http://bugs.freepascal.org/view.php?id=15089

are complete web-images, with the starting Delphi, output Lazarus, and conversion path notes.

This is a good example, of a variation that initially looks serious, but does have common
ground, and code can be written to run in Both Lazarus and Delphi.

Then, I would enter threads in there covering ‘work in progress’ and reports,
and Gary will be setting up a Lazarus download area on his site, and that
will also include porting comments.

It is an incremental thing….

In the last couple of evenings, I have reached 100% yield on the 25 I chose.

This is my framework structure at present:
* I have a common “Rules” .TXT file I add to as I go along..
* Each project has a pathway report .TXT file
* Each project has a \Delphi_Start master sub-directory
* Each project has a WebZip typ batch file, that grabs all this into a zip file.

All this is designed to be clone-able, so users can copy/rename to adapt to any new project.

In theory (ie hopefully) on unzip, you should be able to
a) Load/Compile/Run the Lazarus project
and, for the keener student…
b) follow the rules, and pathway notes, and so reproduce the same port yourself.

Once the setup is done right, (with the good yield ones) it take 3 clicks.

There is some great resource out there, representing years of experience,
and highly accessible.

My favorite example is http://www.delphiforfun.org/default.html

You USED to be able to get free, personal Delphi editions with magazines.
However, you cannot move this to another hard drive, and there is no web-archive
So that resource has to decay over time, rather like radioactive half life …

There was Turbo Delphi, but that too was recently yanked
TD has a nice ide, but rather a fat install, and no chance of running on a flash drive…
There were also cripple-ware aspects, waiting to trap the unwary…

All of Which has left a bigger hole, not made a smaller one..

Indeed, it has now made it VERY hard to generate teaching material, which has to have
a long and stable lifetime, and by very system-portable.
(ie work across a wide range of installed class PCs )

So, we start looking around for alternatives…

Lazarus has been ticking away quietly in the background, so we check that again, and
I started porting some of these nice examples http://www.delphiforfun.org/Programs/index.htm

Some rough edges, and a few ‘rules to follow’, but no real brick walls.

Now, after a few weeks, I have ported ~25 of DFF examples (Delphi6/7) to Lazarus.

First pass Yields are ~80% on the simpler programs, falling away as they get more complex.
Conversion rates improve, as I learn more about Lazarus (& Delphi)

By default, Lazarus removes uses Windows, and mostly that is OK, but on ~20%, that needs
to be added-back, to avoid actually re-writing code.

One project even now runs in Lazarus, (& D7) but fails to run in Turbo Delphi (wry smile here… !)

A couple of others, Turbo Delphi stutters on, (gives errors, but works)
but I think that is the cripple-ware kicking in ?

These projects include (simple) Graphics, some quite complex algorithms, and some do sound.
All of them use programming to solve, or do, something, so are great for teaching.

So I reported some of the rough edges into Lazarus, and one is already fixed.

Lazarus does fire some windows type events just a little differently, which can cause some
head scratching, but usually a Delphi-compatible common ground solution can be found.

Perhaps most important, in NONE of the porting cases, did I need to rewrite actual code structures.

Some needed additional tests, or an Init order changed/improved, or a type-cleanup,
but the puzzle algorithms did not need to be touched.

So far, is it looking viable. Lazarus can be snapshot stored anytime, has a relatively small
download footprint, is very OS agnostic.

It is looking reasonably student friendly, especially if starting with a working project.

Lazarus might not have the ‘twinkies’ and gloss of Delphi, but it is quite usable, and indeed, a good
complement to Delphi.

A teaching lab can run both and show students the details of porting, testing, and how to create a common
code base. They can leave with a knowledge of (at least) both and choose in future employment.

The education prices of Delphi are reachable for many (but not all) at ~$100, but again with
Lazarus, you can make Delphi go further

Students can freely take Lazarus home, and a single-Delphi copy/license in-class allows test/contrast.
Lazarus allows the faculties to keep within the Delphi education license, if they need to release
working software (as in student projects)

Importantly, with Lazarus as a common denominator, there are no ‘drop dead’ teaching situations.

“it is an ill wind that blows no good”

The internal worker-Thread of the Component caused problems when Destroying the Form that contains
and this could randomly crash the Application.

Research revealed that there have been threading issues with the component for ages. And the comments on the buglist didn’t convince me that anyone there really has a clue about multithreading. (The author was talking about using flags here and there to make the problem less likely…)

-> I removed the component immediately from my code.

Running 2 technolgies in parallel is hard. Especially targeting the same environment – Windows Desktop … – we will see how long it will last WPF is at least an invention – fairly makes sense. Maybe looking at Freepascal and Lazarus can help to retain the memories. Visual Studio does not lack a lot and resharper is excellent.

The argements for selling a product are different to the real technolgical underlying. The VCL for Windows native development is still leading. Rewriting Software is expensive. Delphi is in maintenance imo accompanied by

a) movement of existing code to the .net technology – they are the only that try
b) providing integration solutions for existing software

Maybe a new kind of “Powerbuilder”. .net and VS are very architecture centric. Still the C++ 90s mindset, this is why those guys feel at home in C# and .net, and it works. Sometimes the .net classes do not provide the functionallity at the moment than the regualr C APIs, the majority wont need.

Enterprise Solutions: Maintenance costs have to be cut … competitors bring Development Environments, especially in combination with AJAX technologies where the “IDE” is fully integrated into the web based development envirionment … in order to compete existing, extensible products. Several Stacks do already exist, a lot more to come (beside the clouds). Here the Windows + Native approach is not the answer, imo, in a near future. The answer is quick architecture orieneted development … or something ready to use and extensible like SAP which is available on 2 technology stacks mainly focusing on the Java Stack. A web based application in the meanwhile runs on Apple, Linux, Windows,…. and can simply be brought to any other device … in this area a discussion about VS vs. Delphi funny but not accurate. ASP.net and Intranet still fit.

@PiotrL Says –> I see it very simlar. Lazarus … There are few things not possible, maybe somehow you came to same conclusion. The problem is not the IDE nor the language. I also like the GLxxx stuff solving problems on totally different level… a new approach… fits perfect. I personally see the developer of the 90s here.

Mike

That all being said, i don’t think we’ll move to C# or Java. We have talked about it actually, but there’s still some advantages in using Delphi and no actual reason why we shouldn’t continue using D2007.

I disagree strongly with the lazy “fix only the most complained about bugs” attitude of Embarcadero. Many of the outstanding IDE and language bugs could easily be fixed and would make a big difference to me. The guy who sorts through the bug reports in QualityCentral seems to me to be an obstacle to improving the product. By now, the IDE could and should be polished to perfection. They could take pride in a close-to-zero bug product. I’ve tried teaching my 7-10 year old kids how to program in Delphi (as well as VB.Net) and though they cope well with the language syntax, sadly I often have to dig them out of IDE issues or “Find Declaration” not working etc etc – bugs that could have been fixed years ago. (Regular users work around these problems, perhaps unconsciously, but they show up with beginners).

For all these problems, Delphi does have some unique and important qualities.

1) Delphi code runs fast. Really fast. I’ve written a huge suite of medical simulation software in Delphi and any attempts to port it to C++ (when Delphi looked like it was dying) have resulted in slower code. VS C++ does some crazy things such as creating unnecessary local structures for the return value of a function, then copying each element across, rather than using the return value pointer directly. Even using “inline”, the C++ was version slower than the Delphi equivalent, in our tests. C++ has a reputation for being fast. Delphi deserves at least an equal reputation. VB.Net and C#.Net applications are very very slow by comparison (unless they are mostly calling native code/libraries).

2) Delphi compiles fast. Really fast. For small or stable projects this doesn’t matter much. But I write huge applications, or suites of applications, and I wish to fix all problems wherever they occur, right now, not write lists of them. In C++, C# or VB.Net this means I spend literally half by day watching the compiler (or using the exercise bike next to the computer). With Delphi, I don’t get a moments peace. It works me hard, and I get a lot of work done. Single specification + single implementation applications are always horrible for the end user. Iterating the specification and implementation phases is the key to great software, and no language allows this more than Delphi (though unlike VB.Net, Delphi does not allow “Edit and Continue” which can be useful for putting the final polish to applications).

3) The Delphi (Object Pascal) syntax is absolutely brilliant! Neatness is crucial when working on huge projects single handed. For programming snails that crawl across the code reading one word at a time, this is not important. But with the right sort of brain (perhaps visual rather than auditory?), you can “speed read” a page of neatly laid out Delphi code in literally just 1 to 5 seconds. Comments – which by virtue of the fact that they are not checked by the compiler or run are invariably inaccurate or out of date – are almost never necessary because the code reads so well (provided good identifier names are used). Better still, you can spot bugs and anomalies at 10 paces – they just jump out at you. I programmed C++ for years, and was once right at the head of all the latest C++ language developments, and I never got to the point of reading C++ anything like as quickly as I can read Delphi. I believe that programming should not be about syntax and coding; it should be about problem solving. Delphi gets very close to that. C#, by contrast, is almost as messy and visually inefficient as C++.

4) They provide the VCL library source code. This was a very clever and sensible move. I do not believe that I would have switched to it if they had not done this. Thanks to this, I’ve found modifying control behaviour very much quicker and with better results than trying to achieve the same with VB.Net or C# where the library source code is not provided.

I need Delphi to live on. For the sort of work I do, Delphi is the best language/IDE available.

Rather than charge so much for Delphi upgrades, Embarcadero should take steps to both save money and improve their product:

a) Strip out all their weaker programming and other staff.
b) Disband their (completely unnecessary) international infrastructure.
c) Concentrate on polishing their product with a small but capable team. It really can’t be that difficult (unless they’ve allowed their source code to get into a mess).
d) Eliminate most of their support services, because with a bug free product programmers will be able to get enough help from others on the internet.
e) Eliminate the entire Delphi help system. It used to be good, but it doesn’t work anymore. It takes so long to start and is rarely of any use.
f) Replace the “guard” that monitors CodeCentral bug reports with a senior programmer trying their best to perfect their product by getting as many bugs fixed as possible. Let’s face it, VS 2008 VB/C# are well polished, and Delphi must at least match that.

Delphi should be targeted more towards high level developers. It deserves a lot more respect than it has. In many ways C++ is a complete mess compared with Delphi. But Delphi has been tainted by too much of the short sighted “One click ActiveX control” type nonsense. That sort of marketing attracts people who will never really get very far with programming and ultimately gives the product a bad name. Similarly, the early days of Delphi were tainted with an obsession with overly simplistic database programs. All that should go, because those tasks can be done very easily and well with VB.Net (which thanks to cross fertilization from C# has matured into a very good language). Instead Delphi should play to it’s strengths. With Delphi a very small team (even one person) can write huge, polished, high performance applications, that can readily compete with software from much larger teams. Embarcadero should improve the product to help existing Delphi programmers do this, and encourage new programmers to do likewise.

The fact that Delphi pricing is thoroughly screwed up has been made clear to Borland/Inprise/Codegear/Embarcadero for years. They just don’t get it.

They need to wake up to a few simple realities

a) Their friends aren’t in the “enterprise”. The “Enterprise” is currently on .NET and J2EE – they have their own IDEs from IBM, Sun, MS, Oracle etc.

b) They are NOT interested in investing in a product line which has very few developers – just count the number of people on Monster who are listing Delphi as their primary skill. They don’t want to have to put out ads for weeks before they find a single good developer. EVERY company that I know of who was making Delphi products have transitioned to .NET – EVERY company. They have either laid of their Delphi workforce or trained them to work on .NET.

c) Delphi developers are fast moving out of Delphi into other languages. I moved out into C++, .NET and now Java. That has been the same story with almost all of my former colleagues who were in Delphi. If you need to find a Delphi job, it is very difficult right now because of (b).

d) I am very sorry but the truth is that most companies can develop pretty much any “enteprise” solution right now using free IDes. Eclipse is very, very popular. Intellij Idea isn’t very expensive. The libraries are all available free and so also the app servers. If you need the commercial libraries, they can be purchased without one having to change the IDE.

Delphi should have been made into a cross-platform tool which would have allowed development of DB apps for platforms from the desktop to the mainframe. Their current tool support for this has come in too late – the enterprise already has good products for this and they don’t need one more. Kylix was ok, but it wasn’t sufficient. Also your prices were too expensive.

I run a software development company now. I am NOT going to invest in Delphi when you consider that NONE of my clients would want me to develop their solution in a language/platform which looks like it is going to die out. If I suggest Delphi as the language for their solution, they will walk out on me and go to a company down the road willing to develop in .NET, J2EE etc.

The only hope that Embarcadero has is to:

a) Increase the number of programmers available who have Delphi skills. They can only do this if they make the product available very cheap (sorry – no “enterprise” pricing). The barrier of entry should be kept really low. If not, just forget it.

b) Have a single base IDE with “enterprise” functionality as additional modules which can be purchased. This would lower the barrier of entry and also provide companies and hobbyists the option of getting additional functionality when they require it. Have a better marketplace for Delphi plugins and components and additional functionality.

c) Encourage businesses seeking to develop software solutions to use Delphi. Again,this must be priced really low – huge discounts should be given for licenses and even greater discount for volume licenses, incentives for developing solutions and a market place for the same.

Win back your friends and please do remember – your friends aren’t in the “enterprise”.