Category Archives: development

development, microsoft, mono, software development

What does Xamarin’s success say about open source versus proprietary? Miguel de Icaza says he has never been happier

2 Comments

image

Yesterday Xamarin, which offers tools for targeting iOS, Android and Mac with C#, announced a partnership with Microsoft, an announcement which I wrote up on The Register. It drew a few comments, several complaining about the cost:

So it cost more then Visual Studio Pro.

And that is for 1 target platform?

or

Not so useful for little indie developers at those prices.

or

From open source to $999 per developer per year. Monetising Mono seems to have worked, so perhaps PCL being open sourced won’t be such a bargain either.

If you check Xamarin’s pricing you will see that the tools are not cheap for casual users; of course, if you are selling thousands of apps or developing corporate apps at normal rates the tools soon pay for themselves.

Xamarin is doing well as far as I am aware; CEO Nat Friedman told me of rapid growth in the number of customers and I have seen for myself the high interest in the tools at events like Microsoft BUILD earlier this year in San Francisco.

This gives me pause for reflection. What does the success of Xamarin, and the relative lack of success of Mono (the open source C# compiler and .NET Framework on which Xamarin is based) say about how well the open source business model works in the real world?

I was reminded of a conversation I had with Miguel de Icaza, creator of Mono and co-founder of Xamarin, Friedman back in February of this year, when Xamarin 2.0 was launched. I asked de Icaza if the new company publishes the source code for all its products?

“No. Our company does proprietary tools for iOS and Android apps. The entire iOS and Android support is proprietary as well as our commercial Mac support. All those three pieces are proprietary while the IDE and the Mono runtime are open source. Whether the code is open source or not depends on whether it is part of core Mono or core MonoDevelop. Otherwise it tends to end up as proprietary.”

Friedman added: “Mono has a thriving open source community around it, and Xamarin has a thriving community of developers who are building commercial mobile apps. We have 12,000 customers, many of them have never heard of Mono. They came to us because they had a problem to solve, they were C# developers and they wanted to get an iOS or Android app built. We solved that problem and that was worth money to them. The reason we have a business is that Microsoft developers do pay for tools, unlike Web developers for example. It’s been a great market for us. It allows us to invest.”

I asked de Icaza if he gets any grief from the open source community for having proprietary code in his company.

“Actually no. We started doing the proprietary bit at Novell. In fact we’ve been doing proprietary for a long time, even before we were acquired by Novell, at Ximian. We didn’t get a lot of grief from people. I can tell you though that when I was working in the Linux world, they were very stressful days for me, because people constantly complain about a “secret conspiracy” and that thing just went out of control. There are some advocates in the Linux world that don’t like anything that has the label Microsoft on.

“Ever since we did Xamarin which meant we focused on Mac and Windows, all that stress is gone, I don’t think I have ever been happier. In the past I was enduring this constant barrage of senseless attacks, and now I never hear about this.

“One thing that happened in the Linux world is that I was very proud of the four or five big apps that were built with Mono. F-spot that we built, Banshee, and a couple of others. Now with Xamarin I can’t keep track of them any more because they are measured in the thousands. There are thousands of very large apps, over a millions lines of code, that people send us. It’s a very different world, it’s just so much larger than all the work we did in Linux days back then.”

development, microsoft, professional, visual studio, windows

Visual Studio 2013 is released. What’s new?

Microsoft released Visual Studio 2013 yesterday:

VS 2013 can be installed side by side with previous versions of Visual Studio or, if you have a VS 2013 pre-release, it can be installed straight over  top of the pre-release.

I installed over the top of the pre-release and I’m happy to say that this worked without incident. This is how it should be.

image

Oddly, the launch of Visual Studio 2013 is not until November 13th, proving that in Microsoft’s world products can “launch” before, at or after general release.

So what’s new in Visual Studio 2013? Tracking Visual Studio is difficult, because many important features show up as updates and add-ons. After all, at heart Visual Studio is just a shell or platform in which development sit. The Visual Studio LightSwitch HTML client, for example, which made LightSwitch into a strong tool for rapid application development of mobile web apps, appeared as part of Visual Studio 2012 Update 2. Now in Visual Studio 2013 we have LightSwitch support for Cloud Business Apps, though the new project type is shown under Office/SharePoint rather than under LightSwitch:

image

A Cloud Business App is an add-on for SharePoint typically running on Office 365. In the new model SharePoint apps do not really run on SharePoint, but are web apps that integrate with SharePoint. This is great in an Office 365 context, since you can write a web app that is accessible through the Office 365 site and which is aware of the logged-on user; in other words, it uses Azure Active Directory automatically. There’s more on the subject here.

What else is new? Here are some highlights:

  • Better ISO C/C++ compliance in Visual C++
  • Upgraded F# with language tweaks and improved performance
  • .NET Framework 4.5.1 with minor enhancements
  • Support for new Windows 8.1 controls and APIs in Windows Store apps – these are extensive.
  • “Just my code” debugging for C++ and JavaScript, and Edit and Continue for 64-bit .NET apps
  • Graphics diagnostics for apps running remotely
  • Sign into Visual Studio with a Microsoft account. Microsoft pulls developers further into its cloud platform.
  • Windows Azure Mobile Services – build a back end for an app running on Windows, Windows Phone, iOS, Android or web

Does that amount to much? Compared to the changes between Visual Studio 2010 and 2012, no. That is a good thing, since what we have is a refinement of what was already a capable tool, rather than something which gives developers a headache learning new ways to work.

.net, development, microsoft, professional, software development, windows, windows 7

ComponentOne’s TouchToolkit for Windows Forms: another approach to the Windows tablet problem

1 Comment

Software component vendor ComponentOne has released Studio Enterprise 2013 v2.5, the latest in its suite of components, with support for Windows 8.1 and Visual Studio 2013.

The piece that caught my eye is the TouchToolkit for Windows Forms.

image

Here’s the problem. The Windows desktop is poor with touch control, which is why Microsoft created Windows 8 with its alternate, touch-friendly Windows Runtime platform. However users are resistant to the changed user interface, and it does not help with existing desktop apps.

Developers are also faced with a question of simple mathematics. Develop a Windows 8 Store app, get a market of x. Develop a Windows desktop app, get a market of many times x, since Windows 8 can run desktop apps, but Windows 7 cannot run Store apps.

Embarcadero approached this problem with a framework called Metropolis, for Delphi and RAD Studio. It builds apps that mimic the Windows Runtime look and feel, but which are actually desktop apps. Of course they do not run on Windows RT, the ARM version. It is a confusing solution in my opinion, leading users into what Martin Fowler calls the Uncanny Valley, where stuff works almost but not quite how you expect.

I prefer the thinking behind the TouchToolkit. Take your existing Windows Forms apps, or write a new one, using these controls to make them more touch-friendly. They will never be as well suited to touch control as a Store app, but they might be good enough, and of course will run on Windows 7 and earlier versions.

The controls include a magnifier, support for zoom gestures, and a touch event provider that adds gesture support to any control.

Windows Forms, we all know, is not as good as WPF if you want an application that scales nicely and supports modern design. On the other hand, Windows Forms is pragmatic and easy to use framework that remains popular for line of business apps.

delphi, development, embarcadero, mobile

Embarcadero RAD Studio XE5 (Delphi) for Android now available

3 Comments

Today Embarcadero released RAD Studio XE5 which lets you build apps for Windows, Mac, iOS and Android. You can also buy Delphi XE5 separately if you prefer.

Embarcadero’s release cycle is relatively rapid. It was only six months ago that RAD Studio XE4 with iOS support appeared.

The big deal in this release is Android support. If you use the FireMonkey framework, you can build apps for all supported platforms.

There is also a new REST client library and some other enhancements – see here for a list of what’s new.

Embarcadero’s approach to Android development is distinctive. In keeping with Delphi’s tradition of native code compilation, Android apps are compiled using the NDK (Native Development Kit). Embarcadero’s developer evangelist John Thomas told me that this delivers excellent performance. I can believe it, though note what Google says:

Before downloading the NDK, you should understand that the NDK will not benefit most apps. As a developer, you need to balance its benefits against its drawbacks. Notably, using native code on Android generally does not result in a noticable performance improvement, but it always increases your app complexity. In general, you should only use the NDK if it is essential to your app—never because you simply prefer to program in C/C++.

Delphi developers are largely shielded from the complexity of the NDK, since you code using the high level abstraction provided by the runtime library (RTL) and the FireMonkey framework. If that is all you need I should think everything will be fine. If you have a Java library you need to call from your Delphi Android application, you need to use JNI (Java Native Interface) which is not so much fun.

Another point to note is that FireMonkey emulates most visual controls like buttons and lists, by drawing them itself. Users might not notice, if they look and behave exactly like the native controls, but this is hard to do perfectly. Embarcadero’s approach is native in respect of the code it generates, but not in respect of the controls it uses.

I installed RAD Studio XE5 on a Windows 8 machine and set about building an Android app. I already had the Android SDK installed so I asked the RAD Studio installer to skip the SDK but to install the NDK. As it turned out, I am not sure whether it did or did not (I could not find it quickly), but it was easier to download the NDK manually.

I have previously tried Delphi for iOS, for which the usual approach is to run Delphi (or RAD Studio) in a Windows emulator on a Mac, since the Delphi IDE is Windows only. This approach is not so good for Android development because its hard to attach Android devices to an emulator for debugging. Therefore, a real Windows PC is a better platform for Android development. If you want to target iOS as well, you can still do so, by using the remote agent running on a Mac.

Setting up for Android development is a little harder than setting up iOS development. The Android device I used for my test is a Sony Xperia T, and I installed the Sony PC Companion to be sure of having the correct USB drivers for debugging. With Java, the SDK, the NDK, RAD Studio itself, and getting the device connected, that is a fair number of moving parts.

It worked though. I created an Android app, connected my Xperia, and it showed up as a target in Delphi (it is also called the LT30p).

image

I threw a label, a listbox and a button onto my app’s main screen.

image

As it turned out, I should have taken a little more trouble. Here is my app running on the phone:

image

Something has gone wrong with the list, but it looks easy to fix.

According to Embarcadero, a recent survey of over 1300 Windows developers showed that 85% get requests for mobile apps. But what mobile platform is most requested? Android is apparently at the top of the list:

  • 83% Android
  • 67% iOS
  • 33% Windows Phone
  • 17% Windows RT
  • 14% Blackberry

Is that really Windows RT (ARM) or could it include WinRT (Windows 8 Store app) I wonder? Neither are supported by Delphi yet; but at least with Android it now supports the most highly requested platform.

Cross-platform mobile development is critical today, and the new capabilities in Delphi and RAD Studio will be welcome. Is it the best approach? The trade-off is this. On the plus side, you get a cross-platform GUI framework that lets you share the maximum amount of code across all the targets you support. On the minus side, that might not be a good idea; see this post for some thoughts on that. You also get a native executable that should perform well, certainly better than an HTML/JavaScript approach, though I’m not convinced that using the NDK on Android is ideal.

How big is your Delphi Android app? Using the Hello World example above, this is what I got in debug configuration:

image

24.52MB storage. I changed to release configuration and got this:

image

That saved nearly 3MB, to 21.62MB.

Here is the RAM usage:

image

I would be interested in hearing from developers using Delphi or C++ Builder for Android development. How is the quality of this first release? Is the fact that you are not developing in Java a problem in practice?

cloud computing, development

Microsoft’s MSDN changes: too much marketing

5 Comments

Microsoft has updated its developer site, MSDN (Microsoft Developer Network). Simple, relevant and community driven, says Product Manager Brian Harry in a post introducing the new site.

image

Developers are not convinced that it is an improvement. The comments to Harry’s post are pretty damning, in fact. The core complaint is that the new MSDN home page is not a developer site, but a marketing site. Click on those “Use your skills” buttons and you get not technical references, but marketing pitches. They are misleading too. Click on “I build web apps” and you get a pitch for Windows Store apps build with HTML and JavaScript – and no mention of ASP.NET or Windows Azure.

image

What do developers want? Here’s a good suggestion:

Why don’t you have your team work on changes to MSDN that people would actually want? Like editable blog comments with formatting and code snippet options? Or Connect/UserVoice sites that aren’t horrific and disjointed? Or adding the TypeScript blog to the Dev Tools blog feed (something I’ve mentioned a half dozen times already to various people at MS, you’d think this 5 second update might make it into someone’s queue after TypeScript being out for almost a year now)?

I just cannot fathom how you guys are so adamant that listening to developers is important to you when 99% of feedback is ignored in lieu of changes that no one asked for and no one wants.

Most people come to MSDN in search of technical information. I suspect the strong marketing pitch for Windows Store apps will be counter-productive.

On the other hand, present developers with a fast, coherent, logically organised technical resource so that you can easily find both the API reference for the platform you are targeting, and comments and help with using it, and this will attract them and make them want to use the platform more often.

development, javascript, mobile, software development, Web

Appcelerator plans to rethink Titanium architecture, standardise on WebKit JavaScript engine

3 Comments

Appcelerator CEO Jeff Haynie has posted about his plans for Titanium, the company’s cross-platform mobile development toolkit.

The plan is to completely rewrite the core engine, while maintaining a mostly-compatible API. Central to the plans is the idea of using one JavaScript engine on all platforms:

With Ti.Next, we’ve created a small microkernel design that will allow us to have minimal bootstrap code in the native language (C, Java, C#, etc) that talks to a common set of compilers, tools and a single JavaScript Virtual Machine. We have found a way to make the WebKit KJS VM work on multiple platforms instead of using different VMs per platform. This means we can heavily optimize the microkernel (herein after called the “TiRuntime”) and maintenance, optimizations and profiling can be greatly simplified. We’re talking about ~5K LOC vs. 100K LOC per platform.

This will make it possible to share almost all the Titanium code itself across all platforms. The Titanium runtime itself will be shared code written in JavaScript.

Appcelerator says that Titanium code will be “faster than native code in most situations.”

No date for Ti.Next is given though according to this slidedeck the plan is to have the “first set of developer builds available soon to GitHub repo – possibly in the next 45-60 days”. It adds, “production builds are a ways away.”

Using a WebKit JavaScript engine on Windows Phone, for example, sounds interesting.

.net, development, microsoft, professional, software development, visual studio

Anders Hejlsberg says C# 6.0 to use Roslyn compiler, coming in next Visual Studio after VS 2013

A disappointment at Microsoft’s Build conference last month was lack of news about the next version of C#, version 6.0. C# architect Anders Hejlsberg did present a session, but it was on TypeScript, a language which compiles to JavaScript.

Aside: Hejlsberg talks about the new Xbox music app in Windows 8.1 (and Xbox One) which is written in JavaScript. It is a large app with 500,000  lines of code, and new features are now implemented in TypeScript (30,000 lines so far).

However, Hejlsberg did also talk about C# 6.0 at Build, during this Channel 9 Q&A, though you have to scroll through to reach the C# content (about 34 minutes in).

image

He confirmed that C# in Visual Studio 2013 is the same as before, but there will be new previews of the forthcoming “Roslyn” compiler soon, and that C# 6.0 will be in the “next Visual Studio after” which suggests Visual Studio 2014, presuming Microsoft sticks to its annual release cycle.

“We are at a point where the Roslyn compilers are done,” he said.

Roslyn, Hejlsberg explained, is the new compiler for “C#, and VB, and the language services in the IDE.”

Roslyn performance will be at least as good as the existing native compiler, says Hejlsberg. It is better suited to parallel processing so will take advantage of multi-core machines, “particularly for large projects.”

You can read more about Roslyn here. Microsoft describes it as “opening up the Visual Basic and C# compilers as APIs.” Practical benefits include features like instant porting of VB code to and from C#, and the use of C# and VB as macro languages within a .NET application.

Hejlsberg also says that Roslyn will enable a faster pace of evolution for C# in future.

Another aside: Xamarin, which provides a compiler for C# targeting iOS and Android, gets a nod of approval from Hejlsjberg. “I’m a great fan of their work,” he says.

Blogger (and former Microsoft Excel developer) Wesner Moise provides a transcript of the key points.

development, mobile, professional, software development

Mobile developer survey: Apple iOS most profitable platform, but even Windows Phone is viable

Vision Mobile has released the results of a survey of 6000 mobile developers, sponsored by Blackberry and Mozilla.

Reading through the survey reminds me that despite the critical importance of apps to mobile platforms, surveys which look at developer intent are poor predictors of the future health of specific platforms. High interest or even affection for some new platform tends to dissipate quickly if platform adoption is poor.

Developers influence the success of a platform by developing (or not developing) desirable apps, but this is only one among many factors. Others include:

  • Mobile operators: which devices are they promoting and subsidising most?
  • Devices: which has the right blend of looks, usability and features?
  • Fashion: which smartphones are my friends using?
  • Price: which devices are best value?
  • Marketing: which vendor is doing the best job?
  • Enterprise: in business, security and manageability are important

The report confirms the dominance of iOS and Android and is generally down on Windows Phone while there are more optimistic remarks than I had expected about Blackberry 10 and Firefox OS (but note the sponsors).

Here is a stat that caught my eye though:

Monthly Revenue per developer per platform:

  • $5,200 iOS
  • $4,700 Android
  • $3,600 Windows Phone
  • $2,900 HTML5
  • $1,200 Blackberry 10

There is more money in iOS and Android, but Windows Phone and to some extent HTML5 is financially viable too. On niche platforms like Windows Phone, I guess there is a benefit in having less competition.

adobe, development, microsoft, software, software development, visual studio, windows

Adobe Creative Cloud giveaway at Microsoft Build: sign of a new alliance?

1 Comment

At Microsoft’s Build conference last week, one among a number of giveaways to all attendees was a year’s subscription to Adobe’s Creative Cloud. This was announced by Developer and Platform Evangelism VP Steven Guggenheimer during the day 2 keynote.

image

Creative Cloud costs $49.99 (or £39 + VAT) per month for an individual subscription.

Guggenheimer in effect said that the gift came from Adobe; he suggested that “you guys should send them a nice email”.

The stated purpose of the giveaway is to promote the Digital Publishing Suite (DPS). Creative Cloud comes with DPS Single Edition, which lets you create rich, interactive magazine content and publish it to the iPad and Android. Announced at Build was DPS support for Windows 8.1 in “late 2013”; a sign if nothing else that Adobe is taking Microsoft’s tablet platform seriously.

The odd aspect though is that Build is a developer conference and not quite the right target audience for DPS. On the other hand, there are numerous tools in Creative Cloud that are well suited to developers, including the Edge web content tools, PhoneGap Build, and of course the mighty Photoshop for image editing.

Adobe’s Adam Lehman was interviewed at Build about the Creative Cloud tools, especially Edge, here.

Microsoft is no longer trying to compete with Adobe on design tools. Expression Web and Expression Design have been discontinued. The Flash versus Silverlight wars are also consigned to history, making it easier for these two companies to work together.

development, microsoft, open source, professional, software development, visual studio

Visual C++ will implement all of C++ 11 and C++ 14, some of C99 says Microsoft

Microsoft’s Herb Sutter spoke at Microsoft Build in San Francisco on the future of C++.

image

Microsoft has been criticised for being slow to implement all the features of ISO C++ 11. Sutter says most features are now included in the public preview of Visual Studio 2013 – which has a “Go Live” license so you can use it in production – including the oft-requested variadic templates. The full list:

  • Explicit conversion operators
  • Raw string literals
  • Function template default arguments
  • Delegating constructors
  • Uniform int and initializer_lists
  • Variadic templates

More features are coming in the RTM (final release) of Visual Studio 2013 later this year:

  • Non-static member initializers
  • =default
  • =delete
  • ‘using’ aliases

A technical preview will then follow and Sutter listed possible features of which there will be a subset. Full conformance will follow at an unspecified time.

Microsoft is also promising a full implementation of C++ 14, the next update to the standard, even though the exact specification is not yet fully agreed. Some C++ 14 features will be implemented ahead of C++ 11 features, if they are considered to add high value.

Two other points of interest.

Async/await (familiar to C# developers) will be implemented in the post-RTM CTP because it is such a useful feature for Windows Runtime app developers, even though it is not part of the ISO standard.

Finally, Microsoft will also several C99 features in the RTM of Visual Studio 2013:

  • Variable decls
  • C99_Bool
  • compound literals
  • designated initializers

The reason for implementing these is that they are needed to compile popular open source libraries like FFmpeg.

I asked Sutter why Microsoft is not planning full conformance to C99. He said it was a matter of priorities and that work on C++ 11 and C++ 14 was more important. If there are particular additional features of C99 developers would like to see implemented, contacting Sutter with requests and rationale might eventually yield results.

image