Tag Archives: mono

C# vs C++ and .NET vs Mono vs Compact Framework performance tests

A detailed benchmark posted on codeproject investigates the performance of basic operations including string handling, hash tables, math generics, simple arithmetic, sorting, file scanning and (for C#) platform invoke of native code. These are the conclusions:

  • There is only a small performance penalty for C# on the desktop versus C++.
  • Mono is generally slower than Microsoft .NET but still acceptable, and all the benchmarks ran without modification.
  • The Compact Framework, an implementation of .NET for mobile devices, performs poorly.

My observations: this matches my own experiments. Why then do some .NET applications still perform badly? When Evernote switched its application from .NET to native code it got much better performance.

The main reason is a couple of issues that this kind of benchmark hides. One is the GUI layer, which involves a ton of platform invoke code under the covers. Another is the large size of .NET applications because of the runtime and library overhead; a lot more stuff gets loaded into memory.

One thing to like about Silverlight is that it is truly optimized for client programming and load time tends to be faster than for a desktop .NET application.

Note that for mobile these benchmarks suggest that C++ still has a big advantage. It would be interesting to see them applied to Silverlight apps on Windows Phone 7. As I understand it, the Silverlight .NET runtime in Windows Phone 7 shares code with the Compact Framework on Windows Mobile, so it is possible that the poor results for the Compact Framework would also apply to Silverlight on Windows Phone 7. Unfortunately developers do not have the option for C++ on Windows Phone 7.

Mono splits from Novell/Attachmate to form basis of new company

Mono is an open source implementation of .NET, formerly sponsored by Novell, and its future following Novell’s acquisition by Attachmate has been the subject of speculation.

Today Mono leader Miguel de Icaza has revealed new plans. In a blog post, he announces Xamarin, a new company focused on Mono. This company will build new commercial .NET tools for Apple iOS and Google Android, to enable .NET development on those platforms. Note that they will not be called MonoTouch and MonoDroid, the Novell offerings for this, but will be “source compatible”. I am sure there are brand and intellectual property ownership issues here; but de Icaza is no stranger to negotiating tricky issues of this kind, bearing in mind Mono’s relationship with Microsoft .NET. However I am not sure why the new company cannot acquire the existing brands, since it seems that Attachmate will no longer be able to support them.

The plans are not exactly new, but have been forced by Attachmate’s decision to lay off the entire Mono team:

We have been trying to spin Mono off from Novell for more than a year now. Everyone agreed that Mono would have a brighter future as an independent company, so a plan was prepared last year.

To make a long story short, the plan to spin off was not executed. Instead on Monday May 2nd, the Canadian and American teams were laid off; Europe, Brazil and Japan followed a few days later. These layoffs included all the MonoTouch and MonoDroid engineers and other key Mono developers.

Apparently Xamarin has “angel funding” but is looking for more.

The advent of MonoTouch and MonoDroid has been good for Mono, since it gives the project a stronger business model than it had previously. These mobile platforms are hot, and the ability to code for them in C# is great for Microsoft Platform developers. This factor could enable Xamarin to succeed.

On the other hand, Novell’s name gave Mono enterprise credibility as well as the backing of a large company, and these it now lacks.

The curious thing is that Mono is valuable to Microsoft. The company seems at times to hate Mono, because it removes the need for Windows, and at other times to love it, because it extends the breadth of .NET to include Linux and now iOS and Android. Microsoft gave some sort of official status to Moonlight, the Mono implementation of Silverlight, though the company’s support for Moonlight has always seemed hesitant.

So can we expect now that the company which can afford $8.5 billion for Skype, could expend a few million in support of Xamarin? Or will it stand by and hope that Mono fades away?

I have no idea, though I would like to see both Mono and Xamarin succeed. It is no threat to Microsoft, but does take .NET to places that Microsoft will never support. Without Mono, C# is merely a language for programming Windows.

Adobe AIR 2.6, MonoMac 1.0, cross-platform is not dead yet

It is a busy time for cross-platform toolkits. Adobe has released AIR 2.6, and reading the list of what’s new you would think it was mainly for mobile, since the notes focus on new features for Apple iOS, though AIR is also a runtime for Windows, Linux and desktop Mac. New features for iOS include GPU rendering – a form of hardware accelerated graphics – access to the camera, microphone, and camera roll, and embedded Webkit for apps that use web content. On Google Android, you can now debug on devices connected via USB.

There is also a new feature called “owned native windows” which lets you have a group of windows that remain together in the Z order – this lets you have things like floating toolbars without odd results where toolbars get hidden underneath other applications.

Asynchronous decoding of bitmaps is another new feature, allowing images to be processed in the background. This seems like a stopgap solution to overcome the lack of mullti-threading in AIR, but useful nonetheless.

Since the Flash runtime does not run on iOS, Adobe has a packager that compiles an AIR application into a native app. This is now called the AIR Developer Tool or ADT. You can use the ADT to target Windows, Linux or Android as well; however platforms other than iOS still need the AIR runtime installed.

Adobe is dropping support for the original iPhone and the iPhone 3G. iPhone 3GS or higher is needed.

If you want to build a cross-platform app but prefer .NET to Adobe’s Flash and ActionScript, the Mono folk have what you need. I’d guess that the Mono team has a small fraction of the resources of Adobe; but nevertheless it has delivered MonoTouch for iOS and is working on MonoDroid for Android. Just completed in its 1.0 version is MonoMac, for building Cocoa applications on Apple OSX. Mono is not fully cross-platform, since the GUI framework is different on the various platforms, but you do get to use C# throughout.

I am happy to agree that true native code is usually a better solution for any one platform; but at a time when the number of viable platforms is increasing the attraction of cross-platform has never been greater.

Mono project: no plans for cross-platform WPF

Miguel de Icaza’s report from the Game Developer Conference is upbeat, rightly so in my view as usage of Mono is continuing to build, not only in game development with Unity, a development tool that uses Mono as its scripting engine, but also for mobile development for Apple’s iOS with Monotouch and for Android with Monodroid. These mobile toolkits also give Mono a stronger business model; many sites use Mono for serving ASP.NET applications on Linux, but without paying or contributing back to the project.

Mono is an open source implementation of C# and Microsoft’s .NET Framework.

That said, it is interesting that Mono is still struggling with an issue that has been a problem since its first days: how to implement Microsoft’s GUI (Graphical User Interface) framework on other platforms. Mono does have Gtk# for Windows, Mac and Linux, but this does not meet the goal of letting developers easily port their Visual Studio client projects to Mono. There is also an implementation of Windows.Forms, but de Icaza mentions that “our Windows.Forms is not actively developed.”

Apparently many tools vendors asked the Mono team at GDC when Windows Presentation Foundation (WPF) would be implemented for Mono. WPF is the current presentation framework for Microsoft.NET, though there is some uncertainty about where Microsoft intends to take it. I remember asking de Icaza about this back in 2003, when the WPF framework was first announced (then called Avalon); he said it was too complex and that he did not plan to implement it.

This is still the case:

We have no plans on building WPF. We just do not have the man power to build an implementation in any reasonable time-frame.

That said, Mono has implemented Silverlight, which is based on WPF, and there are some signs that Microsoft might merge WPF and Silverlight. What would the Mono team do then?

Miguel de Icaza says:

Silverlight runs on a sandbox, so you can not really P/Invoke into native libraries, or host DirectX/Win32 content inside of it.
There are other things missing, like menubar integration and things like that.

Of course, this is no longer true on Windows: Platform Invoke is coming in Silverlight 5.

Perhaps the Mono team will knuckle down and implement Silverlight with desktop integration, which would be good for cross-platform Silverlight and compatibility with Microsoft .NET.

Then again, it seems to me that Mono is increasingly divergent from Microsoft .NET, focusing on implementing C# in places that Microsoft does not touch, such as the mobile platforms from Apple and Google.

That is actually a sign of health; and you can understand why the Mono team may be reluctant to shadow Microsoft’s every move with Silverlight and WPF.

Trying out MonoTouch – C# for Apple’s iPhone and iPad

I’ve posted an article on trying out MonoTouch, which builds on the open source Mono project to provide a means of developing apps for Apple’s iOS using C# and the .NET Framework.

It is easy to assume that since the .NET Framework is Microsoft’s technology, using a non-Microsoft implementation is risky. Then again, Mono is open source; and the more usage it gets, the better it becomes. MonoTouch is an important development for the project, since it is a commercial project which might actually be making some money for Novell/Attachmate. While it would be nice to get it for free, it is important that Mono makes business sense as well. MonoTouch has given the Mono project a significant boost.

The top Silverlight feature request: implement on more platforms

One of the things mentioned by Microsoft VP Scott Guthrie in his Firestarter keynote yesterday was that Silverlight 5, the new version set for release in 2011, implements some 70% of what users have voted for. I presume he means the feedback forum here. But look what the top request is – as noted by a comment to yesterday’s post:

image

Looking at the comments, Android is a common request, and relatively easy for Microsoft to achieve given the open nature of that platform.

This was apparently not part of the 70% though. Instead, Guthrie introduced more Windows-only features – showing that concerns about divergence between Windows and Mac implementations when Microsoft announced COM support at the 2009 PDC were justified.

What if Microsoft had purchased Novell, or purchased Mono from Novell, instead of letting it go to Attachmate? It would have enabled Microsoft to unify the Windows and Linux implementations as well as building on the work the Mono team has done on compilation for iOS.

That dream is over though; the Silverlight application strategy seems focused on making it better for Windows-platform corporations.

The Java crisis and what it means for developers

What is happening with the Java language and runtime? Since Java passed into the hands of Oracle, following its acquisition of Sun, there has been a succession of bad news. To recap:

  • The JavaOne conference in September 2010 was held in the shadow of Oracle OpenWorld making it a less significant event than in previous years.
  • Oracle is suing Google, claiming that Java as used in the Android SDK breaches its copyright.
  • IBM has abandoned the Apache open source Harmony project and is committing to the Oracle-supported Open JDK. Although IBM’s Sutor claims that this move will “help unify open source Java efforts”, it seems to have been done without consultation with Apache and is as much divisive as unifying.
  • Apple is deprecating Java and ceasing to develop a Mac-specific JVM. This should be seen in context. Apple is averse to runtimes of any kind – note its war against Adobe Flash – and seems to look forward to a day when all or most applications delivered to Apple devices come via the Apple-curated and taxed app store. In mitigation, Apple is cooperating with the OpenJDK and OpenJDK for Mac OS X has been announced.
  • Apache has written a strongly-worded blog post claiming that Oracle is “violating their contractual obligation as set forth under the rules of the JCP”, where JCP is the Java Community Process, a multi-vendor group responsible for the Java specification but in which Oracle/Sun has special powers of veto. Apache’s complaint is that Oracle stymies the progress of Harmony by refusing to supply the test kit for Java (TCK) under a free software license. Without the test kit, Harmony’s Java conformance cannot be officially verified.
  • The JCP has been unhappy with Oracle’s handling of Java for some time. Many members disagree with the Google litigation and feel that Oracle has not communicated well with the JCP. JCP member Doug Lea stood down, claiming that “the JCP is no longer a credible specification and standards body”. Another member, Stephen Colebourne, has a series of blog posts in which he discusses the great war of Java and what he calls the “unravelling of the JCP”, and recently  expressed his view that Oracle was trying to manipulate the recent JCP elections.

To set this bad news in context, Java was not really in a good way even before the acquisition. While Sun was more friendly towards open source and collaboration, the JCP has long been perceived as too slow to evolve Java, and unrepresentative of the wider Java community. Further, Java’s pre-eminence as a pervasive cross-platform runtime has been reduced. As a browser plug-in it has fallen behind Adobe Flash, the JavaFX initiative failed to win wide developer support, and on mobile it has also lost ground. Java’s advance as a language has been too slow to keep up with Microsoft’s C#.

There are a couple of ways to look at this.

One is to argue that bad news followed by more bad news means Java will become a kind of COBOL, widely used forever but not at the cutting edge of anything.

The other is to argue that since Java was already falling behind, radical change to the way it is managed may actually improve matters.

Mike Milinkovich at the Eclipse Foundation takes a pragmatic view in a recent post. He concedes that Oracle has no idea how to communicate with the Java community, and that the JCP is not vendor-neutral, but says that Java can nevertheless flourish:

I believe that many people are confusing the JCP’s vendor neutrality with its effectiveness as a specifications organization. The JCP has never and will never be a vendor-neutral organization (a la Apache and Eclipse), and anyone who thought it so was fooling themselves. But it has been effective, and I believe that it will be effective again.

It seems to me Java will be managed differently after it emerges from its crisis, and that on the scale between “open” and “proprietary” it will have moved towards proprietary but not in a way that destroys the basic Java proposition of a free development kit and runtime. It is also possible, even likely, that Java language and technology will advance more rapidly than before.

For developers wondering what will happen to Java at a technical level, the best guide currently is still the JDK Roadmap, published in September. Some of its key points:

  • The open source Open JDK is the basis for the Oracle JDK.
  • The Oracle JDK and Java Runtime Environment (JRE) will continue to be available as free downloads, with no changes to the existing licensing models.
  • New features proposed for JDK 7 include better support for dynamic languages and concurrent programming. JDK 8 will get Lambda expression.

While I cannot predict the outcome of Oracle vs Google or even Apache vs Oracle, my guess is that there will be a settlement and that Android’s momentum will not be disrupted.

That said, there is little evidence that Oracle has the vision that Sun once had, to make Java truly pervasive and a defence against lock-in to proprietary operating systems. Microsoft seems to have lost that vision for .NET and Silverlight as well – though the Mono folk have it. Adobe still has it for Flash, though like Oracle it seems if anything to be retreating from open source.

There is therefore some sense in which the problems facing Java (and Silverlight) are good for .NET, for Mono and for Adobe. Nevertheless, 2010 has been a bad year for write once – run anywhere.

Update: Oracle has posted a statement saying:

The recently released statement by the ASF Board with regard to their participation in the JCP calling for EC members to vote against SE7 is a call for continued delay and stagnation of the past several years. We would encourage Apache to reconsider their position and work together with Oracle and the community at large to collectively move Java forward.  Oracle provides TCK licenses under fair, reasonable, and non-discriminatory terms consistent with its obligations under the JSPA.   Oracle believes that with EC approval to initiate the SE7 and SE8 JSRs, the Java community can get on with the important work of driving forward Java SE and other standards in open, transparent, consensus-driven expert groups.   This is the priority.   Now is the time for positive action.  Now is the time to move Java forward.

to which Apache replies succinctly:

The ball is in your court. Honor the agreement.

Measuring start-up time for .NET, Java, C++

A comment here points me to this comparison by Decebal Mihailescu of start-up times for processes on Windows using different runtimes: .NET in several versions, Java 1.6, Mono 2.6.4, and Visual C++ 2010 (native code).

image

It is notable that native code is much faster than the runtimes, and that .NET is ahead of Java – to be expected, perhaps, since Windows is its native environment.

The test application is tiny and when we are talking about a time of less than a tenth of a second, users are unlikely to care much. I would like to see tests based on a larger application; on the other hand it becomes increasingly difficult to make a fair comparison.

It’s a shame Embarcadero Delphi was not included, though I’d expect similar figures to those for Visual C++.

Start-up time is critical to perceived performance. One reason I use an ancient version of Paint Shop Pro for simple image manipulation is that I can double-click an image and be editing it instantly, whereas alternatives like Adobe PhotoShop or the free paint.net make me wait. Unless I need the extra features these provide, I stick with fast and simple.

Microsoft – make up your mind about Moonlight

I’ve been trying out Microsoft’s Office Web Apps, as provided for the release version of SharePoint 2010. The cross platform story is uneven, whether across Mac/Windows/Linux, or across different browsers, or even across different versions of Windows and Office. So far it does mostly work though, even if there are problems with certain tasks like printing or opening an online document in a desktop application.

The biggest problem I’ve had is on Linux. Supposedly Firefox 3.5 on Linux is supported. I ran up Ubuntu and Firefox 3.5, and went to look at a document in Word Web App. When I selected the document, Firefox quit. Every time.

After checking that Firefox was up-to-date it occurred to me that the problem might be related to Moonlight, the Linux implementation of Silverlight done by the Mono team. I disabled it. Suddenly, everything worked, even Edit in browser.

Moonlight is not just an open source project like the original Mono. It has a certain amount of official blessing from Microsoft. Here’s what VP Scott Guthrie said back in September 2007:

Over the last few months we’ve been working to enable Silverlight support on Linux, and today we are announcing a formal partnership with Novell to provide a great Silverlight implementation for Linux.  Microsoft will be delivering Silverlight Media Codecs for Linux, and Novell will be building a 100% compatible Silverlight runtime implementation called “Moonlight”.

Moonlight will run on all Linux distributions, and support FireFox, Konqueror, and Opera browsers.  Moonlight will support both the JavaScript programming model available in Silverlight 1.0, as well as the full .NET programming model we will enable in Silverlight 1.1.

You would think therefore that Microsoft would test the Firefox/Linux/Moonlight combination with its shiny new Office Web Apps. Apparently not. Here’s what the user experience is like for Office Word App. I figured that the solution might be to upgrade Moonlight to the latest version, so I did, installing what is now called Novell Moonlight 2.2. I went back to Word Web App. Firefox no longer crashes, but I now get a blank area where the Word document should be shown, and an error if I resize the browser window:

Now let’s see what happens if I disable Moonlight:

Everything is fine – except now there is a banner inviting me to “Improve my experience” by installing Silverlight. If I follow the link I eventually get back to the same Moonlight install that I have just enabled, which would actually break rather than improve Word Web App.

It is obvious that if users have to disable Moonlight to work with Office Web Apps, this will not help Moonlight adoption on Linux.

Office Web Apps are new and I’d hope this is something that Microsoft, Novell and the Mono team can soon fix between them. One reason for highlighting it now is the hope that something could be done before the full roll-out of Office and SharePoint 2010 on May 12th.

The real point though is what this says about the extent to which Microsoft cares about Moonlight and Linux users, and how much or little communication takes place between Microsoft and Novell. Silverlight isn’t required for Office Web Apps – as you can see from the above – but it is used to good effect where available, and this Office release is therefore an important release for Silverlight as well.

Microsoft should make up its mind. Is Novell really a trusted partner for Silverlight on Linux? Or a third-party product that has to take its chances?

Mono Tools for Visual Studio: code on Windows, run on Linux

I have just com across Mono Tools, a Novell add-in for Visual Studio that lets you test Mono compatibility. It adds a Mono menu which has options to run locally or remotely in Mono, analyze for compatibility issues, and create deployment packages. No sign of Mac support, which is a missed opportunity, but understandable given that Novell owns SUSE Linux.

For those few still unfamiliar with Mono, it is an open source implementation of Microsoft’s .NET Framework, enabling your .NET applications to run on other platforms. One compelling use is to have your ASP.NET web applications run on the free Apache web server, rather than Microsoft’s IIS.

image

Mono Tools works with both Windows Forms and web projects.

image

This is just the sort of thing Mono needs to move it further into the mainstream, though another less welcome sign of business acceptance is that this is a commercial product, currently costing $99.00 for an individual or $249.00 per seat in an organization. There is also an Ultimate edition at $2,499, which comes with a commercial non-LGPL license to redistribute Mono.

The Mono Tools team is now looking for testers for its 1.1 edition, which supports Visual Studio 2010.