Clarion.Net update

The AppGen for Clarion.Net is making steady progress (despite a few wild rumors to the contrary).  It’s a given that the progress is much slower than we would like, and in moment I’ll explain why that is the case. The core of Clarion.Net just like Clarion Win32, is not the AppGen, but is the Clarion language and Clarion runtime function library.  For those who are willing and capable of writing their own code (as opposed to generating it), Clarion.Net has long been both productive and complete. Let’s take stock on what we have now.

Well first and foremost we have the Clarion language, extended to be able to take full advantage of the .Net Framework while maintaining compatibility with Clarion syntax and language features. Let’s not underestimate the importance of this. I know there are those developers who claim they don’t care what the language is, but they are the minority. For most of us, the language we read and write to build our applications has great significance.

Next we have the Clarion Runtime, every function we have in Clarion Win32 we also have in Clarion.Net. The entire runtime has been ported to 100% verifiable .Net code with the exception being Window structures and the corresponding Accept loop and its related functions. So we lost Window structures, but did we gain anything in the trade? Yes we did, we gained the ability to seamlessly work with any add-on UI component library available. As you know Window structures are constrained to the core set of controls that we have had since the introduction of the Clarion for Windows product. Not so for Clarion.Net, the world of add-on controls is now wide open. And when we talk about the Clarion runtime let’s not overlook that this support encompasses the Clarion Driver technology. Driver support is bread and butter for most of us.

Is there anything else? Can we ignore what might well be the most important reason for some? Clarion.Net enables you to use you existing Clarion knowledge to build and deploy .Net applications for all of the platforms supported by .Net (Desktop, Web, and Mobile). Of course you could pick up a new language, and in a few years be a master of it. But your knowledge of the Clarion language shouldn’t be taken lightly, it’s a valuable asset.

Is that all? You have access to the entire.NET Framework and all of its classes and methods, including XML Web service classes. You have painless interoperation with both COM and native code Libraries. Clarion.Net programs and code can be integrated seamlessly with any other .Net languages; inheriting from classes created in other languages, invoking and being invoked by methods written in other languages, in other words you can use any library written in any .Net language, the same as if it were written in Clarion.Net. And the reverse is true.

So that’s what we have now, but what we’re missing is the code generation that separates Clarion from all competitors. As you may know the template language used in the new AppGen.Net is a complete departure from the syntax used by the templates in our Win32 Application Generator.  In place of the Clarion template language we use an extended set of T4 template directives. The extended set of T4 template directives include equivalents for our #PROCEDURE, #CODE, #GROUP, #CONTROL, #EMBED and #AT and #EXTENSION directives as found in our Win32 template system. And there is the root cause of the delay. The T4 engine has many advantages for generation of .Net code – but it’s primarily capability by design is for generation of individual classes and code snippets. The base premise of Clarion code generation is one of round-trip code generation, embedded code within the generated code block, generation of entire applications, and at the core, an Application tree that represents the structure and relationships of procedures within the application.  We very badly underestimated the complexities of imposing this required functionality onto the T4 engine, and it cost many more man-hours than we projected in our most generous estimations. We apologize for that, as we know a lot of Clarion users are purely AppGen developers, and are not going to be productive without the code generation functionality. But on the plus side, we now have the prerequisite functionality working, almost completely. And our stated plan to make the templates available on a public subversion repository so that users have instant access to updates and fixes, and have the ability to contribute new features and fixes (or even new templates and template sets) remains central to our plans (both for .Net and Win32).

The long-term goals remain unchanged; to create a code generation environment that opens the doors to WinForms, WebForms. Compact Forms, WPF and Silverlight. And in fact beyond those platforms, the new AppGen will open a path to Android and iOS. In order to get it into your hands faster, the initial templates will scale back on some of the optional features like the option to generate test cases (nUnit) for aspects of the APP. That can added later, by us, or by the community.

In the interim before the release of the AppGen for .Net we’ll be releasing an update to Clarion.Net that has a number of fixes and delivers the new technology that allows you to easily expose Clarion.Net methods to your Win32 Clarion applications.

2 thoughts on “Clarion.Net update

  1. many thanks for the update Mr R Z. the current version of clarion sharp has allowed a good deal of fast prototying and has given us all a good start at adding another language tool to the dot net arsenal. Carionsharp is a much easier to use language for many developers since it is not as wordie as vb dot net and much simpler then csharp.

  2. Hi Robert,

    First, thanks for the update … and the candor regarding the status of

    Althought I believe there was mostly good news on it … can you expand on this sentence: “But on the plus side, we now have the prerequisite functionality working, almost completely”. The almost completely is what cought my eye … is this on of those cases that the last 10 yard of a marathon will take as much as the last 25 miles? A bit more information on a “realistic” timeline for the appgen will be appreciated.

    Again, thanks.

Comments are closed.