some touchy MS devs.
The big, successful open source projects are bankrolled. Hobbyists play a part but there are a lot of academics and professional devs funded by businesses. Xamarin is doing that for Mono. But, in the long term, it's hard to see them competing against Microsoft. I hope they survive. But if they don't, and some future Microsoft CEO, scrabbling for revenue, decides to close the doors on public development, then there will be no funding to develop the code. Perhaps, in that situation, interested parties would fund a fork; who knows what the future holds? (Hey, Microsoft could go bust and leave Xamarin.) But while I don't think Microsoft are engaged in a deliberate E3, it could still end up looking like that in a decade's time.
Yeah because using the Apache license means MS retain TOTAL CONTROL and if you use their code they can LOOK IN YOUR PANTS AT ANY TIME and then EMBRACE EXTEND EXTINGUISH ALL OF OPEN SOURCE
or alternatively, you've got serious prejudice problems.
Ah,I see you're today's MS shill. One word: Kerberos.
"Well .Net is faster and more efficient, easier to install and manage, has had a couple of orders of magnitude fewer security holes, and supports multiple programming languages....And it's not made / supported by Oracle."
Have an upvote for all points except being easier to install and manage. Managing Java on 'nix, even at scale, is a walk in the park compared to .Net runtimes and service packs. Hopefully when Windows introduces their new package management system (as planned) then we can start seeing some parity in that respect.
the main issue we have is that nearly EVERYTIME there is an update for Java (just about every week!) it fecking breaks something! So I generally stick with the version we're running which is currently 7 update45 then pick a new release and test all the permutations of applications\OS that we run with said new Java update, And then repeat, grrrrrrrrrrrrrrr
>>"Well .Net is faster and more efficient, easier to install and manage, has had a couple of orders of magnitude fewer security holes, and supports multiple programming languages....And it's not made / supported by Oracle."
Alright, alright. But apart from all that, what have the Redmonds ever done for us?
"Well .Net is faster and more efficient..."
That may be true.
"...easier to install and manage..."
I don't know if I'd agree with that exactly. They both have their weaknesses. One thing that always gets me about .NET is that if something goes wrong with the installation, generally the only way to fix it reiiably is to reinstall the operating system. Otherwise, you can try to uninstall, mess with the registry, reinstall, and hope for the best.
"...has had a couple of orders of magnitude fewer security holes..."
I think you're comparing apples to oranges here. Most of the security holes related to Java have been security holes in the browser sandbox. There is no widely deployed browser sandbox for .NET (one has not been practical because it has not previously been cross-platform, and you can make a pretty good case that security concerns make such a sandbox impractical in any case).
"...and supports multiple programming languages..."
Actually, you can run multiple programming languages on Java's virtual machine as well, though the others are not widely used.
"And it's not made / supported by Oracle."
I'll give you that one, though being made and supported by Microsoft brings its own set of issues.
To be honest, I'm not sure I like the idea of a cross platform .NET better than Mono and the alternatives already available (Python, Ruby, etc.). When talking about Mono I really feel like I can say it's easier to install and manage (though Java on Linux is pretty easy). However, though this is clearly a smart move for Microsoft to better support its Azure efforts, it's also a good move for developers. An open .NET is much better than a closed .NET for us as well as Microsoft. I just hope it doesn't adversely affect the Mono project over the long run.
"Fair play to MS for providing a life-raft for developers deserting the legacy Windows platform"
.Net already runs on more recent Windows versions than XP.
This is about providing cross platform options for other niche OSs like OS-X and legacy UNIX type platforms and derivatives like Linux. So developers can for instance code / test on a Linux platform that provides the environment they might like to use for day to day coding, but then can be run in production on a standardised Windows Server / Azure platform where you need features advanced features like clustering, and with advantages like high security features (for instance constrained delegation, expression based ACLs, etc.), lower vulnerability counts, automated scale out ability, lower TCO, etc. etc...and you can likely support multiple client OS versions with only a few tweaks and a recompile.
Interesting so you claim this is so developers can go home to develop on their Linux boxes at home (or at work on the desktop) while only using windows servers for production. Might have a point for some time if we are talking about .Net but in general in industry there are probably now nearly as many Linux/Unix boxes in production as there are Windows. Especially if counting web servers and such. By the way all those advanced security, reliability, and performance features you mentioned appeared fairly late comparatively in the windows world. Keep on trollin' though or better don't.
".Net already runs on more recent Windows versions than XP."
Sure it does, but it doesn't change the fact that Windows is a legacy platform. Windows 8.1 ticks all the boxes that OpenVMS did on the day of it's release... Single vendor, a 'mature' ecosystem, failing to gain significant market share in the fast growing market segments of the day, it's a legacy platform.
Choosing to post as AC speaks volumes for the quality of your unsubstantiated claims of "lower vulnerability counts". If you believed your assertions you wouldn't have a problem with associating them with your account name, did the 19 year old remotely exploitable vuln spook you ?
For the record I reckon C# and .NET are better than Java, so I am pleased to see that there is an escape route for users of that tool chain.
'"Fair play to MS for providing a life-raft for developers deserting the legacy Windows platform"
.Net already runs on more recent Windows versions than XP.'
Ah, but this is the salesman's definition of "legacy": that which we don't sell or support. (Or the developer's: that which I can't stand to maintain.)
"This is about providing cross platform options for other niche OSs like OS-X and legacy UNIX type platforms and derivatives like Linux."
Niche? Legacy? This is the shit that runs the world! I don't see Google, Facebook, Apple, Amazon, Vimeo, Dropbox RackSpace, et al. running all their major services or infrastructure on Windows. Or most of the world's supercomputers for that matter. In fact outside of Microsoft or internal enterprise stuff, who actually uses Windows for running large-scale computing at all?
There must be a reason for this.
>>This is about providing cross platform options for other niche OSs like OS-X and legacy UNIX type platforms and derivatives like Linux. So developers can for instance code / test on a Linux platform that provides the environment they might like to use for day to day coding, but then can be run in production on a standardised..."
I really don't think that is the case. For one, GNU/Linux is not a "niche" OS. It's hugely successful on the server side. And OSX is not a "niche" OS. It's hugely successful on the client side.
But to your main point, I don't think that makes sense at all. You develop on a system that is as close to your target environment as possible. Even minor differences can have drastic results. (In fact the minor ones can be the hardest to find and debug - from recent experience, alas).
I don't doubt that MS want to get as many people as possible using their platform but the above as a method doesn't make sense to me.
I guess they got tired of totally being locked out of anyone really using their stuff on all the latest trends like OpenShift, CloudStack, Amazon EC2, Google Apps etc..
Nearly every new cloud and PaaS platform is based on OpenSource, because lets face it, what cloud provider wants to deal with software licensing when building platforms to offer their customers?!!
Microsoft has been totally shut out of this ecosystem outside their own Azure stuff, and no one even talks about anything Microsoft these days in the world of the cloud.
"because lets face it, what cloud provider wants to deal with software licensing when building platforms to offer their customers?!!"
Hyper-V Server is completely free. With no feature limitations. No license fees for the fully featured product...
http://www.microsoft.com/en-us/evalcenter/evaluate-hyper-v-server-2012-r2
What I mean by the somewhat strange title is whether Microsoft will not just open source the .Net stack with resulting releases on Linux, but will there also be a "enterprise pays for support" option?
You know the companies I'm talking about. The ones who, faced with CentOS vs RHEL will always go for the latter because that way "we can call someone if things go wrong". And consequently forbid deployment unless they can actually send an annual cheque for support to somebody.
I think you mean will Microsoft offer support contracts for DotNet on other platforms like Oracle does for Java?
They might, but I doubt it will make much difference. Mono never saw much interest on server or desktop platforms, and I don't see why Microsoft's version would be any different outside of the Windows platform.
Back in the early days most of the major Linux distros packaged up Mono and offered support for it, but developers weren't interested. The distros then gradually dropped it from their standard installs and supported repos, although you can still install it if you want.
Given the above history, I'm quite honestly struggling to see what Microsoft's DotNet is going to offer that Mono didn't already. If it's just that you don't like Java, well there's loads of other languages to choose from. There are no core software components written in Java on any Linux distro that I can think of. For example you don't get even get Java in a standard install for Ubuntu, it's an optional install if you happen want it. It's something that is mainly used for enterprise server applications, and quite frankly I can't see those guys changing tack now. Java people will continue to use Java.
I think that what Microsoft is really hoping for here is to keep mind share active for C# for use in deployments on Microsoft Azure. The thing about interlocking vendor lock-in is that when your platform is on the rise, all the proprietary bits and pieces work together to reinforce each other. When your platform is in eclipse though, then the inter-dependencies tend to work the other way, dragging the platform down even faster.
So, I see this as a defensive move on the part of Microsoft. They know that Windows is a legacy platform. It's not the thing that is driving the future of the industry, on mobile, tablets, or cloud. Try to think of a big well known web corp that is using Windows - other than Microsoft themselves, they're rare as hen's teeth. To keep relevant, they have to follow the developers, and the developers who are defining the future of the industry are not using Windows. If Microsoft did this ten, or even five years ago, it might have worked. Now though, I think they've missed the boat.
I'm quite honestly struggling to see what Microsoft's DotNet is going to offer that Mono didn't already.
If it's released under the Apache 2.0 Licence, as implied in the article, then what this offers is an explicit patent licence grant.
Mono always seemed to come with threats of patent prosecution if you didn't get it from Suse...
Vic.
>>You know the companies I'm talking about. The ones who, faced with CentOS vs RHEL will always go for the latter because that way "we can call someone if things go wrong".
I don't know why you put that last part in quote marks. You can call RedHat if things go wrong - it's a major and good thing for the Enterprise customers.
Wake me up when I no longer need a gazillion .Net runtimes to run apps.... unlike Java, where I can run 1.3 code in a 1.8 JVM... might need a little tinkering, but mostly it works.
You have v3.1 .Net runtime and the app wants 2.x, you're fucked. This sucks golf balls through garden hoses.
I think this is a desperate effort to promote an outdated, useless piece of crap... consider .net dead in 5years.
"where I can run 1.3 code in a 1.8 JVM"
"You have v3.1 .Net runtime and the app wants 2.x, you're fucked"
Apples to oranges!
You're comparing v1.3 to v1.8 - both with a major version of ONE - to v3.1 to v2.0 - which has two different major versions.
Damn, have you seen SDL v2 - it's nothing like SDL v1.2 (though they are building a compatibility layer IIRC). Allegro v4 to Allegro v5.
Take your strawman, and go home
It's not even true. The 3.x runtime will happily run code compiled against 2.0 unless the idiot who wrote that code has explicitly told it not to run without 2.0.
3.5 will run any previous code. 4.6 will run any 4.x code.
The maximum number of .NET framework versions you need is two. So we have learned that for Hans, "two" equals "a gazillion".
Colour me very, very sceptical. Microsoft does not have a history of being honest about anything, from secret APIs to dishonest dealings with "partners". I can only wonder what hidden traps they'll have in their definition of "open source", or their implementation of the "open source" version.
Colour me very, very sceptical. Microsoft does not have a history of being honest about anything, from secret APIs to dishonest dealings with "partners". I can only wonder what hidden traps they'll have in their definition of "open source", or their implementation of the "open source" version.
Yup. That said, the comments here seem to be divided between MS shills and those who seem to be bad at history, and who are consequently doomed to repeat it. I would avoid creating any corporate dependency on Microsoft code unless you're planning to flog or float your business soon - it's not good long term planning if you want to keep your IT costs down.
I think this may originate in reality in the UK. Current UK government has thrown out the Microsofties that the previous administration allowed to infest everything, so they now need a different set of clothes to show that they're oh-so-supportive of "open" or totally be out on their ear as Cabinet Office appears to be using People With A Clue™ (come to think of it, I may met a few in Brussels a few weeks ago).
Microsoft have been "nice" with bad consequences far too often. Trust must be earned, and they have a *lot* of catching up to do.
Are we allowed to compile System.Data namespace back into VS 2013 for windows and develop coherent Metro Apps with local db's. Just when I was starting to love microsoft again, they pull this and gimp their API so significantly that I'm stuck in VS 2010 for the foreseeable future.
"The real question is .Net going to still be relevant by the time its been fully developed and is truly production ready on non MS platforms.? After the Silverlight debacle Microsoft has in past given mixed signals about .Net's long term future. In some ways this may turn out just like it did with Symbian."
To be honest, i've done some test programming with C#, it's fine. I must agree about the mixed signals though, I had though .NET was considered essentially defunct, that Microsoft was done with it and .NET 4.5 was it.
I don't really develop for Windows, other than dipping my toes now and then. But, I certainly do hope for Windows developer's sake that Microsoft comes out with SOME kind of clearer roadmap.
When I did some Windows development recently, I decided to find out what the "best practices" were; since I was starting from scratch, why use outmoded and no longer recommended toolkits and techniques? Well, I found no recommendation, every desktop toolkit was either formally deprecated or not really deprecated but they were making it clear(ish) that development was done. Some developers said the newer toolkits weren't feature-complete and to use the older formally deprecated ones. Very confusing. The only formally recommended development path at that point was to develop Metro apps, which I was obviously not going to do since a) I don't think enough people had Windows 8 then or indeed now, and I don't have a Windows 8 VM either for that matter. b) I didn't like any aspect of any Metro app I've ever seen, since it's not really a desktop interface but a tablet interface forced onto a desktop.
So I decided on WPF (despite it also being considered to be kind of on life support on that point) and found it acceptable; XAML and WPF reminded me the most of the old Swing toolkit for Java. However, I can see where people could find it inflexible for their purposes. But again, even WPF was not recommended at that point, and indeed it sounded like even .NET was kind of being placed on life support (although specific addons like Entity Framework were even then clearly still being actively developed.)
At least Microsoft has indicated they will continue to work on WPF and .NET, giving SOME kind of guidance.
It is incredulous for Microsoft and it's supported development community to believe that developers for Free/ Open source Software (FOSS) tools in Java, OpenStack Cloud Computing, KDV/Xen Virtualization, Apache Hadoop big data analytics and other development like Docker Containerization will want to transition to Windows development platform when every one of these FOSS projects senior development managers have stated quite unambiguously that their software is developed and runs more reliably, more efficiently - faster, more scalable and flexible, with far less bugs and base OS security vulnerabilities on Linux or BSD UNIX-like than any Microsoft foundation development environment.
I applaud Microsoft CEO Mr. Nardella on his aims of embracing the FOSS” type” development model, probably through his many years experience as a true technologist, but the Microsoft hegemony of software development has always been very strictly proprietary and non-compliant to Open, International standards, and it would take years without any guarantee of success to amalgamate the Microsoft kludge systems to new standard of (true - as defined by fsf) FOSS development.
All that stale Microsoft baggage just won't fit in.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
