I think Delphi simply needs a plan. I mean: Borland - Inprise - Borland - We're going to sell! - No, sorry, we're not going to sell, it's CodeGear now - Ok, we ARE selling, it's Embracadero now. That did not really help building confidence...

Then, they seem to do too much with too little manpower, and the result are products like Delphi 8. Oh yes, they were people who actually built stuff with it, but that does not change the fact that it was really bad, sorry guys. The new IDE (I think it was called Galileo) took until Delphi 2006 to actually be usable, but at that point they already lost a lot of reputation due to two less than stellar versions (D8 and D2005).

Then, they did leave out Win32 too long. Unicode VCL? Bah, just use the shiny new .net! Now, if their .net implementation would not be soo far behind Microsoft that would may be work, but with a Win32 Delphi that starts to age and a .net Delphi that is 2 years behind Visual Studio, I did not really have the opinion that the guys had a clue what they want to do.

I do not want to sound too harsh though, the Delphi Guys were quite open and available in the newsgroups, and they did not really try to hide the fact that it was Borland Management who is to blame here, but as a customer that does not matter to me.

So basically, a period of bad, old and overpriced products really cost them.

I think they are now moving strong again. They seem to be under the hood of a company who actually cares now, they are start making more interesting Products again (to my knowledge, Delphi 2007 has a good Vista Support, working AJAX and Delphi 2009 FINALLY adds Unicode, even though I did not try and use them). Blackfish SQL looks really interesting as well.

I think they still need a bit more polish, a clearer road map, re-think the pricing. Compared to the Visual Studio Professional Pricetag of 899 €, Delphi is not really competitively priced, and the full Developer Studio is even less so.

The Delphi Language is still quite attractive in my opinion, so I would be sad to see them go down, but in the last years, Delphi reminded me of the Amiga, with all the positive and negative thoughts associated with that.

Delphi is still the best development tool around for Windows. It's better than anything Visual Studio has to offer, but it's not better enough for people who have learned programming on the C family to feel an incentive to switch.

You want to make Delphi stand out? Fix the obvious annoyances in Delphi and just let it shine. There are two separate sets of features that need fixed: IDE annoyances and language annoyances. First, the post-D7 IDE is a horror to work with. Delphi made its reputation by being better than Visual Studio. Why, why did Borland decide to consciously imitate an inferior product?

• The D7 helpfiles were wonderful. I'd be hard-pressed to imagine a better documentation system for a programming environment. What in the WORLD has happened to it? Microsoft's document reader is just plain awful, and the help for the last two versions weren't even complete! Bring old-school help back please!
• Slow, slow, slow! There's just no excuse for Delphi being as slow as it is. D2009's a lot better than the last few have been, but there's still room for improvement.
• The dev team needs to remember that an IDE is a text editor, and a text editor never blocks. As long as there isn't a dialog box open, there's no reason why you should ever be unable to edit your code. That means that if CodeInsight needs to sit around and spin my HD (for up to 5 minutes, in some cases) to display some popup or tooltip, it should do it in a background thread. Same goes for whatever it's doing the first time I hit F1 in a session. Until it's done and that help options dialog box actually appears, I should be able to write more code.
• Someone already mentioned making all the random squigglies go away. If it will compile, it shouldn't be marked in the code as an error, period. Again, D2009 is a lot better at that, but it's still got a ways to go, especially with being unable to find units listed under uses. (And generics tend to screw it up too.)
• Class completion (CTRL-SHIFT-C) has some annoying corner cases relating to properties that need to be fixed. Check out what happens if you press that while you have a write-only property in your class, or a property whose read or write declaration refers to a protected member of a parent class. The first one produces code that won't compile. The second is subtler and more dangerous: its result compiles, but can break working code.
• It's just too expensive. It's a great program, but it's not worth nearly as much as they're charging. Turbo Delphi was a great idea. Update it to a 2009 version and they'd make a lot of people very happy.
• I can has multiplatform support plz? Being able to write code on Windows and cross-compile it to Mac and Linux (or to write the code and compile on other platforms) is a feature I'd willingly pay extra for. Lazarus is nice, but it's just missing too many critical features, and its debugger is a mess.
• EDIT: Just found another annoyance that ought to be fixed. If you leave Delphi open overnight when Windows changes between DST and standard time, (for example, if your computer is "asleep" at the time and you turn it back on again after,) it "alerts" you that a whole bunch of your files have been changed and asks you to reload them. Repeatedly. That needs to get fixed.

The language needs a bit of work too. Object Pascal is a wonderful language, and I'd pick it over VB or one of C's misbegotten descendants any day, but it's got a few rough edges that need to be polished.

• Properties are wonderful. Why are we stuck with an implementation that's only halfway complete? Give me one good reason why you should ever be unable to pass a read/write property to a var parameter. If they both refer to the same data member, it's simple. If not, retrieve the read value and make a copy of it, pass the copy, get the result, and send it to the appropriate write.
• Likewise, array properties need to be fixed. Why should they require get/set methods instead of direct access? If I wanted to work with mandatory get/set methods to access the private members of my objects, I'd code in C++. I use Delphi to get away from that sort of syntactic diarrhea.
• Bring some of the syntactic sugar introduced in Oxygene/Prism over to native Delphi. Some of it's dependent on a managed code framework, but stuff like the colon operator and double comparison ("if 5 < x < 12") are just automagical compiler tricks, and I'd love to have access to them without all the overhead the CLR imposes. Also, if they could find some way to bring parallel FOR loops and Futures in without the CLR, I'd be ever-so-grateful.

Delphi's already the best there is. But fix the above things, and it would be so far ahead of the pack that people would consider it worth switching, instead of where it's currently at, struggling to maintain an ever-shrinking lead.

I learned Delphi for a job I held recently (for a short period). I honestly think that Win32 Delphi is the best natively compiled (ie, not Python, Java, or .NET/Mono) language in existence. If you compare a feature list of C# and Delphi (pioneered by the same intelligent being) you will be amazed at what Delphi does while still compiling to native code. The Algol/Pascal/Delphi syntax did not become popular, but I honestly think that it is a shame because it is very well structured. I'd much rather use Delphi than C++ for writing native apps. That said, I am a Linuxer, and these days I use Python, but if Free Pascal were a standard part of Linux distros, I would happily use it for non-trivial applications.

To answer the question: Linux support (not Kylix, that was a mistake) and less of an emphasis on .NET development or up-to-date .NET/Mono support would help. Unfortunately, I don't think anything will rescue Delphi at this time.

It's too bad. Delphi code compiles insanely fast compared to C++. It is a statically typed language with the option of having dynamic types (Variant) like Boo, and possibly C# 4.0. Really, this language was so far ahead of its time. People are excited now about getting dynamic typing in their non-dynamic languages, but Delphi had it going on well over a decade ago. Interfaces, abstract classes/methods, properties, etc. I could go on, but you get the point.

The original architect ^[1] of Delphi moved to Microsoft and created C#. So to answer your question, I guess he'd have to quit and go back to CodeGear.

When a product no longer has the founding team, it's very difficult to ever get the "magic" back. Corel acquired Ulead and now my favorite graphical editing software is slowly turning into crap.

Borland as we knew it is gone, and I've switched to C#. The size of the community alone makes all the difference in my daily work. Just look at the number of Delphi questions in here, for example. EEK!

Funny though, Delphi still rules Win32 in a world where Microsoft WPF makes people nervous. I still use Delphi 7 at work for our Windows client apps.

I think that is the key to a recovery. I really don't see anything else doing it at this point.

We all agree that Delphi remain a wonderful development tool.

Now Delphi needs StackOverflow: a good place to meet great Delphi developers !

We need to feel that we are still alive :-)

What is needed is:

1) Attitude - for companies to start adopting development tools by technical preference of their developers rather than following the political hype. We have stuck firmly with Delphi for 12 years, in spite of occasional doubtful looks from customers. This has resulted in significant performance and maturity in our codebase. Had we switched to the fashionable language du jour every 2 years, we would be nowhere near this level today. So yes, Delphi is "legacy" but also (for us) an indispensable basis for future advances.

2) Evangelism - show the youngsters today how Delphi completely outperforms the "big guys" in runtime performance, productivity, and community support. Keep the pricing competitive and make sure students do their first programming courses with Turbo Delphi.

3) CodeGear should go in and embrace the efforts of the FPC team to gain access to native 64-bit Windows, Linux and MacOS cross-compilation support.

4) Take the lead in support for Multicore processing (compiler level optimizations, framework library support, debugger improvements, design tools, etc).

5) More effort to be put into quality assurance tools specifically for Delphi - code analysis, coverage tracking (branch and line based), timing profiler, unit and UI testing, etc.

regards, Kristofer

I also think that they should forget about the .NET version, and focus on Win32.

A Linux compiler would be a great feature. I don't think a Linux IDE is necessary, just the compiler.

And I second targetting universities and schools. My first programming class was with Turbo Pascal, so adopting Delphi later was a natural path. Today kids are beggining with Java.

Delphi needs a technical overhaul as well as what others have alluded to, long-term plans and stability.

That an service pack upgrade to Delphi is followed by nothing less than an 18-point plan on how to manage the upgrade means that somewhere, somehow, a lot of people just didn't do their job.

Even Microsoft, who just released SP1 for Visual Studio 2008 covering many different editions in one service pack manages with a normal installation wizard that just does the job. Period.

This kind of problem is symptomatic to how Delphi has been created and managed over too long a time.

Delphi needs to be fixed:

• stability issues
• make the actual compiler and the intellisense processor behave the same way, or even just combine them, we have been forced to just ask the editor to forego any kind of advanced intellisense because it just plainly gets it wrong, every time. Things like red squigglies hovering in whitespace between methods and similar doesn't build confidence.
• make the entire eco-system more stable, the way component packages are loaded and built just means that the day just one of these files have a problem, Delphi simply gives up, entirely
• and then there's the entire long-term planning department, which seems to have taken a permanent hiatus. Stop bundling new products along with Delphi and start making the product itself better, then perhaps people would start taking it seriously enough to warrant long-term commitments to using it.

I believe the illustration shown above, with the miracle, describes the current state of affair exactly.

Delphi is a sinking ship, and nothing less than a miracle will save it. Good intentions and well-meaning is not enough.

I loved Delphi and developed in it for years...even attending several Borland conferences. They lost me when I realized (at one of the shows) that Inprise's management really had no interest at all in standard development and was hoping to pull an 'IBM' on the market: showcasing bloated "enterprise" application frameworks that were way overpriced. They had this jewel in Delphi but they decided it just didn't make them enough money right now so they starved it of the resources it needed (Delphi2005 anyone?) but kept pushing the price up anyhow.

I was also put off at the need to update my reporting tool and all of my third-party controls every time a new Delphi was released. It wasn't just the cost - although that was a factor - it was knowing that everything would break when I upgraded and that I'd have to spend days getting it running again.

I still maintain a quite successful Win32 product using Delphi7. Delphi 2009 just doesn't have any appeal - especially since I'm not even sure that the 3rd-part tools used in my app are still around.

Delphi was always the choice by developers in the know and clearly a superior development tool than Visual Studio in so many ways.

Then came .Net and the situation became much less clear. Borland put a lot of effort into making Delphi a .Net player, but by producing a product that was always going to be one or two steps behind Visual Studio .Net it was doomed. Visual Studio .Net was always the best choice for .Net development. Borland should have just produced a Delphi plug-in for Visual Studio instead of leaving it to RemObjects.

But now that CodeGear is in charge and free from the shackles of Borland, things seem a lot better. CodeGear has listened to customers who have always said that they want more work done on the native development side of Delphi.

This is were Delphi clearly scores, MicroSoft has now left Delphi as the primary tool for native Win32 (and hopefully soon Win64) development. For simple applications Delphi is a winner. This is why I am still using Delphi.

As an example we are having great success in deploying Delphi applications using thin client computers running XP embedded. These systems use flash drives rather than hard drives and are fanless and hence very quiet. The flash drives we use are 512 Mb so there is plenty of space to install a Delphi application on top of the small footprint of XP embedded.

I think Delphi will be around for many years to come.

1) Provide an educational version that's completely free.

2) A strong emphasis on stability. There are all too many of us who haven't upgraded because we don't like what we see.

3) The DOJ to do it's job on anti-trust. Microsoft's development tools aren't earning their keep.

4) While Kylix was a major error, making it possible to compile for a Linux target would be very worthwhile. While nobody is going to buy Linux tools that does not mean there isn't a market for a tool that does both Windows and Linux.

Delphi needs:

1. Better help & documentation. Abandon the MS Document Explorer. Go back to CHM or an in-the-IDE-html-browser. Better documentation means 100% coverage of the VCL, and RTL, with extensive examples and demos. Hundreds of demos of everything under the sun. Load a file, process text and parse expressions, process pictures, process audio, handle database tasks, everything.

2. Simplification. There are too many features in the IDE. The thing is a monster. A beast, a behemoth. Like Visual Studio, the learning curve on this thing is a b----.
   Remember Turbo Pascal? Remember Delphi 2.0? Every version of delphi has gained more and more features.

Unfortunately you probably can not take most of them out. Someone recently showed how to put Delphi into a "delphi 7 lookalike mode". I think that ought to be an option given to users out of the box "go to new user learning Delphi mode", versus "go to full interface/expert mode".

Delphi was harmed by many false starts. Delphi.net was a false start. The help debacle ever since Delphi 7 was a giant leap backwards. The move from a single product IDE (Delphi 7) to the "RAD Studio" environment, is now, in retrospect, a debacle. Almost 100% of Delphi users do not want or need any C++ Builder help in their IDE, and this ruins the experience for 90% of the users. C++ Builder ought to be a separate IDE, that shares common BPL files, but launches separately, with a separate help TOC and index.

I say one thing that Delphi needs to do is offer an express edition or something. At such a high price to even begin to learn the language, it's very discouraging for new developers.

Provide a way for me to learn it without having to skip around to old tutorials on the Web where I don't know if the author is teaching good technique appropriate for the current release. I want a book, or video, or anything that will help me understand the language and "the Delphi way", and one that's not several versions old.

I want something I can trust to teach the language, the common idioms, and if possible, tradeoffs for given techniques. I wouldn't mind a book directed at new programmers, but I want one that gives me some confidence that it’s not just an enthusiastic beginner's take on the language, which would be admirable, but may teach me bad form. An official or "CodeGear approved" book or video tutorial would help me commit and actually use the software I paid so much for, but have left sitting unused for lack of confidence. It seems like every book and tutorial I could find was two or more versions old, and since I'm new to Delphi, I don't know if learning from an older source would be the right way to do things on the current version.

A clean product with good help that simply works.

Price. With VS2008 Std at a price of less than US$250, Delphi would need to be significantly less expensive than $250 to get people to take a chance on it. The great thing about Turbo was that it was $50 when it came out and anyone could afford it to just play with. Delphi IIRC came out at $100 (probably wrong about that, but I'll get corrected :-), again cheap enough to take a chance on in the face of MS.

New developers - not sure how to attack this one, but there needs to be either a way to get it taught in HS or undergrad college -or- some really sexy version for game development to hook kids. That, and as mentioned before, some new, updated beginner books on using Delphi.

Just my (devalued) $0.02 worth.

There needs to be a Delphi site that promotes, sells and supports Delphi with only a passing reference to the other complementary codegear technologies. These should have sites of their own and all should have similar design to emphasise that all belong to the one family As a template for this codegear could do a lot worse than to emulate Realbasic ^[1]

Realbasic has a smaller userbase than Delphi but the forums ^[2] are very active, responsive, easy to use and tightly integrated with the site. The product is explained in simple english with introductions tailored for people with differing backgrounds and people new to programming ( see especially the demo video) It's simple to both try out and purchase the product from the Real site no matter where you are located in the world and the price is the same however far the download has to travel.

They have a free trial and a 90 day money back guarantee.

By contrast the codegear site is a mish-mash of design and content and arguably should carry a warning for epileptics. The different products serving different markets (interbase, jbuilder, delphi for php, blackfish) are lumped together as though they were interchangeable choices or were from an attic sale. Even within the object pascal delphi product there are a confusing array of offerings (Delphi 2007 for win32 R2, really slips off the tongue). Navigation around the site is difficult. Purchasing involves redirection to reseller sites where the product often needs to be located again and where prices are likely to differ. The forums being newsgroups(?) provide a barrier in themselves to people without a university background with horrible names like delphi non.technical. And to cap it all people now have to get their head around if Delphi is from codegear then what's that reference to embarcadero doing everywhere? Maybe it makes embarcadero feel good but it does nothing to sell the product

The impression given is that if you're new to programming then Delphi is much too complicated for you. And I would imagine that if you weren't new to programming but were having a first look at Delphi then you'd pretty quickly conclude that this crowd were clueless

The company behind Delphi isn't even able to produce a valid formal grammar ^[1] for the language. That means no third-party tools for code browsing, graphing, refactoring, static code property checking, etc. And the builtin Delphi capabilities are lacking.

I still use Delphi and intend to do so. The biggest benefits I find are:

1. Quick and easy deployment ( no need for .NET and GAC )
2. Easy to debug - No .NET stack overflows that are increasingly difficult to debug
3. Better documentation (and working examples) than MSDN
4. Several thousand working and tested components

I should point out my clients do not care for the latest fad., they just want something that works reliably and keeps up-to-date with their business. Delphi is still the fastest way to created reliable executables that I know about.

What Delphi needs is not more features but better marketing of their product. Focus on needs of large corporations rather than fads like RoR/BlackFish/.NET you can only sell a fad for a short amount of time.

Latino ^[1] - for Linux look at FreePascal and PowUtils

Effectively coding in Delphi is not easy since we're not taught it in school. To me, that's the biggest problem, no one is going to switch to Delphi and for the most part all us junior programmers are scared to go to a place where Delphi is the major programming language because about 2.5 percent of development houses use it and it seems to be useless experience.

However, Delphi is awesome, if you read the manual it gets all the more awesome. So if Delphi were taught as a benign non-controversial teaching language like it's predecessor Turbo Pascal was (I'm not sure if there no controversy here), it would be used way more than Java is.

Any info on 2009 would be appreciated, we're planning on updating from Delphi 7 to 2007 (and it may well be 2009 by the time we do).

There has been such a lack of focus (Kylix !) and stability in the Borland / Inprise / Borland / CodeGear / Embarcadero mess over the years that large companies were never going to invest in that technology suite, because who knew where it was going to be in six months time ? Even now, with the new owner in place, who knows where it's going to end up ?

I started using Delphi when D3 was the current version, and have since used 5, 7, 2005 and 2006. In that time, I evangelised somewhat to the VB developers that I was in fairly close contact with, when it really was an excellent alternative to VB (up to VB6 I guess). But corporates bought into MS technology then and they do now.

If Embarcadero are going to change Delphi's fortunes then there needs to be continued support of free versions (the Turbo range), somehow getting Delphi into the hearts and minds of new developers (get 'em while they're young !) and fix a problem that MS isn't fixing.

Oh, and Jim, I am using Delphi, because I love it ! And as of today, reading your blog too ;^)

Delphi needs compelling advantage over the existing ways to build apps on Windows. Since the .Net runtime is commonly found on client PC's, Delphi's advantage of being native feels less significant. At this point, Delphi provides a way to build Windows applications in a less-friendly and arguably less-productive manner than VB.Net, C#, IronPython, etc. It seems cornered into the niche of maintaining legacy applications.

What is needed to get Delphi back in the mainstream?

The hand of God?

I think Delphi was pretty nifty originally (as was their C++ equivalent). At this point though, I'm not sure what Delphi offers over comparable products except a smaller developer base. If you want to build for .NET, it seems that C# is the obvious choice. If you want to build native, you're probably going to go for C/C++

So what niche exactly does Delphi fill, other than the we've-already-got-this-code-written niche? And yes, that's an important niche, but it won't do much to get Delphi into the mainstream. It just relegates Delphi to legacy code. I don't think Borland wants Delphi to be the new Cobol.

For the record, I do wish that Borland was more competitive with Microsoft. It worries me that development tools for Windows are almost exclusively produced by Microsoft.

What's gone wrong with Delphi is a case of mismanagement. They shouldn't have released Kylix and they cannot compete with MS on the .NET front. Delphi 2005 was a disaster of a product and the damage had been done in previous with 8 and 2005 being the most notable.

They are moving in the right direction with releasing the turbo range for free but I think the best things they can do is target schools and universities and get people switched on sooner rather than later. Perhaps they could open some of their source?

Many of the comments have come from either fans or detractors. Let me say this as someone who strongly considered Delphi but decided against it because...

1. There were no up to date books. The learning technique seems to be: read a few old Marco Cantu PDF's then visit a half dozen web sites for the topics you need to cover. I think the last book I saw was Mastering Delphi 2005. The newer texts are geared toward Delphi veterans. It's hard to take a development tool seriously without at least one comprehensive text on the subject.

2. There hasn't been a suitable trial version. I just checked their site and noticed that they finally have one. Nothing I could find said how long the trial lasted. (IMHO, trial versions for IDE's should really be at least 90 days.) Still, a low cost lite/express version would be nice.

I already answered this once, about IDE problems and language features, but there's something else Delphi needs, from a more psychological angle.

I attended a Delphi users' group meeting last week, and if I wasn't the youngest guy in the room I was pretty close to it, and I'm closer to 30 than I am to 20. Young programmers are the future of programming, and they're all learning the C syntax family in school.

Delphi needs to start attracting high school-age hobbyist programmers! Add in a good, solid game-development framework like Asphyre ^[1] to the standard Delphi package, release it as part of a free hobbyist edition, (like Turbo Delphi of a few years ago,) and let all the kids start to learn how easy it is to have fun with programming.

Microsoft understands this. Just look at Visual Studio. There was really no reason for them to put VB (pre-.NET) into VS. It was just a toy language, without enough power to justify placing it at parity with "serious" languages like C++. But from a psychological perspective it was a brilliant move.

How many of you out there first got into coding with those little "write games in BASIC" books? I know I did! VB drew those kids in as they grew older and brought them around to the Visual Studio way of coding. I'd have been the same way if a high school friend hadn't introduced me to a brand new "visual Pascal" program called Delphi. Embarcadero needs game-development tools to get young programmers on board. And then, as they get a bit older, they realize that oh by the way, this fun language they've been playing around with can also do heavy, serious stuff like multitier databases and VOIP programs.

Marketing, Positioning and Perception. Code Gear tends to send mixed messages when it talks up Delphi. I recently had this discussion with Nick Hodges and expressed my concern that Code Gear is missing a critical opportunity to get a strong, consistent message out there.

Delphi is not C# and it never will be. The fundamental difference is that Delphi provides developers who need to write real code with one of the best tools available. Other platforms such as C# turn the developer into an assembler, developing applications from other people's code.

I think that Code Gear is starting to realize that it's market is in native code development and has recovered significantly with the release of Delphi 2007/2009. The waters were muddied with Delphi 8.

Delphi's strength is in native code and Code Gear's marketing should aim to educate the development community that core code will always be required, despite Microsoft's attempts to convince developers otherwise.

Delphi puts the real power of native coding in the hands of the developer whilst MS keeps that power for themselves and relegates 'developers' into the role of assemblers.

Code Gear should be working hard to establish itself as one of the leaders in this distinctly differentiated field of development.

MS promote C# as a develpment platform that uses MS-magic to make things happen (check out the new features in C# 4.0). Code Gear should be promoting Delphi as the platform that is used to develop the 'magic'.

i know im too late, but we need books on delphi ,just check amazon for books on java, c, or even python, perl and compare 'em with books dedicated to delphi .

The biggest thing I can think of is cross-platform support, and standardizing the language extensions they have made with ISO, etc.

The reason I use Pascal over C in general (besides its greater readability and dislike of having to deal with pointers for everything) is because it has a bigger standard library. That means I can write something once and compile it on multiple platforms. I have the speed and simplicity advantages of something like C, with the run-anywhere abilities of Java or C#. (yes, it has to be recompiled, but that's not such a big deal most of the time).

Turbo Pascal was taught in schools because it was a good implementation of an open language (Pascal). Borland briefly called the Delphi language ObjectPascal, but reverted to calling it "the Delphi language" in more recent versions, and so far as I know hasn't pushed to standardize it.

On the other hand, Delphi compatible compilers are available for many platforms, with the most prolific being FreePascal. In fact, FreePascal will run on PowerPC, Intel, and other processors, bringing its support close to C. The other half of cross-platform support is the GUI - which FreePascal is reasonably good at as well. Whereas Kylix required QT to be installed on Windows if you used CLX, Lazarus allows you to compile native to GTK, Win32, or even Mac OS GUI these days.

Although Delphi is a little more polished, I mainly use Lazarus these days for exactly the reasons listed above. In fact, if you want to write a GUI program that runs natively on Linux, Mac OS, and Windows - what other good options do you have? For individual programmers, I highly recommend looking into Lazarus. It's true that the debugger is still sub-par, but things have come a long way.

Delphi should catch up by taking the strengths of Microsoft's tools and using the tradeoffs made as weaknesses. Where Microsoft tools support only Windows, Delphi doesn't have to. Where Microsoft's tools support "managed code", Delphi's can generate native code. Microsoft can tightly integrate into their own codebase, whereas Delphi can offer better support for Open Standards.

a few answers here mention advertising.

to embarcadero : why not consider a delphi sponsored tag in "StakOverflow" ?

Of course it's not the most important thing needed for Delphi,
but little things can sum up to something Big...

As a Delphi developer, I'd say focus on Win32 development. Borland/Inprise/Borland - Codegear/Embarcadero really lost their focus when they decided to do Delphi for .NET. It was and is guaranteed to lag behind MSFT's offerings in that area.

Now that Microsoft is focused on .NET and web development, there is an opening for truly great Win32 environments. Even though that's quickly becoming a niche market, it's one they could dominate.

Delphi needs us - me and you (and them). Having been to some Delphi presentations, it is great to see the product in the hands of people that actually seem to care about it, and who look like they have hitched their wagon to Delphi and have actually developed the product in the right direction. The next release (2011 maybe?) should push things again in the right direction.

NOT .net or JVM support. Delphi was great as it was a faster way to build GUI apps that could be statically linked into one executable (more or less). I would like to see it go back to that. It had an awesome "rad" environment too (man that term has gone out of vogue).

I think it should play to its (historical) strengths. There aren't many tools on windows where you can build an executable and ship it and it will just work anymore... delphi was one of the greats.

In our case? An inexpensive, low-end version of Delphi, with prices prominently displayed on the website.

We use InnoSetup, an open source installer package written in Delphi, and we'd really like to add some new features to it, and send our patches to the author.

Back when Delphi was Inprise, it cost something like $800. Today, the website doesn't show any prices ^[1]. This usually suggests that either (1) Delphi is embarrassingly expensive, or (2) a salesperson wants to talk to me on the phone and figure out how much money I've got. Then, when I say no, the salesperson will call me back every 3 months.

Microsoft has quite a few inexpensive, low-end options for C# and C++. If Delphi is going to survive as anything but a niche product for legacy enterprise code, it will need to do the same.

What would not hurt Delphi is frankly a free version with sqlite database. I started using delphi when Delphi 1 was in beta and totally loved it. I stopped using Delphi for totally unrelated issues when Delphi was in version 5. I might be going back soon. But a free version doesn't hurt.

What about a Mac version ?

Delphi needs a couple of things:

1. Price and confidence
   These two go together. You simply cannot charge so high price for something without solid future track.(Borland-Inprise-Borland-selling-not selling-CodeGear-selling-Embarcadero). No software house is going to spend so much money on the risky product!

2. Innovative product
   Programmers need to develop application in less time. They need to have reason to upgrade to the new Delphi version. => Buy DeveXpress and offer all their VCL stuff in new Delphi. Come up with some Delphi framework... simply do something what give us a strong argument over .NET Visual Studio.

3. Community
   Borland simply did not appreciate the community enough. The best way about Delphi was to use components by someone else. I didn't need to develop my own components. I either bought or simply took the code of someone else. If I needed help... I used google and I got the answer... It's changing now. All these developers are going to C# and their post are desperately old. :(

Support for more operating systems and architectures. Support open-source communities financially. Increase media exposure.

And educate the people that Pascal is powerful.

Also, there really should be more articles on Wikipedia about Pascal, Object Pascal, and Delphi, and Embarcadero.

Maybe they should rehire Anders Hejlsberg.

Unicode VCL, Generics, new Datasnap enhancements etc. was right things although much late. I believe if they can have progress like 2007->2009 for two more Delphi versions, they can be on track again. They lost many years around D8-D2006 but I appreciate their progress with D2007 and especially D2009 and I know there are still need for native apps so I think, Delphi might not be mainstream (it never was, anyway) but can be a strong alternative again, if they will do keep doing right things. IMHO, they just need x64 and more support for multicore and web-enabled ,distributed applications.

Delphi Needs great help (D2009 getting better) and lots of supplied demos. Although there are lots of stuff on the 'net nice readable code should be available for download from Codegear and which compiles and works. Take the PIC micro as an example. It sells in millions due to lots of simple examples of how to use it.

Delphi could use a few more tools that are common-place outside the Delphi community :

• NDepend ^[1] (dependancy analyzer, .Net only)
• Valgrind ^[2] (dynamic code analysis, Unix only - Valgrind can be used by FPC ^[3] though!)

Both very powerfull tools.

Sure, we've got Code Healer ^[4], Pascal Analyzer ^[5] and AQtime ^[6] amongst others, but they are not common-place (yet - maybe because they are commercial instead of free software? I don't know.)

Free version.

i've used Delphi 5 for the last 10 years, but now we're switching to Visual Studio C#.

Delphi is a clean language. The compiler is much faster than VS, and is more stable. Its class library source is open. It has had an active and helpful forum community.

But recent versions of Embargadero Borland Codegear Developer Studio are very Visual Stuido-esque. It's slower, bigger, bulkier. The open and thriving community forums are now closed.

And if i'm going to use a Visual Studio clone, why not just use the original? Especially when the original is free?

Perhaps we would come back if Borland stopped spitting on it's customers.

What I don't understand is why Delphi is always singled out, like it is here, and why C++ Builder hardly gets the same attention. Truth be known that both Delphi and C++ Builder use the same underlying VCL. And the IDE for Delphi is the same as that of C++ Builder with the property editors, controls, and drag-n-drop components.

My company uses C++ Builder almost exclusively. We even developed our own visual layout tool for our accounting system so that we can build or change screens on-the-fly. We do this by reading in and streaming the VCL DFM files and connecting the resulting GUI to an embedded Perl interpreter.

I can't say enough good things about the VCL. My only wish is for a Linux port such that I can just recompile for Linux. Kylix was a disaster....

my 2 cents ;)

A VB6 to Delphi converter and allow Delphi to support VB6 third party controls.

From my limited experience, VB6 is a lot more similar to Delphi than it is to .net. So, if folks can create a VB6 to VB.net converter then perhaps someone can create a VB6 to Delphi converter. Not an easy task. But an easy sell if it works. And that's the kinda of technology you want to work on, right? A technological challenge with high marketability.

I think the new managmenet with Embarcadero is a big plus. There is still a market for non MS tools and I think we will see Delphi gain ground. While it has a popular following outside of the US, it definitly needs to gain steam on the US side. IMO Delphi needs to:

1. Continue to publish their roadmaps and stick to them
2. Quality releases
3. Bypass .Net and stay native
4. 64 bit processing
5. Take advantage of Embarcadero tools and integrate them into a future IDE.
6. Tighter web development integration. Maybe VCL for Google apps.

Delphi has a large base and the first two items should allow them to focus on preventign the base from migrating to .Net (it is insane in my opinion to move a large Delphi system to .Net from a cost perspective). As a long time user I see a very positive future with the Embaracero purchase.

There are places where I find Delphi to be a better tool than C#/VB.NET. For writing Win32 services, Delphi works better for me. The footprint of the .NET Framework is huge and for many cases, just overkill for a service application. I have written a few services with Delphi and it's a really good tool for that sort of work.

Delphi 2009 finally adds unicode to the VCL, that should make it more attractive to the world market. I was glad to see that the 2009 version is Win32 only, I really had no use for a non Micrsoft .NET IDE. I think Delphi will have it's place, but it's just going to be a tiny sliver of the market place.

Let's see what the Embarcadero era will bring. It is no secret that the Inprise/Borland period was a period of missing focus when it came to the Delphi/C++Builder tools.

Having used Turbo Pascal and Delphi since their conception, I still haven't managed to find other tools that equally rich in function, performance and ease of use. It is even harder to find a language that surpass Delphi code in readability and maintainability.

Yes - there are areas where Java, Python, Ruby, C# or even C++ work better, but if you are doing native Windows applications - there simply is no better tool available. It is also a glittering tool for writing data servers.

With Delphi 2009 and Generics, Unicode and other assorted tidbits - Delphi has enough vitality to hold it's own against the steady flow of "new and shiny". With the next versions heading for parallelism and 64-bit - the future looks solid.

Some tool designs just keep working and cuts through your work.
Delphi is nostalgic in the same way as a knife is nostalgic.

People, Please..... You can upgrade from really old versions of Delphi (5) and the upgrade price for the pro version is less than 400.

Delphi is not dead, it's still better than C# and winforms for desktop applications and developing database apps is about 100 times easier and more intuitive.
If your using C# instead you are missing out BIG TIME. C# users are like lemmings.......

Codegear just need to be a bit quicker on their feet.

Clearly Microsoft have as good as abandoned customers developing native code, this creates a good opportunity for an alternative. However we have only just got support for Unicode and there still isn't any 64 bit support.

Another thing that would change Delphi's status, is if universities went back to teach Pascal. If the next current crop of fresh programmers would be weaned on Pascal, end enjoy that language, its syntax, and its idiosyncrasies, Delphi would become a more natural choice. I dont know of any major (or not major) university that has been teaching pascal as the main language in years... in fact, havent seen it taught formally since I was in high school. Everyone is teaching more C-type-syntax languages now (C/C++/C#/Java/Perl/etc).

Maybe if Delphi would stop being supported altogether, Pascal would be considered a historical language and could be taught instead of LISP ;-)

Seems that if they could cater to a missing feature in the Microsoft toolkit, they could gain some market share (Such as making UI's unit testable).

I originally used Delphi because VB wasn't ready for prime time outside of IT apps, and C++ MFC was a real PITA to develop UIs with.

64 Bit Compiler, please

Give the compiler away for free. Then, sell the IDE.

Invest in the compiler, trying to reach for every platform posible, prioritizing each one: i.e. x86 for windows, mac, (maybe) linux, then ARM, etc.

Also, take a strong bet for Linux. Linux still lacks of a real visual and RAD development platform for the masses. Is there a way of making Delphi to be a real portable (compilable everywhere) platform, including compiler, IDE and libraries?

Kylix initiative was partially the way to go, that was the cause of its forced failure.

What are people referring to when they use the word Delphi?

• Are they referring to the Pascal language, the Object Pascal dialect in particular?

• Are they referring to the facilities in the IDE?

• Are they referring to the libraries that come with it?

• Are they referring to the use of components?

• Are they referring to its static typing?

• Are they referring to the need to go through the compile, debug edit cycle that is required at every bug?

• Are they referring to its non interpteted nature?

• Are they referring to its use of compiled code?

• What exactly do people mean when the say Delphi?

Cut through the sarcasm and smart remarks... there was something actually intelligent said in the 10 messages:

"The desktop is not where it's happening anymore".

That may be true. Remember, though, that today's big web apps are not developed by single developers anymore. They're centered around large databases and require large teams of developers to build and support them. Often this involves PC desktop support apps that aid in loading, data modification, testing and diagnostics. Some of these systems get their data from scanning and coding operations that use desktop apps to drive the process. As long as there are PCs there will be a need for PC applications.

Remember also that the desktop app is the simplest way for a beginner to learn programming.

Delphi needs to market to the grass roots again. Their biggest problem is a total focus on industry. Turbo Pascal became popular when it sold in bookstores for $50. When Pascal was declared dead in 1990, Delphi made a huge comeback with a product that offered a "personal" edition for $99 in computer stores. In both cases they mass-marketed to the general public. Selling the cheapest option in your suite for $1,999 is not the answer.

Fortunately, the latest incarnation has a "starter" option for $199 (new user), $149 (upgrade). This is certainly more reasonable than recent releases, and may win back some popular support. Still, I'd offer a freebee download or a dirt cheap personal edition if I were them.

And Delphi does keep up with modern technology quite nicely. Their latest suite builds applications for tablets, cell phones and other handheld devices as well as PCs, and they do offer new tools for building web applications. So it is not totally "desktop-centric" as some contend.

Delphi is as powerful as any other language out there, and is easier to use than most. It builds applications for both native Win32 and .NET, and has growth potential to include Linux and Mac; something Microsoft will NEVER do!

It has all of the constructs of a modern programming language, with the option to still use old-style constructs when it's more practical to do so. You don't have to use templates, interfaces and object lists just because they're there; sometimes it makes sense to use an old fashion array or record.

Don't think of it as an out-dated product that had its day. Think of it as a fine wine that's aged and seasoned.

It's too late for Delphi, which is sad as I love working in it.

1. Kylix was a disaster. No one will pay for Linux developer tools.
2. The .net versions are redundant and only usable for transition to Microsoft's .net tools. Company name changes and being for sale with no buyer give a feeling of instability.
3. Every place I have worked that used Delphi has completed or is in the process of moving to .net.

What I believe Delphi needs most urgently

• Rock solid IDE.
• Radically cut prices
• Native 64bit
• Good documentation (it amazes me how bad it is in D2006)
• LINQ alike features and syntax.
• Competitive multithread solutions.
• A completely free for all usage lite-version (like VS Express)

All this is needed just to make Delphi on pair with Visual Studio. After that they need to start innovate if they want to reach the top.

Some things that could make Delphi reach the top

• Some kind of cash refund if the IDE ever crashes.
• Everything .NET has but native.
• Some way of directly being able to use manged .NET code from inside native Delphi.
• Advertising, convincing old custumers that Delphi is back on track.
• Cross platform support. Not just Linux and Mac. All sort of embedded OS and small microprocessors. Mobile phones, Ipods, toasters, you name it! (skip the VCL if it can't be ported).
• More luck than mankind has ever seen.

Also, I think moving Delphi.NET to Visual Studio would be a good idea.

I think it's telling that (as far as I can see ) no CodeGear employee answered on this question up untill now;

To me this is a clear sign that they have full confidence in their product!

I've been an enthusiastic Delphi 6 user. I've switched to C#/C++ because my job requires me to use those tow.

I've been playing with D2009 for two days now and I think it's fantastic. However, I'll probably not buy, the steep pricetag beiing the main reason. So what would bring me back?

• lower pricetag
• native Linux support
• native Win64 support

I'm not that fond of .NET as everybody else seems to be. Native is the way to go.

Borland lost us with Delphi 6. Our million+ line application, which has successfully been transitioned from 2 to 3 to 4 to 5, would not longer compile on D6, without drastic code rewrites. Someone thought it would be a good idea to break DesignIntf, but it cost Borland in the end. Too bad, since Delphi is still to me the best way to build a Win32 application, but we are currently slowly moving from Delphi5 to VC++. In my opinion, nothing will bring back Delphi.

I still using delphi 6 on some of our products.

i belive .net delphi compiler must be free, maybe core part of .net framework, but free. And anyone with a notepad shoud be able to implement a delphi solution without paying for de ide. and anyone who bougth visual studio shoud be able to develop a delphi solution.

besides love to the language, there aren't any other reasons to develop a delphi application instead of a c# application.

borland folks just destroyed delphi. maybe is just time to let it die

Moving back in the direction of cross-platform support.

Back when I was a Delphi programmer, our desktop app had been written in Delphi mainly because at the time (the mid 90s) it was the only decent win32 RAD tool on the market. As things moved on, they stayed with Delphi mainly because it wasn't Microsoft under the hood - knowing the libraries and widgets and whatnot didn't come out of Redmond gave us an extra level of confidence that it would actually work.

Also, we were in the rare position of if we could have gotten a Linux version out the door, we probably would have doubled our annual sales. Even though Kylix never actually worked, the promise of of being able to essentially just recompile our win32 GUI app for linux was the best thing we'd ever heard since we didn't have anywhere near the manpower to rewrite the thing.

It was a coincidence that the company went out of business about the same time Kylix was officially canceled and Delphi moved to .net, but it sure felt appropriate.

But seriously - if Delphi rolled in transparent support for Mono (or something) and went back to being three years ahead of Microsoft instead of two behind, I think their market share would come back overnight.

Delphi Prism already has some sweet language features that hardly anybody knows about, like the colon operator for avoiding nil checks, expressions in property read/write clauses, and the Observer-pattern "notify;" modifier for properties.

Put some of those features into Win32 Delphi. The colon operator would be a fantastic addition to the language. Property read/write expressions, same deal. Add an Observer pattern to the VCL that comes with built-in language support that makes it dead simple to do it right.

Steal some ideas from C# 4, too, like passing named parameters (something I've been missing in Delphi ever since I saw Smalltalk). And for God's sake, give me a "using" block to replace the tired old "try..finally..FreeAndNil".

In other words, pour on the syntactic sugar.

Then hype those features. Blog about them. Put how-to videos about them on the front page of codegear.com. Give conference sessions (and keynotes) that are just about language goodness. Put a low-cost compiler out there, lurk in C# forums, and tell people "Oh, that would be easy with Delphi, maybe you should check it out." Get mindshare. Show people that the language has built-in support for the things they bang their heads against on a daily basis.

Microsoft does a release every two or three years, and they're very cautious about adding new language features. CodeGear could (and should) run circles around them.

Two important aspects to stay completive are support and marketing. Embarcadero has done a good job at both in a short time.

Another aspect is the integration with DBs. An integration as tight as IBX is needed for popular databases such as Oracle. Embarcadero can also take advantage of being a specialist at database tools and increase Interbase's presence in the market by providing a high quality version. In effect, giving people an alternative which is less complex and less costly than the competition.

It's still early to judge things. It sure is getting better and better.

Embarcadero has bright people working for a bright future. I will stick to Delphi and Interbase for now.

faster IDE, 64 bit support.

Used Delphi from 1995 but switched to Visual Basic 2008 C#/.NET because Delphi could not consume this .NET/C# SaaS applik from e-conomic

https://www.e-conomic.com/secure/api1/EconomicWebService.asmx?WSDL

When the Delphi environment gets the opportunity to consume that webservice I will consider switching back.

The documentation is like MSDN: Sometimes over-explains easy and normal stuff that barely you care but the important issues just left a single cryptic line.

Add multi-plataform support (Delphi prism could be a start but there's something about .NET that makes it so slow so buggy and only supports Windows fully, GNU/Linux has limitations and some bugs) There's Lazarus and wxForms (wxWidgets for Delphi) but that's unofficial.

• The language is pretty good, Delphi can do almost everything what C++ can with less effort, even supports operator overloading!!! (I know some JAVA developers that would appreciate that)

• Supports components for 3D without having to know what an interface is (glScene) and without dealing with the DirectX.

• There's a free IDE (the turbo 2006 following the microsoft's way) for commercial projects

• Together integration (another Borland's revolutionary tool) for visual code generation and auto documentation.

...Sometimes i think Borland started hiring Microsoft's employees since KYLIX announcement, my fears were confirmed since i saw the .NET requirements for just running the IDE (i would prefer JAVA... ¿does this means that IDE was made with Microsoft's Visual J#?)

At the end i think Delphi's future will be: 1. Microsoft Delphi(tm) once Embarcadero finish the Delphi Microsoftization (this term doesn't exist... officially) and like C# (without Native support). 2. Lazarus, the last hope for those in need of stability (well... lazarus has it's own problems too).

And for Builder and KYLIX... that's just another sad story. There are less demos and docs in builder than Delphi.

Why should I? Specifically, what does it offer? At best, it looks like an average language: not the highest-level, not the fastest, not the most standardized, not the most portable, not the simplest. For anything I want to do, it seems there's a better option.

Also, it's not entirely clear what the situation is with free compilers. They exist, but seem to lag on features/compatibility. I'm with Bjarne on this one: "It will take a lot to persuade me that the world needs yet another proprietary language".

So the answer is to fix these things, i.e.:

1. make the compiler free, or at least the spec
2. find a niche, and become great there

For #2, that niche really can't be "desktop Windows apps". I don't know that any third-party language/compiler/toolset has really been able to hold onto "native $(OS) developer platform", on any platform, for very long. The people building the platform aren't going to depend on a third-party, so they'll either buy you, or squash you.

Delphi is a very good product, but with to less unique selling points. I think, a Kylix for Mac OSX would be a such one. If a software developer has the possibilty to add 10% Mac users to his customer base by using Delphi, then this is a strong argument.

Loved the old Borland when they were doing Turbo / Borland Pascal. The compiler was tight for the days. The executable was small, compact, and it ran fast vs something from other vendors that produced bloated executables. Barrier to entry of using their product at the time was pretty low, price-wise, which was a good way to attract new / existing developers. Borland nowadays is just a pale shade of what they used to be. Go back to the old days and maybe they can success in the IDE / SDLC tools / Language market again.

Delphi was a good language design for early 1990s.

Over time, it added more features to try to stay current. By 2002 or so it had some Byzantine and messy parts. I know this because I built a parser for Delphi source. I had to do it by trial and error, since there is no complete formal definition of the language grammar. People still come with corner cases that Delphi compiles but my parser doesn't.

Delphi's value proposition was that it has all the advantages of C++ as a Win32 programming environment, without the extra complexity. That's if you regarded Multiple Inheritance, templates etc as needless complexity, which quite few people did. But Java and .Net fill that niche, and other niches, and also does away with the memory management complexity that Delphi had, and allows Generics as a simpler form of templates. And runtime metadata, and some platform portability, etc.

C# was a good language design for early 2000s.

It too is adding stuff, but seems to be handing growth a bit better, so it may last longer than Delphi. However, it is inevitable that it too will be superseded sooner or later.

Delphi's remaining niche is for maintaining existing Delphi code, and possibly for Win32 applications that for some reason can't be .Net applications. Delphi.net never had much of a niche - if you're going to write for .net, why not learn C#, or if you know VB, use that?

What is needed is for you to get over your nostalgia. Delphi has had its day.

For one they can release and maintain a viable *nix version of their development suite.

Another, they can make it less bulky. The amount of disk space required for .NET and Visual Studio is ridiculous, and Code Gear has certainly followed suit with their enterprise-y crap.

Make it less enterprise-y. Small, fast and capable compilers and IDEs.

That would get me to switch/stay. Lazarus and Free Pascal just aren't there (for me) yet.

I used Delphi quite a few years ago, back when Borland owned it. 1999/2000 or so. I was just learning programming beyond QB and was trying to figure things out.

I found it very much superior to VB6, but the Pascal syntax bugged me and I thought it unintuitive compared to C++, even VC++ 5( Pascal syntax still bugs me).

For me to choose Delphi today for a new project compared to C# would require several things:

1. Speed of software. I use emacs. It's fast and lean for an app today. I like that.
2. Straightforward libraries, both for UI components and for database access. I've been 100% unhappy with .NET database access and display on screen and consider it a complete disaster of architecture astronauts compared to, say, Perl's DBI or PHP's builtins. I would switch to Delphi very fast if they could make native W64/.NET DB/UI access trivially easy to do.
3. A CLI for scripting. I use make for my compiling projects.
4. A stable corporate situation where bugs would be fixed, software maintained, etc. That's a big advantage when the software isn't open-source.

Arguably, if a Free Delphi non-commercial CLI-only version was available, it would provide a tremendous edge in breaking into the marketshare again. Certainly I'd give it a spin.

IMO Embarcadero needs to move away from being desktop/Win32 centric. Support for XML, generic webservices, ISAPI etc (particularly in the areas of mapping and debugging) in Delphi are light years behind MS.

Borland dropped the ball BIG TIME back in 2000-2002 - they had the ability to conquer the world with Intraweb and their early support for ISAPI, webServices, etc but they went off in the wrong direction (NO direction...)

Embarcedaro needs to understand that the desktop isn't where it's happening any more - Current Delphi support for XML, webServices, ISAPI, Intraweb,etc (particularly with respect to mapping and debugging) even in XE is light years behind what MS has done with C#/.NET. IntraWeb could have been THE KILLER APP for internet development but it languished and remained buggy and with very poor documentation and support, so much so that it was virtually ignored.

Even though I'm a Borland fan and Delphi geek since Delphi 1.0, I'm afraid it's going to be too little, too late, for Delphi to be anything more than a niche/legacy product now. The world has moved on, Delphi has not.

Bribery? End times? Hurricanes?

I loved Delphi years ago. When what seemed to be the most "mainstream" choices were C++, VB, or Delphi (not that it was really ever mainstream), Delphi was the cat's meow.

However, I do not miss Delphi for a second. At all. There was something about it that never seemed quite right. Weirdness with windowing on Windows (hehe) and seemed to wreak of an ocean of badly-designed apps.

As much as I loved Delphi back in the day, I couldn't care less what happens to it now that I have my beloved C#.

Delphi's time has been and gone. The programming world has moved on, and is moving ever further away from it. It was a great tool in its day, but that day was yesterday.

And I think it's funny that people are still suggesting a *nix version of Delphi after the Kylix fiasco. What's that old saying about those that refuse to learn from history are doomed to repeat it?

What the company could do, and it would be a game changer is to support C# for VCL as a third language. There is a market for win32 apps, and C# for the moment is superior language.

I don't use Delphi because my dad used Delphi...

You might as well wish for the return of PowerBuilder, Gupta SQL Windows, and, while we're at it, Lotus 123. Delphi, while it was great during its peak, is forever lost to its better managed, better marketed, and more widely known competitors. RIP Delphi.