PB - the road ahead

[Hedley Moses]

I just came back from a UML overview course today which I found extremely interesting. Mention was made of how the model in changing especially in big corporate environments away from the standalone mammoth applications to more of a service model. Example mentioned was that no longer will we be developing an application to build and all encompassing application but rather services which perform a specific task. The developer or systems designer will in future assemble the various services required to build up the required functionality rather than develop the application from code. This is a scalable model which can be built on or taken away from depending on the organisations changing needs over time. Of course its has to be built on OOP be to able to do this. Another comment the presenter made which I found most astounding was that .NET was designed to eventually replace COM. The benefit is that the programmer is able to develop applications which are no longer platform dependant while having access to many thousands of tried and tested framework classes.

With this is mind I was wondering how the future of Native language programmers will be impacted and especially how products like PB, which are not framework based but rather compile to native code, will be impacted. Will it survive if it does not embrace concepts like .NET. Should it be looking at concepts like model driven architecture as the way forward. In terms of .NET Microsoft definitely seems to be going the full .NET route

[Greg Turgeon]

Speculating about the future invites madness. Here's something to stir this particular pot a little more: http://news.cnet.com/2100-1046_3-5148148.html.

A sample:

Microsoft software architect Don Box said the company will not invest much more in Component Object Model (COM) and Distributed Compound Object Model (DCOM)--Microsoft's mechanisms for sharing objects between programs. Instead, Box said, programs will use managed services based on the Extensible Markup Language to communicate with each other. Box is leading the work on the "plumbing" part of Longhorn, called "Indigo," which is effectively the successor to Microsoft .Net and as such will dictate how programs are written in future Windows platforms.

This, from January 2004.

I'm willing to bet only that "The developer or systems designer will in future" do what most of us do, and that's means exploiting whatever targets of opportunity come along.
The smart ones also will pay little attention to Microsoft's declarations about the future of anything.

[Joe Byrne]

I have to agree with you Greg. I've been in the programming business for 25+ years and if I only had a nickle for every prediction I've heard.... oh how rich I'd be.

I clearly remember people saying that PB wouldn't survive because the future was OS/2. Then when PB/DLL came along, there was talk of doom and gloom for PB because VB was "so much easier and powerful".... yea, well, I'll keep putting my bets down on PB.

There is still lots of talk about "everything going to the cloud". I write Web based apps as well as "stand-alone". Both have their place and both have tools specifically well suited for the purpose. But when I have a customer with a problem, they really don't care how the problem is solved. The good programmer will analyze the problem and develop the best way to solve it. That is why its a good idea to know more than just one technology and more than one language.

The future? I doubt anyone has the ability to really predict it accurately with any consistency. At best, some of the ideas will pan out, but more often than not, things will go a totally different way.

[Chris Boss]

Software development costs can become very expensive and companies have to find ways to speed up development.

COM (Component Object Model) was designed to be a solution to the problem.

It is not because COM is somehow better than any other method of programming, but because COM does a very simple thing:

Code Reusability

Thats the key.

The thing is though, that code reusability does not require COM, DOT.NET or any other "leading edge" method pushed by Microsoft.

Another term for a reusable block of code is a :

Library

A DLL is a Library of reusable code blocks (subs/functions).
You can write a library which is simply used as source code (aka. Include file).

PowerBasic will not somehow become obsolete because of new technologies. As long as one follows the simple principle "find ways to reuse code", Powerbasic will be a powerful tool for programmers.

The advantage of using well designed libraries is that once they are well tested and fully debugged, they can be used over and over again to produce reliable software quickly.

Libraries can come in many forms, some low level and some quite high level. Low level libraries will always be valuable, for those who need to do a lot of customizing. High level libraries though make software development much quicker. The building blocks of bigger projects.

I have long felt that quality library code (reusable) makes a programmers life much easier. If a programmer doesn't have the habit of converting often used code he/she writes into reusable library code, they won't be programming for long (it takes too long) or at least won't be that productive.

I am a big fan of both Library code and code generation tools.

Can anyone remember a tool for DOS programming called:

SoftCode

It was an amazing software tool. The code generation engine and parser language was just fantastic. The screen designer was good.

The only mistake they made, was the templates (code used by code generator) were not very good, IMO.

I liked SoftCode so much, I wrote my own code templates for it. It took quite awhile to do, but once I was done, I had a code generation tool which allowed me to build DOS applications rapidly. My code templates had the ability to do drop down scrolling windows (like a drop down Grid control with multiples rows and columns), popup windows and it had a proprietary database engine which not only created a multi-user database (for networks), but also used a proprietary mirror imaging database engine, which allows you to break up the work and have multiple key servers with mirror images of the database.

Reusable code libraries and code generation tools are the key to fast, quality software development. With such well designed tools, one should be able to develop software ten times faster.

PowerBasic is not lacking because it is not dot.net ready.

Powerbasic has all the features necessary for developers to write quality reusable code. Some will do this using DLL's. Others will do this using the new OOP features. Some will do this using code generation tools and source code libraries.

As long as Powerbasic offers a number of ways to reuse code, it won't be made obsolete by COM (ActiveX) or dot.net.

While other methods of code reusability will continue to produce bloated, slow applications, PowerBasic programmers (with the right tools) will be able to produce fast, lean, reliable software.

[Hedley Moses]

I guess Bob must have grey hairs trying to plan a roadmap for the development of PB products. The big boys like google etc are pushing cloud computing and that must have quite a significant influence on how software houses develop their products

[Vincent van der Leun]

What I always found interesting is that Microsoft tells everybody to port their apps to .NET as soon as possbible (note that I love .NET and I find it a joy to program in C#. I definitely have nothing against it)...

But if one looks at Microsoft's own tools supplied with their operating system or their commercial applications, they don't seem to use the .NET framework themselves that much for key applications. I don't know if Windows 7 will change this, but I believe only a very minimal of executables supplied with Vista are really using the .NET framework to do their work for example.

Therefore, I guess, it would seem to me that native coding is perfectly save for the coming years. I can hardly believe Microsoft will stop documenting their API, as a lot of big companies, like Microsoft, are also still not using the .NET framework for their applications.

[Edwin Knoppert]

> I can hardly believe Microsoft will stop documenting their API

This may become an issue.
Windows7 users may in some cases have to use a Virtual PC session with XP on it.
I found a topic on that:

http://community.winsupersite.com/blogs/paul/archive/2009/04/24/secret-no-more-revealing-virtual-windows-xp-for-windows-7.aspx

This is imo one of the biggest changes MS made.
I don't mind but i suspect we may have to resort to com programming much more as anticipated if we program for Windows7.
Luckily PB gave us very good support on that in the last release.

>PowerBasic is not lacking because it is not dot.net ready.
The different object types in .NET are simply not easy to be accessed from a non-.net programming language.
For example, i stress the PwrDev users to write a custom assembly as bridge and establish 'contact' using easier variables and methods.
Trying to access the internals of the CLR is imo a waste of time.

A tool like MS c++ (VS) does both worlds, it must have been a major job i guess.

[Stanley Durham]

The future is in the clouds.
At least, that’s the current big hype.

cloud computing

Microsoft Office will float to the cloud

[Hedley Moses]

I would guess that DOTNET wont go down very well with Powerbasic where the emphasis is on smaller and faster. By its very nature DOTNET would be slower than native code compiling. The newer technologies do however have some real advantages which cant be ignored like platform independant development as an example. There's real advantage in that. The question then is where should Powerbasic pitch itself especially in the corporate environments where the system landscape includes a mix of all technologies, ERP, MS and web all interacting with each other

[Rodney Hicks]

in the corporate environments where the system landscape includes a mix of all technologies, ERP, MS and web all interacting with each other

This is only a niche market, for there are lots of corporations that have no need to be in the limelight with their head in the clouds.(couldn't resist the pun)
This coding business is not unlike the wheel.
One doesn't want to reinvent the wheel, but at times a new wheel is needed. The bicycle wheel was invented long before the airplane wheel, but when the airplane wheel came on the scene, it did not make the bicycle wheel obsolete.

The need for a platform independent technology does not obviate the need for the present technology.
Think of it another way, do you want everything your company does in the clouds?

[Stanley Durham]

...question then is where should Powerbasic pitch itself...

PowerCLOUD?

[Cliff Nichols]

PowerCLOUD?

Would that not be a recreation of discussions in the past about VaporWare???

[Joseph Cote]

Would that not be a recreation of discussions in the past about VaporWare???

Pedantically I would have to classify this as CondensateWare.

[Scott Slater]

or water vaporware.

[Bud Durland]

Whenever I read the term "platform independent" in this context, I giggle a little...

[Mr. Kevin Diggins]

An observation:

Most web applications that I'm exposed to are doing "databasey" things - which is fine, if that's what you need done. But in my little world, I'm presented with lots of problems that require coded solutions, most having nothing to do with databases, except perhaps cleaning them up. And while some of those tools are short lived, others remain useful year after year.

Moral of the story:
When your only tool is a hammer, everything looks like a nail.

[Marcel Kollenaar]

[...], which are not framework based but rather compile to native code, will be impacted. Will it survive if it does not embrace concepts like .NET. [...]

Even the software behind .NET is written in native code. PB wil survive, it has nothing to do with .NET but everything with user demands. What is wrong with COM, DCOM. It is one of the best concepts they ever invented. Use it and keep using it. The only thing is that Microsoft didn't want to lose was the Internet. They build a new common internet layer for their languages .NET. But it smells all the same. If you code in C# or VB .NET the result is the same.

Over a few year there will be world wide data crises (The Cloud) and eveyone is abandoning the continous web access. Falling back to "stand alone" programs and only make contact with web in very secure short burst mode programs.

[Bruce Atkinson]

I believe that the most telling thing about .NET and whether or not you should worry about abandoning native code is to look at what the Windows programmers are doing. A good example is to look at is the new API for the new features in the Windows 7 taskbar. The API for that is implemented using COM. A special toolkit that is a wrapper for this API is provided so that .NET apps can use these features.

So you don't have to worry about it for a while.

You also don't have to worry about everything as a service. While people like to create them and think of doing things that way, the reality is when you need to get things done in a timely way and you have to deal with large datasets, it isn't a practical solution.

[Brian Chirgwin]

I believe that the most telling thing about .NET and whether or not you should worry about abandoning native code is to look at what the Windows programmers are doing. A good example is to look at is the new API for the new features in the Windows 7 taskbar. The API for that is implemented using COM. A special toolkit that is a wrapper for this API is provided so that .NET apps can use these features.

So you don't have to worry about it for a while.

You also don't have to worry about everything as a service. While people like to create them and think of doing things that way, the reality is when you need to get things done in a timely way and you have to deal with large datasets, it isn't a practical solution.

If I remember correct Windows 7 is the first version of windows that DOS and Windows 16 applications won't run on. Your great great grand kids will be retiring before native code is done away with.

.Net does just in time native code compilation (on the fly compilation). So any .NET app is really just like any other language. I don't see MS disabling application written in other languages C, C++, PB in some way. Possible, of course, but highly unlikely. No one would move to that version of Windows. Actually, it wouldn't be Windows anymore. Every application that wasn't .NET and some that are only partially .NET would stop working. 99.99% of all applications ever written.

No worries for the next 150-200 years.

I'd be more worried of Linux (or other OS) taking over market share and becoming the OS of choice.

[Scott Slater]

If I remember correct Windows 7 is the first version of windows that DOS and Windows 16 applications won't run on.

Vista 64-bit is currently that way.