Does it really matter any more? The speed (ie cost) of development is far more important than speed of execution nowadays. IMHO this comes down to a scripted / web page GUI and a server back-end.
Embarcadero Technologies, having itself been been bought by private capital in 2006, acquired Borland Software's CodeGear division just over a year ago. CodeGear's developer products include Delphi, a RAD tool that creates native code Windows executables, and the JBuilder Java IDE now based on Eclipse. These products have an …
Delphi still is the easiest way to make very nice, very fast fully compiled Windows applications. If you are selling off the shelf software, it has to be zippy. Eclipse is a good example of interpreted "off the shelf" software. It's clunky as hell and gets in the way of development more than it contributes to it.
I love C# and .NET and it is good for what it is primarily designed for: intra-business applications and web sites. If I were looking to write the next greatest anything (HTML editor, database tool, word processor) I would hesitate to write it in .NET, it's just a bit too clunky for that.
and Builder can be used for web applications or other applications, it is just an HTTP stream nothing more complicated than that.
I am betting Irrelevant is a PHP bod, poor sod.
Cross platform has been solved rather well with wx, sdl and openGL though, and of course Python, Java and Perl interpretors are available on many platforms, and a compile itself is not a big deal.
So, yeah good luck to them, but it won't ever be the same as the heady days of C++ Builder, the desktops are fragmenting and the Windows figures are largely bogus, the claim that every PC sold is windows is where the fraud of that starts.
There is interest in the Pascal world for Delphi to work on multiple platforms.
Take a look at the Free Pascal Compiler and Lazarus. Both open source, and source compatible with Delphi to quite an extent.
The only thing that lets them down is the lack of components, and the bizarre, sometimes impossible installation hoops you have to jump through to get what components there are onto a Linux IDE.
@Proto-robbie - Rapid development always was Delphi's party trick, it is an incredibly fast RAD environment. It has supported cgi-bin/ISAPI development for years, and being a compiled language they really fly.
Well I bought Kylix - wrote most of our hotel management and pos software with it and still use it although Lazarus looks to be less bug ridden and more importantly not obsolete.
One of the main problems with Kylix is that is was only certified to work on distros of Linux which were old when Kylix came out and are now really old. With a bit of fiddling it's possible to get it to work on other distros, but it's an uphill struggle. Any software that allows Pascal/Delphi programming on Linux is welcome here!
I wrote quite a few little DOS tools in Turbo Pascal, maybe trading program size and speed for the time to write the stuff. Delphi looked good, but the Windows GUI, even with Delphi, killed that side of development.
And fast code is more than just compiling instead of interpreting. It's lucky I'm using a Linux box.. I have lots of choices.
Even if some of them look more like line noise than a language.
Borland's strength and beauty was always being the best toolkit in the desktop developer's market. Pushing Delphi and siblings into the middleware & enterprise markets was a typical "idea in a suit" abortion. Going head-to-head with the likes of IBM, java etc was doomed from the start and everyone knew it.
Delphi was so good at those kind of small bespoke jobs that even a developer like myself who would rather stick needles in their eyes than code Pascal would use it for certain problems. That and the incredibly rich ecosystem of small shops providing components and libs, the ease of drilling down from RAD design, through the windows API, all the way to win32 asm if needed meant one could have performance-critical things like - for example - audio plugins looking beautiful and performing well in no time at all.
So Borland was maybe at home in a lot of niche markets, but so what? - it's still money in the bank at the end of the day, and better than having no money, no future, no developer base and being kicked around like a corporate football.
Nope, I still have a copy of Delphi before it got numbers (£250 in the late 80s?), and wrote some useful software in it. It was good for building tabbed dialogs, but not for serious WIMP stuff, 'cos it got a lot harder to eg drag off a large window and scroll arbitrary graphics or components on it.
As you suggest I can write PHP, but can also do Perl, Python, Algol, Fortran, Java, PL/SQL, VB, C, C++, C#, Z80, 6502, 8088/6 ... Pentium, 6800 / 68000 family & ARM Assembler, Pascal, Delphi, Ruby, C++ Builder, Rebol, Lua, Lisp, Forth, Prolog & Cobol. I have probably missed a couple, but I am getting on a bit. My first job involved writing in 8-bit machine code, and I went to the Macintosh launch on my 25th birthday.
Yes, Delphi is a great programming language, fast as crap and not a bad choice for writing a desktop application. It's well understood by one percent of the programming population and for that reason unsuited for enterprise use despite its technical merits.
ps how is DLL hell these days? Haven't had it since moving to web pages! To be fair there are other issues, mostly caused by M$oft.
As of Delphi 2009 (released in 2008!) we now have...
- Unicode is now the "natural" string type for the IDE, compiler and language
- Anonymous methods
- Parallel programming support in the form of a Monitor synchronization object supported by every object (rudimentary at this stage but a clear signpost for some future language infrastructure support)
Delphi HAS moved on and kept up (and in some ways gotten ahead again), even if some peoples perception of it hasn't.
This is going to be a hard re-ignition. I don't see potential market in cross-platform IDE. Moving forward, I see people would not mind to write their application in any language and any platform, especially for server application. Why? Because we have virtualisation and services interopability. I also don't see any major advantage of having a cross-platform IDE for developing desktop application. Why? Because every platform has their own best IDE. If you want to develop a Windows application, use Visual Studio, there are tons of .NET libraries available that will work really well in Windows 7. If you want to develop a Mac application, use XCode, again, tons of libraries available. Creating "many" shares library may not maximize the platform capability.
Additionally, Microsoft optimize their .NET framework to improve the "performance" and "re-usability" - with quote as I don't know how true this is, but I am sure that with faster hardware, they will definitely improve something. Today, the .NET framework is working with C++, C# and VB language, and for developing desktop, client/server, web services, web application, etc.
I agree with the word "betting" used in the article, I think they should re-assess the success probability by looking at MONO and Java.
REALbasic ( www.realsoftware.com ) has been offering cross-platform development, Windows, Linux and Mac ( both native IDE's and cross-compilation ) for ages. Not always brilliantly ( I binned an early edition in absolute frustration ) but it has really improved in recent years.
It carries the "Basic" stigma which drives 'real programmers' [sic] apoplectic but it's easy to use and a natural progression for past VB6'ers who flinch at dotNet. Of course, that's another reason for 'real programmers' to diss it, but what do I care about narrow minded attitudes when my cross-platform apps are making real money and keeping people in jobs.
Good luck ( seriously, not sarcastically ) to anyone who wants to widen the cross-platform market.
Delphi rocks, but the price sucks balls. If they would just make one version and then bring the price down to about $2-300 per copy they'd probably sell 10 times as many equalling the profit they're aiming for now.
But the difference is it would make it WAY more attractive to new-comers (and people who left for cheaper pastures), therefore bulding the user-base and community back up, making it MORE appealing for even MORE new-comers.
It just sucks that it has to cost like $1300 for the average single developer to be able to use Delphi for things like micro-startups and internal usage in a small company, which is where the roots of Delphi took hold from my experience since TP7.
Prices like $250+ for quality add-on component sets wouldn't be such a kick in the face after dropping $1300 for the IDE. If Delphi was cheaper there'd be more money left for the component creators, so they'd sell more, and make more/better components, growing Delphi's abilites further and expanding the community (again back to more sales of Delphi, repeat)
I think alot of why people think nobody is using Delphi is becuase everyone goes "WTF kind of price is that?!" and decide to priate it instead of buying it, and from there keep it on the down-low as to the tool they're using to avoid the cops. ;)
I've got to agree with marc 9 on that point. RAD studio feels like it was rushed out the door.
BDS (Borland Developer Studio, the predecessor to the latest CodeGear offering) had some issues (ctrl-f6 not working, buggy under Vista) at least it did what it was supposed to (compile and link) fairly well.
I've been using CodeGear C++ Builder for about a week, and I'm impressed with everything except its ability to successfully compile and link. And that, at the end of the day, is the most important feature.
Biting the hand that feeds IT © 1998–2019