The Case For Supporting and Using Mono 570
snydeq writes "Fatal Exception's Neil McAllister argues in favor of Mono, asking those among the open source community who have 'variously described Mono as a trap, a kludge, or simply a waste of effort' to look past Miguel de Icaza and Mono's associations with Microsoft and give the open source implementation of .Net a second chance, as he himself has, having predicted Mono's demise at the hands of open source Java in 2006. Far from being just a clone of .Net for Linux, McAllister argues, Mono has been 'expanding its presence into exciting and unexpected new niches.' And for those who argue that 'developing open-source software based on Microsoft technologies is like walking into a lion's den,' McAllister suggests taking a look at the direction Mono is heading. The more Mono evolves, the less likely Microsoft is to use patent claims or some other dirty trick to bring down the platform."
Qt (Score:5, Interesting)
With Qt 4.5 going LGPL in March, one would have to wonder why you would use Mono over Qt or Java.
There are legitimate reasons - the CLR for instance or the multi-language support. But Qt has a Java API if you're addicted to virtual machines, and the C++ toolkit compiles anywhere with a modern C++ compiler. It supports Javascript (QtScript) and Python bindings. But unlike Mono, which is Microsoft derived, there will be no patent worries. Nokia really does want Qt everywhere.
The picture is getting more and more complicated when it comes to software development, and I think that's wrong. I liked .Net as an idea. We could all code to one platform, but the business/IP aspects prevented that technical utopia. I am hoping that LGPL Qt will, while a little more limited be that multi-platform toolkit that everyone can use to solve new problems, instead of continually recoding the old ones.
Re:Objective Review (Score:5, Interesting)
Ohhh, yeaaah. I remember this guy. This is the same nitwit who used logic from 1996 to try and convince us all to burn our webapps [slashdot.org]. I see he's back with even more faulty reasoning.
I guess there's only one thing to say. Slashdot, meet the new John C. Dvorak [slashdot.org].
WPF Support (Score:5, Interesting)
Until Mono gets WPF support there isn't going to be much cross-compatibility. Any Windows .NET developer with any sense is writing in WPF already. WinForms is dead.
But Mono seems quite content to ignore WPF for now. One can't help but think it was part of that Novell/Microsoft deal.
The subset of WPF in Moonlight is useless for non-web development. It's great way for MS to pretend their Flash-killer format is multi-platform though.
Re:The thing is... (Score:4, Interesting)
Re:But the political reasons... (Score:1, Interesting)
I don't agree here at all. If you look back at my comments you see that I am a Mono critique. But frankly I have changed my perspective.
Mono is actually all right...
What Mono did and it surprises me is that a Mono developer = Microsoft .NET developer, but a Microsoft .NET developer != Mono developer.
What I am saying is that if you learned Mono you can use your skills on Microsoft .NET, but not the other way around. This is because of the libraries that they use. Frankly this is good since it means they are independent and will adopt what they need to adopt no more no less.
Re:The thing is... (Score:3, Interesting)
Yeah, because you can compile standard C and standard C++ on just about EVERY platform. Unfortunately the phone OS makers do not want us to make native code and instead want us to program Java apps. With that we will never be able to trash the idea that phone apps are slow and pseudo-3D.
Re:Mono just miss one thing (Score:4, Interesting)
It's unfortunately not easy to develop on Mono right now, but IMHO only due to the debugger.
I disagree. MonoDevelop is the bane of my existence. It's not even that it's missing features - it's that the damn thing crashes randomly and the basic features (like code completion) are broken. It's been this way for me for years... so long that I almost wanted to start contributing to the project. But then I just installed Visual Studio 2008 on a Windows VM and it solved everything.
I swear, I haven't really hobby coded too much since I started using Linux years ago. Part of it is because I have everything I need and don't need to change much. The other part of it is that I haven't used a single goddamn IDE in Linux that doesn't make me want to shoot myself in the face. Fanatics can gab on about how a real developer doesn't need a a decent IDE, and that's true - but what's also true is that once you've had access to elegant debugging, code completion, and compilation, you don't ever want to go back.
Re:Qt (Score:5, Interesting)
That's a rather rosy way of looking at it. In truth, Sun sort of messed up the marketing on a variety of different levels. The only reason why it caught on was that developers had a chance to try it out through Java Applets. (Originally supported through a partnership with Netscape.) Developers who tried the Java platform were so impressed by it (compared to the standard C library of the day) that they lobbied for its use everywhere. In fact, there was an entire news site devoted to Java promotion called JavaLobby. In its day, it handily competed with sites like Slashdot for readers.
Sun (thankfully) wasn't stupid. They took a look at where developers were trying to use Java and started supporting them. Some of the ideas took flight (e.g. servlets) while others floundered (e.g. Java3D). But at the end of the day, Sun managed to produce a superior platform that the vast majority of the market wanted to use.
The only reason why it gets so much criticism today is because other languages and platforms have invested considerable effort in catching up to where Java is today. And even then, it's hard to argue the rich availability of libraries for the Java platform. If you use Java, you are guaranteed to always be able to use the latest and the most obscure technologies. Nothing escapes its roving field of vision.
When Java gets replaced (and I'm sure it will happen eventually), it will not be because of marketing. It will be because the replacement platform yet again turns the entire industry on its head. Love it or hate it, the new technology will make us all step back and think.
Re:Mono just miss one thing (Score:1, Interesting)
I mean, I can't understand why this guys are loosing their time building another sad IDE instead of getting mono integrated with Eclipse. I'd drop all MonoDevelop efforts tomorrow morning and put the team to work on the debugger, and then get it integrated with Eclipse or SlickEdit.
You can actually build everything from SlickEdit but you can't debug.
Mono, as a platform, is great but:
They must forget MonoDevelop
They must have a proper debugger *everywhere*
They must go *truly* multi-platform: I mean, official releases for all the linux distros plus MacOS X, and the BSDs, Solaris, HP-UX, AIX and so on. While they *stick* to Linux... they're dead.
Everything else will come...
Re:The thing is... (Score:5, Interesting)
I used to code Java in the JDK 1.1 - 1.4 days. It sucked ass. It was slow, had weird dependencies on X11, required a lot of boilerplate code (such that up to half of the LOC could be logging inside exception blocks), had various JRE incompatibilities all over the place (such that some applications just couldn't be run bug-free on ALL of AIX/Solaris/Windows/Linux/Mac), and the reference JDK/JRE was Sun's proprietary property. I left Java and went on to C, Perl, C++, and Lisp. Naturally I used Emacs and SLIME.
Then I found Clojure. And I got a $350 laptop from Walmart last week that had 3 gigs of RAM and a single-core 1.8GHz AMD processor. And I thought, "I wonder if Eclipse will run decent on this thing?" And it does, and it's not all that slow, and it is by far the best IDE I've ever used.
I'm now re-climbing the learning curve on modern Java, and it's looking pretty good now. AspectJ does a good job eliminating a lot of repetitive code, eclipse-metrics warns me when I'm not being decent at OOP design, and the available libraries are top-notch. Java the language isn't so bad anymore, and now with Clojure on top I have plenty of linguistic room to prototype and get to choose the best among many paradigms for each situation.
Give Java a fresh look, it's come a long way.
Re:What's the point with Qt now fully free? (Score:4, Interesting)
Amazing what difference a license change makes, eh ?
Re:But the political reasons... (Score:3, Interesting)
Not necessarily. Intel seems to do pretty well with x86-64, better than AMD even. Apple has finally brought Unix(tm) to the masses. Fujitsu's processors are better than Sun's.
Re:But the political reasons... (Score:3, Interesting)
While that is true it actually rarely matters in this case. How many stable dotnet applications actually require dotnet 3.5? The commercial application I'm using on both MS Windows and linux is using version 1 (and development is continuing with that version), and there's a whole lot of stuff using version 2 - it seems every time you install a dotnet program on MS Windows you need to download all the libraries from Microsoft again. So long as mono doesn't lag by two versions it will still support a lot of this dotnet stuff.
While I haven't seen anything really in the way of fast or even stable dotnet applications on any platform and think even the name is incredibly wanky (can't use it in a readable sentence) and most code of no better than shareware quality - I do think it is a good thing to support this new VB environment with pretensions of C and java on multiple platforms.
Accepting Mono is NOT giving up and giving in to Microsoft vendor lock-in - you can run software developed in this enviroment on other platforms so avoid Microsoft vendor lock-in.
Re:Objective Review (Score:2, Interesting)
I find it interesting how much anti-webapp rhetoric has shown up as of late. It means that sophisticated webapps are on the verge of becoming the next de facto technology. "First they ignore you. Then they fight you." I'll leave you to figure out what happens next.
Thanks man! Keep up the good work!
Re:healthy distrust (Score:3, Interesting)
There's one vital difference. Samba is not adhering to a standard that has been ratified by the ECMA that they can use to say "Ahhhhh, you're infringing!" and where the ECMA has not ruled that it will be RAND licensed forever whatever happens. Mono adheres to something quantifiable that can be pinned on them. Samba doesn't. It is something reverse engineered and where much of the implementation is partially completed and even incompatible with what Microsoft does. You could say that Samba just happens to be Windows Networking compatible in some ways. If you look at any of the patent applications that Microsoft has filed you will generally see something like "This adheres to code running in a CLR........" with a definition of what the CLR is somewhere in the text and a referral to the ECMA. You can't nail Samba down like that.
.Net is what Microsoft is telling people to use to develop applications for Windows, applications are what keep Windows where it is and Windows is where Microsoft's bread and butter is. Don't underestimate their willingness to protect that.
Re:clone or unique, but not both (Score:2, Interesting)
Re:Mono is .NET (Score:3, Interesting)
Yep that's right. I remember back in the day, I worked at a Microsoft "Very Valued" Partner which basically meant they could screw you over whenever they wanted just because and if you didn't comply, no more partner for you (*yank MSDN library and cheap licensing*). I almost brought the company to the brink of losing their partnership just by recommending Linux machines (which we did sell but apparently not advertised) to users that needed a xAMP stack.
Either way, the company had to build their own ERP (simply because they wanted to - this was the .com era) and Microsoft somehow got involved which trumpeted Visual Basic everywhere, the early releases of .NET were not stable yet and according to them .NET was not yet ready for business. Of course, if you ever did VB you would know that once you're done, the code looks like crap, the thing is as slow as hell and crashed every so often. 2 months later, the same people told us that it would be better to do it in .NET (which was then just out as 1.0) and it would all be very well. If you ever worked with VB.NET back then you would know that it was just as buggy as classic VB although 'neater' to code for and the functionality was all well so the company folded to the MS pressure and rebuilt it. All fine and dandy until MS came out with C#. All of a sudden early .NET (preview and 1.0) became largely deprecated (what, this was supposed to be a stable and EXTENDIBLE language!) and with it also VB was heavily modified to fit the 1.1 framework and so there was another rebuild of some code. Of course some parts were still in classic VB and old .NET code and in the mean time new programmers had come and external things had started to happen so just rebuilding pieces of the application started breaking. Also major security holes were found, were patched and broke stuff and thus the application had to be rewritten from the ground up (again) in C# for .NET to fit the thing.
I left the company later after that. In the mean time 2.0 and 3.0 came out. I wonder if they have been rewriting stuff for almost a decade now. Maybe if they just stuck to PHP and C++. I read a good (older) book recently on C++ and it said a good extensible program (algorithm) in an OO language is written twice: once to know what classes you are going to need to build, second time to apply what you got to know in well documented code, classes and libraries. I believe writing something more than 3 times just because the libraries you got to rely on changed is just bad practice.
Re:precisely the problem (Score:1, Interesting)
I've yet to have a problem running a mono app on FreeBSD (ok, it's not Linux, but typically compatibility is worse on FreeBSD).
Install x11-toolkits/libgdi+ (or whatever it was), and mono, and I've had graphics, networking, threading, pretty much whatever I've thrown at it, available in FreeBSD.
Scala (Score:3, Interesting)
Re:Objective Review (Score:3, Interesting)
> Many C and C++ programmers aren't so good at doing shell scripts. *shrug*
But I can assure you that most shell "scripters" aren't good for C, and even less for any non trivial programming project.
That's one of the top reasons IDEs are popular: they hide the build process from the developer.
I just don't understand why people defends autoconf as "the right tool(tm)". Maybe it's right for your case or knowledge, but most users (and developers) find it really annoying, yet is the sanctified open-source-build-tool standard... just because lack of decent alternatives.
Re:Objective Review (Score:1, Interesting)
I just don't understand why people defends autoconf as "the right tool(tm)". Maybe it's right for your case or knowledge, but most users (and developers) find it really annoying, yet is the sanctified open-source-build-tool standard... just because lack of decent alternatives.
Not true, most developers I know (and I know plenty on linux!) loathe autotools, in fact someone liking autotools is enough reason to mock him and treat him like a loonie. New projects usually go with CMake. Many old projects are switching to it, like KDE.
Re:Objective Review (Score:1, Interesting)
Re:But the political reasons... (Score:3, Interesting)
.NET CF has forms support, but it's not WinForms as we know them. AFAIK the libraries are managed by different teams and while they present some of the same interfaces they're entirely different under the hood. Much smaller, too.
I didn't know about .NET Micro, but looking at the Wikipedia page makes me start involuntarily drooling, if this is true...
Re:But the political reasons... (Score:3, Interesting)
(.NET/Mono is already pretty awesome even for projects that require native libraries, though. Package libfoo.dll and libfoo.so in with your executable and assemblies, and it will intelligently grab the right one on Linux/Windows. Not so easy on BSD, Solaris, or OS X, but those really aren't primary platforms for that particular effort.)
Actually, it is just as easy on MacOS X. My company's SDK is .NET based, and we provide a "Mono version" for both OS X and Linux. Of course, the .NET assemblies themselves are the same, but there's just a couple of parts that are platform specific, which here we'll just call "libfoo". I can package in "libfoo.dll", "libfoo.so" and "libfoo.dylib", and it happily works on Windows, Linux and OS X. Not sure about the BSD and Solaris worlds though. Worst case scenario, it should be trivial enough to check for the native library and then manually call the right one, as long as DLLImport can handle it. I haven't stepped off the Windows/Linux/MacOS line yet when it comes to .NET, so I haven't really looked in to what's possible and what isn't.
Re:Objective Review (Score:3, Interesting)
Just sort of have to throw my two cents on to this as well and agree whole-heartedly. I FAR prefer native GUI front ends on whatever platform to a web app. For most big web apps, they tend to be a basic form or two with "complex" stuff being called behind the scenes, in which case all of the real programming work went in to that "behind the scenes" stuff and that's not what I'd call a web app either.
I'm actually even considering going to the trouble of coding my own front ends for some internal web apps at work just because it's SO much faster to have a native client making calls to the back end DB than it is to wait for a buggy IE only webapp to do its thing and spawn a thousand or so annoying windows (especially ones that open with some text along the lines of "please wait", after which it opens another window, and closes the first, at which point IE pops up a requester warning me that the page is trying to close the window and do I want to allow it or not).
Re:healthy distrust (Score:3, Interesting)
Ballmer has said Linux violates 235 patents. That is using their patents.
No, they are never going to identify any patents or do a real court case, unless there is a significant change in the environment (like they are actually threatened by a competitor using Linux). The threat is much more effective.
They refuse to ever identify any patent directly. The reason they do this is that the patent may be shot down for prior art, or that it will be shown that Linux does not violate it, or that it will take about 1 day to patch Linux to work around the patent. If this happens for even one of those 235 patents it will greatly dimish the threat perceived by pointy-haired bosses.
Re:Objective Review (Score:3, Interesting)
as unpopular as it is to say so, I have to agree, I personally haven't found a single useful webapp for what I normally do, I see the usefulness of intranet server based apps, but webapps are generally useless, slow, buggy and problematic
Either what you do normally is pretty intensive of resources (not webapps strong point to date) or you're not doing a lot of work across multiple organizations. Trying to deploy something in a highly distributed and heterogenous setting is an utter nightmare, and using a webapp instead turns it into just plain sucky.
Re:Objective Review (Score:3, Interesting)
We are in such a case.
We are a multi-national, and in one research site tooling was developed for building, first all Win32 C++, later .NET C#.
Recently came the day that we got in bed with people who only work on Linux, so all our tools must now run on Linux. There is NO gui dependency here. In this case, Mono was really our rescue.
Re:Objective Review (Score:2, Interesting)
then you shouldn't be using Linux either (Score:3, Interesting)
A third party implementation of a standard defined by the first-party implementor is always going to lag behind the original.
Linux, GNU, GNU C++, libc, and many other tools that we take for granted all started out as clones of proprietary software from a litigious, monopolistic company. So did many other open source projects. If people had followed your reasoning, free software and open source software wouldn't exist. .. What matters is that when developers and end users pick a technology, they pick the leader, not the follower. Accepting Mono is giving up and giving in to Microsoft vendor lock-in.
Lucky for us, then, that Mono is not a follower. There are dozens of Mono-based desktop applications, and they are not based on .NET, but instead on Gtk# and other Linux technologies.
The situation with C# is really not much different from C++. There is an open language standard, there is a large set of Linux-native libraries, and there is a large set of Microsoft APIs. For C++ on Linux, you get GNU C++, Gtk+ and other FOSS libraries, plus WINE if you want it. For C# on Linux, you get the Mono C# compiler, Gtk# and lots of other FOSS C# libraries and bindings, plus a separate set of .NET libraries if you want them. Mono's .NET libraries seem to be used about as much as WINE, which is to say, not a whole lot. The FOSS C# libraries are just a lot better.
Re:Objective Review (Score:2, Interesting)
Re:Objective Review (Score:3, Interesting)
Crap like Google Maps I find insulting. We had map software 15 years ago. The only thing we needed was periodic udpates, but web apps go to a completely opposite extreme, every single data request is serviced live, nobody finds it acceptable to risk that data might be hours or days old. If I'm actually using my connection to retrieve data (files, audio, video, etc.) all those web-apps slow to a crawl.
Well them, buy your map software or even a dedicated GPS device. Personally I'm happy with google maps. It also makes a nice backup to my gps that I have it installed on my blackberry. If I had enough of a need I'd buy mapping software.
Youtube is an example of a great webapp. For most video needs it works great. There are times I want better than a small flv file, and I will watch a dvd or mpeg 4 in those cases.
Here's my deal... (Score:3, Interesting)
I've given Mono a fair bit of consideration. I should give it more, because I feel like it's important to evaluate carefully these technologies that might be useful to me...
For interoperability I think Mono is a good thing, end of story.
But I don't think it's a technology that I want as an integral part of my Linux system. There is the whole "Microsoft may decide to be bitches" thing, which kind of puts me in the position of liking what Mono can offer but not wanting to rely on it too much... But there's also the fact that it's kind of a little bit alien. I want a Unix system based around Unix concepts... Mono has at least a few Windows-isms to it (.exe extensions, for instance) but I don't know if it goes deeper than that. That's OK for a package I'm installing to gain the ability to run .NET software, but it's not something I want in a package that I'm going to use to build new applications and such for Linux. I wouldn't want to treat Mono as a core part of a Linux system if it has too much Windows flavor to it. It's just not the kind of system that I want. I like the technology (especially now that Mono supports native compilation) but I feel like it's also important for the technology to fit the system where it's to be used.
I do feel there's a need for some of the things Mono has to offer: I love what .NET does in terms of promoting cooperation between different programming languages (the various .NET versions of scripting languages and other programming languages, built to take advantage of the CLR, and so on) - it should be easy to pass objects from one programming language to another, and it should be easy to bind existing libraries to different programming languages - and .NET does that. Scripting languages benefit from having a good, efficient intermediate bytecode form - that is also provided by .NET. This is why, despite being a bit uneasy with Mono (partly, I'll freely admit, due to long-standing prejudice and grudge regarding Microsoft) it's still appealing to me.