Is Linux Ready For Delphi? -- Delphi R&D Answers

Chrismo writes: "Danny Thorpe, Sr. R&D Engineer at Borland, has written a great article addressing "some of the commonly expressed fears, misconceptions, and even misplaced euphoria" heard from the Delphi community since the announcement of their move to support Linux with their development tools."

Comments

More apps!

[RancidPickle]

Great! Having Delphi for Linux will introduce more folks who program/use Delphi to produce apps for Linux. More talented folks entering the Linux camp! Anything that creates more apps for Linux should be welcome news. Hopefully, the R&D division will get it right. Sounds promising, anyway.

Re:More apps!

[uradu]

I think you're trying to force a fixed point of view on the whole matter. Why would Borland have to do any conversion of any sort, either of Linux hackers to Delphi, or of Windows programmers to Linux? While some of that might indeed happen, Delphi can stand on its own legs. It is one of the most elegant development environments and languages around. And it generates commercial grade code, to boot.

Think of it this way: if Delphi for Windows had come to market a couple of years earlier, VB would have never stood a chance. Delphi offers all the ease of use and RADness of VB, while using a potent language that can scale to large projects--something VB certainly can't claim. And it creates real stand-along Windows executables; no p-code, no support DLLs etc. But since VB came first, the hordes of amateurs that jumped on it never had a reason to switch, and they went forth and multiplied. Delphi never really stood a chance.

What Borland has now is a second chance. They are going back in time to the beginnings of a new OS platform and are the first to offer a powerful RAD tool. They can then benefit from a new generation of start-up programmers that jump into Delphi because it's the only choice--like VB back then. Once they become proficient at it, they will never have a reason to switch, even if Microsoft should one day decide to port VB (ha, ha). It's all about money and leverage, not conversion. Don't forget, there are constantly new generations of programmers arising. It's just a matter of catching a big wave of them.

As an aside, I used to be a long-time C/C++ programmer (well, long-time considering my age anyway). I gave Delphi the cold shoulder for a long time--it was one of those things I was going to check out some day, just not today. One day I finally did, and I never turned back. I certainly can't claim that Object Pascal is a superior language to C++. Sometimes I miss templates, very rarely multiple inheritance, never C++'s awkward strings. It lacks the succinctness of C++, and yet I find myself going back to old code and understanding it easier than I ever did old C++ code. Delphi is a Roger Moore to C++'s Shawn Connery--less flashy, more understated, most refined. On Windows it makes working with COM a pleasure--something that can't be said about C++. In the end, I keep finding myself completing projects before my C++ brethern, especially those saddled with VC++.

Uwe Wolfgang Radu

kudos to Inprise

[eries]

Really thoughtful piece. This guy knows his history, and makes some really interesting comparisons between Linux now and the early days of Windows. Too bad more companies don't approach their marketing strategy that way.

Want to work at Transmeta? Hedgefund.net? Priceline?

A Question For Delphi Pros

[quakeaddict]

On Win32 Delphi uses COM to a great extent. How will that translate to Linux? Will they port COM to Linux? Will they jump to Corba? Will they play nice with Gnome/KDE?

Thanks.

Re:A Question For Delphi Pros

[StrawberryFrog]

> On Win32 Delphi uses COM to a great extent. How will that translate to Linux?

I wouldn't say great extent. But COM will not be supported directly under Linux.

Look at the Borland chats, particularly this one on Delphi R&D: http://community.borland.com/article/1,1410,10459, 00.html)
Q: Is COM/DCOM going to be supported in Kylix? What about the future of MIDAS & Kylix?
A: We won't be directly supporting COM and DCOM under Linux. We will have MIDAS solutions on Linux.

> Will they jump to Corba?

Delphi/Win32 already has some CORBA support. IMHO that will most likely carry over.

Also see the live chat on the topic of ADO: http://community.borland.com/article/1,1410,20067, 00.html
Q: Any idea of how Delphi ADO applications might convert to the forthcoming Linux version of Delphi?
A: Unless MS provides ADO for Linux, we don't anticipate providing ADO supporting components on Linux.

Background: ADO is a back-end independant database connectivity package made by MS for MS operating systems. It is implemented as COM interfaces. Therefor chances of MS providing ADO for Linux are effectively nil.

Second article in a row!

[scumdamn]

Again I'm impressed, and I'm not easily impressed. (Simpson's reference: Homer: Wow! A blue car!) This is a thoughtful and well articulated article by a non-marketriod, non-Linux zealot that I feel has increased my understanding of things, has kept me occupied for more than ten minutes, and made me think.

For those of you complaining of lack of "News for Nerds" and "Stuff that matters." you have no farther to look than here.

One thing that occurs to me

[IntlHarvester]

Is that Delphi being available for Linux might actually spur the sales of their Windows product quite a bit.

At this point, more savvy IT departments are aware of Linux, but the common perception is that "it's not ready yet" (or is it "we're not ready yet"?). But, even out on the horizon, the existence of alternative environments has to cast some doubt on the typical Windows Standard Environment policy in place at most shops.

So, when a large internal RAD project comes down the pipe, even if 100% of the clients are currently Windows systems, Delphi may look like a better choice than Visual Basic simply because Borland is willing to consider popular non-Windows platforms. It at least gives you long term options -- in 2002, you'd hate to be the person to tell the CIO that Linux on the desktop is impossible because your department has just completed two million dollars of VB development -- a much worse problem than some poorly converted DOC files.

I'd expect there will be quite a few "We get along fine with gcc and vi, thank you." posts, but keep in mind this move isn't really about you. It's about the people on the S curve the author talked about (who probably care as much about you as you care about them), and my guess that in the short term, it's really about providing a more attractive product to their Windows customers.
--

Re:One thing that occurs to me

[jkaster]

I'd say you're right. Delphi for Windows sales have already increased because of our plans for Delphi for Linux.

Re:Pascal is dead

[TheKodiak]

You know, I hated Pascal all through my DOS days. I kept hating it out of habit until Delphi came back. Delphi has so _totally_ changed my mind - C/C++ is great for the hardcore who either write entire projects by themselves, or who know how to write a comment, but for working on projects with many contributors, lots of whom hate writing comments, Pascal borders on self-documenting.

I know this sounds lame, but...

[god_of_the_machine]

I am VB programmer... despite the bad publicity the language gets, I was born and raised on GW-Basic, and then moved to VB way back in the old Win16 days. Now, I'm doing more and more (non-programming) work in Linux, but I can't start developing for it because I am too lazy to learn C++ when I can do a bloody good job in VB.

I think that it's great that Delphi(Object Pascal) is being ported to Linux -- how about an open-source version of a basic language so folks like me can start developing? With an open source Object Basic tool I would start porting my programs like crazy -- as would a million other VB programmers.

Before dismissing me completely, just consider that not everyone has to be a "programmer" to generate a lot of decent results in an Object Basic type langauge.

Re:I know this sounds lame, but...

[JabberWokky]

> With an open source Object Basic tool I would start porting my programs like crazy -- as would a million other VB programmers.

>Do you think anyone is interested in your crap?

A super HLL like VB has it's place -- I submit to you that Visual Basic is to Windows as Perl is to CLI Unix (including Linux). With one key difference.

Although I don't like VB myself, I have seen or used plenty of goofy small apps that were written in VB. Everything from Sky Charts to conversion programs to those damn sheep that walk on your desktop. Nothing of any importance, but useful when you needed some bizzare little tool.

I have a feeling that none of these programs were written by anybody that you would want near your kernel code -- they weren't programmers as much as hobbyists. They had a vision of what they wanted, or just started banging on the keyboard, seeing what they could do. I know the sky chart program was written by an amatuer astronomer for himself. Maybe the sheep author just liked sheep.

They key points are: VB allowed the unwashed masses to create what they wanted without much effort, some of those people probably went on to become "real programmers", and I couldn't fix the problems because I didn't have source.

Hey! That last one is the difference between VB and Perl (calm down, it's not the *only* difference... just the one important to this missive). I know plenty of people whose most recent accomplishment has been discovering the "WordArt" button go on to fiddle with some shell script written in perl. But I can come back and fix it or at least look it over to make sure it dosen't call "rm" anywhere in it.

The problem with Perl? It's easy for non-programmers... but not for GUI work. They want some sort of hand-holding IDE that pops up help, auto-completes their line, and suggests where the bug might be. Linux could use that. Not for us -- for them.

Now, I'm not saying that if we arm the masses with a HLL for a Linux GUI that we are going to get an office suite or anything terribly useful.

No, we're going to wind up with hundreds of screen savers, star charts done up the way some guy in Topeka wants it to work, and those damn sheep.

But, I had to outlaw those damn sheep in my prior job. They filled up our mail system, and caused about a 10% increase in BSODs... because *everybody* *ran* *them*. 400 WinNT boxes, and at least two-thirds had those sheep on them.

...

Lot's of people talk about how Linux needs a stable web browser, or a better word processor. In at least one financial office, the killer app was some little .exe that made sheep walk on the desktop. Written in Visual Basic, probably by a non-programmer. And it captured a two-thirds market share.

Linux could use a GUI super high-level-language like VB.

Not for us coders... for the rest of the world.

and their damn sheep.

--
Evan

Re:I know this sounds lame, but...

[Acy James Stapp]

Perhaps XBasic is what you are looking for. http://www.maxreason.com/softw are/xbasic/xbasic.html

Re:I know this sounds lame, but...

[elflord]

Its not that they don't know or don't care, its mostly because they're rampant conformists.

Not really true. Mostly, it's been because the target audience were traditional UNIX users who didn't care that much, and that the GUI toolkits they were using simply sucked. This is starting to change with GNOME/GTK and KDE/QT

To denounce the absolutely horrid and disgusting UI of Linux is to spit on their holy ground,

This statement is a gross over-generalisation at best. If you bothered to discuss usability issues with the GNOME/KDE developers, you'd see that they really are listening to the users.

There are some sane Linux users out there but sadly, they're just that, users.

If you're implying that the KDE and GNOME people are "not sane", I'd say this statement is outright slander.

Re:I know this sounds lame, but...

[Pfhreakaz0id]

Geez. Don't apologize for using VB. I hate to break it to all the folks here but there is a REASON businesses use VB: It's often the proper tool for the job. Folks I don't apologize for using Vb. The programing I learned as a kid was Basic and some assembler. I've tried to learn C a couple of times and, frankly, wasn't impressed. I don't understand why I should take twice as long to produce code that does the same thing. I can do multi-tier apps, talk to any database I want (what database doesn't have an ODBC driver?), OO stuff, write my own add-ins to customize the IDE and I don't have to recall syntax of a damn command/function I use once a year (or even a function I just wrote), 'cause VB will pop up the syntax for me. I can write a VB program that doesn't use data binding, that is scalable, that manages it's memory properly to release resources back to the OS and is TRIVIAL to integrate with the other applications that businesses use (not JUST m$ office. Lot's of programs expose a type lib object model these days.) Of course, if I had heavy string manipulation to do, maybe I'd look at a C++ COM object I call. Ok? There is a REASON businesses have become so sold on VB. There is a REASON it's the most-used development environment.
---

Re:I know this sounds lame, but...

[divec]

Don't apologize for using VB. I hate to break it to all the folks here but there is a REASON businesses use VB: It's often the proper tool for the job.

I agree with you about this. I also think it is stupid to bash people because they write in VB.

What I really hate about VB is being forced to use it when it *isn't* the right tool for the job. And also, being forced to use programs written in VB which should never have been written in VB. For string handling, PERL is much better. For large projects with complex data structures, C++ is often better. For projects which are hard to write in any language, any language with a decent debugger is better. It is annoying when your boss makes you write 10 screens of unmaintainable junk with no error-checking, when you could have written a readable 10 line perl script to do the same job better. This isn't a flaw in the language itself, but it is marketed and used for environments which it's not suitable for.

Re:I know this sounds lame, but...

[Cycon]

They key points are: VB allowed the unwashed masses to create what they wanted without much effort, some ofthose people probably went on to become "real programmers", and I couldn't fix the problems because I didn'thave source.

Hey! That last one is the difference between VB and Perl (calm down, it's not the *only* difference... just the oneimportant to this missive). I know plenty of people whose most"WordArt" button go on to fiddle with some shell script written in perl. But I can come back and fix it or at least lookit over to make sure it dosen't call "rm" anywhere in it.

How can you assume that this will change with the advent of Delphi on Linux? Unless Delphi is an interpreted language (which I actually wouldn't know since I've never come into any for of contact with it), there's no assurance that you're going to be able to evaluate the security model of your user's dancing sheep...

At the very least, this article clearly enforces the fact that Delphi itself will be neither Open Source nor Free (as in binaries), so honestly, how would one be any better off if there's a plethora of Delphi-developed applications suddenly available for Linux unless those programs are released under and Open Source License? (aside from side benefits from bringing in new users to Linux, increasing Linux's viability, etc.)

--Cycon

Re:I know this sounds lame, but...

[guran]

No Delphi is not an interpreted language. Nor is C++. Then why does people release c++ source?

Hopefully Inprise will do to Delphi as they did to C++ Builder namely release the compiler, but charge for the RAD tools.

And closed source apps on an open source platform is far better than closed source apps on a closed source platform.

Re:I know this sounds lame, but...

[SurfsUp]

I hate to break it to all the folks here but there is a REASON businesses use VB: It's often the proper tool for the job.

Or at least it seems to be. Don't look now, but most of those businesses are now switching to Java to do these same apps, having been burned just too many times when VB fails to deliver the goods. If you don't believe me, read the job ads.

VB is, at best, a prototyping language, ranking somewhere below COBOL in the fitness-for-mission-critical-use department.

Re:I know this sounds lame, but...

[elflord]

But anyhow, GNOME and KDE are, just as the author stated in his article, are just getting closer and closer to functioning like Windows

KDE and GNOME are primarily development toolkits, not user interfaces. Most of the effort thus far has gone into creating APIs.

Listening to the public instantly means you're clueless.

If it doesn't work for your users, it doesn't work. Obviously, listening to everything Joe user says is not desirable ( for example, Joe user doesn't know much about application design ). However, it's important to be receptive to feedback from your users.

If I went up to the guys that create GNOME and KDE and told them everything they are doing is old-fashioned

Such as what ? If you have any brilliant ideas, I suggest you post them to the KDE mailing lists instead of whining on slashdot. You may be surprised. Since you say "if", you obviously haven't tried to offer any input.

Re:I know this sounds lame, but...

[elflord]

Pretty good summary. Some comments:

1) most of the programming has been by-geeks-for-geeks, which tends not to emphasize gui;

True. Partly though, there was no good GUI framework available.

2) x and its window managers have traditionally been sluggish and offered poor functionality (for a desktop UI -- let's not get into the running programs in Moscow from NYC stuff), so *nix die-hardss don't expect much;

Window managers and X aren't supposed to do very much. The problem is that no one developed much of a GUI to go on top of these , though KDE and GNOME are a good start.

3) programming in traditional *nix toolkits wasn't particularly easy or pleasant, which cut down the incentive to do it;

True as far as the GUI stuff is concerned.

4)Motif was not free.

This was a biggie -- it made Motif a non-starter on Linux. long term, this is probably a good thing, because it forced the Linux people to design replacements -- QT and GTK -- which are much better suited to writing modern desktop applications ( and are unburdened my the runtime license ). In all fairness to the Linux people, GTK and QT only started in 96, so the Linux GUI is really still quite a new beast.

Re:I know this sounds lame, but...

[Zigg]

The vast majority of it is useless, buggy and there are scads of inconsistencies to be found in each and every one when it comes to user interfaces.

And I must say to you that large swathes of Linux software I've seen that's been put out by ``independents'' (i.e. not the large groups like KDE, GNOME, etc. or the more well-known programmers) falls into the same category. The difference is that it generally does not force you to click through a plea for money before you use it. :-)

There are hordes of people out there who will do a bad job with any tool. I believe that a bad tool (and I think VB is a bad tool for quite a bit of stuff) will be passed over by a good person.

Re:I know this sounds lame, but...

[Now15]

You forgot one significant language group, divec. It's called HTML.

I'm the head programmer in a medium-sized wholesale/retail/ass-kissing firm, and I've found that HTML can be the answer to almost any programming problem. And so long as the client-end can parse the page it can be read. With a touch of javascript, the interface became as easy to use as a regular VB app. With a little CSS, the interface became attractive, too.

We have rostering, stock management, address books, calenders, indeed solutions for almost every part of the business running on our intranet. The exact same programming is being displayed on the many Windows machines, the Macs, the weirdos with Linux, and the management with their iBooks running on battery and wireless networking.

I don't think we've got a home-made excecutable left.

We primarily use coldfusion, there's a smattering of perl when it's been needed. In fact there's one script that was made by one of our extra-geeky geeks in c++ and one little brute force maths routine in assembler. It's used to do some extra tricky mathematical stock forecasting and graphing, displayed as a GIF.

The best part about this, is that we successfully avoided a costly tech upgrade scheduled for Jan 00, because the client machines were using less RAM, less HD, less processor power, less network traffic. Software costs are way down, because all a machine needs is the OS and a free browser (mostly IE).

Backups are easier and faster, reliability is higher and redundancy is a given (oops, did you drop you laptop? use mine for now!). Software upgrades are practically unnessicary, and support calls are down dramatically. In fact, the support guy seems to be spending most of his time reinstalling Windows and assisting lame users (hey, we do love 'em) with Microshot Word.

Might I add that application development times are way down, too?

We are currently investigating the feasability of replacing all the desktop machines with Linux, KDE, and Mozilla (once it's released).

The cost savings of not having to upgrade the entire network to 100BaseT alone saved us a pretty penny. We spent that money on another server instead. If you're asking, yes -- the servers are Windows NT 4 boxes, and yes -- we are investigating Coldfusion for Linux too.

In summary, before you start any multi-user software project, ask yourself "can this be done as a web page?", and think about it carefully before you dismiss the idea.

--

Re:I know this sounds lame, but...

[Pfhreakaz0id]

Agreed. That's why I said, for instance, that If I have SERIOUS string crunching to do, I wouldn't recommend VB to a client. VB's string handling is slow. But I'm talking a LOT of strings.
---

Polemic comment in the article

[divec]

Linux is about choice. Any Linux advocate who says Delphi is not welcome in the Linux space is a hypocrite.

That's not true. Some people who advocate Linux do so because it is free software, and would only advocate free software, and consider non-free software to be coercive and immoral. Danny Thorpe clearly disagrees with this view (as do most Linux users, probably) but that doesn't mean that its proponents are hypocrites.

Re:Polemic comment in the article

[jkaster]

I guess you're not considering the possibility that there may be a free version of Delphi for Linux, or a free Object Pascal compiler released by our company in the future. We've already released a free JBuilder and a free C++ Compiler.

Re:Polemic comment in the article

[divec]

My point is that the Linux space is vast and very "free" in many senses of the word. There is room in the GPL to get paid, so why take the ambiguous point that "non-free" software is immoral. First tell me what you mean by "free"...

Ok, maybe I could have phrased it better. I'm saying that some people only advocate software that is "free" in the Debian Free Software Guidelines sense, and hence would not advocate Delphi (unless it is open-sourced). They may be right, they may be wrong. But my point was that their view is not hypocritical, and I think that Danny Thorpe is not fairly describing such people.

Re:Polemic comment in the article

[divec]

I guess you're not considering the possibility that there may be a free version of Delphi for Linux

You're right, I wasn't. If this does indeed happen, and the license is DFSG-free, then I would agree that Danny Thorpe's comment is quite right.

Re:Polemic comment in the article

[divec]

You should probably test your theory out by starting a company, hiring 50 developers and giving away the product for free. That should work out well.

Actually, I wasn't claiming that shunning non-free software was a practical viewpoint, just that some people hold such beliefs and that doesn't make them hypocrites (which is what Danny Thorpe said).

(But in fact there are successful companies who do produce only free software, so it is possible even in today's legal climate. One good way to make money is by selling support for your product).

Until you realise that OSS/FS is just a new way to develop software and a new way to model a business, you are going to end up dissapointed over and over again.

I believe this is the same kind of thinking which was behind Danny Thorpe's claim. To many people, OSS is just a design methodology - ESR and Linus to name two examples. But there are other people - say RMS, Bruce Perens and Alan Cox - who think it is more important than just a design methodology and that there are freedom issues involved too. You or I may disagree with either view, but you can't call either view hypocritical, which is what Danny Thorpe did.

Re:Polemic comment in the article

[thimo]

Hmmm... Interesting, *very* interesting. Now the question that pops into my head is: Are *you*, being the company Inprise, seriously considering what you are suggesting here?

/me would ask Inprice to mary me if they would... ;-)

Thimo
--

Personally, *I AM* ready for Delphi on Linux.

[torpor]

As a long-time Linux user, and a recent religious convert to the ways of Delphi (having, wrongly, abandoned my old faithful Turbo Pascal skills when Windows 3.0 came along in favor of C/C++, which was a *HUGE* mistake because I'd written about 3 million lines of Turbo Pascal code during the late 80's/early 90's), I am totally ready for Linux Delphi.

Delphi rocks, as a RAD tool. There really isn't much out there for Windows that can compare - Delphi *ACTUALLY* made Windows programming fun again - specifically the extremely well-designed VCL.

Prior to Delphi, Windows suffered the same fundamental problem that Linux currently does, at least for me anyway, which is that there are a large number of API's, and multiple different ways of doing things, from a developer standpoint.

The Delphi VCL changed all of that for me as a developer who cares about getting things done fast, as rock solid as possible - it encapsulates a lot of the dreck that is the Windows GUI API, and makes it productive.

Now, I'm not saying that Linux is the same - certainly, the GNOME/KDE efforts are very well designed projects, but there is still a last-step of organization that is required to make RAD a reality for those GUI environments, and I sincerely hope that Delphi can bring that into the Linux mix. Either way, Linux will still be a great platform to deploy apps on, and I use it every day regardless - its just that the Delphi way of doing RAD is going to make for a *huge* shift in developer focus away from such mundane things as library dependence, text-based GUI design, towards rapid application development.

And, since Linux needs apps, rapid app development can only be a good thing.

When I can use Delphi to build apps on Linux, I will ditch whatever last vestiges of control Microsofts operating systems have over my current development environment/requirements, and happily be a full-time Linux developer. Right now, I'm *forced* to use Windows as a client software deployment platform, because Delphi makes Windows programming so damned fast...

Re:Is Delphi Ready for Linux?

[Speed Racer]

You mean products like Delphi, which is made by Borland (now owned by Correl, a merge resulting in renaming them Inprise).

The only thing worse than an incorrect post is a post correcting it that is also incorrect.

In the beginning, there was the land called Bor, and Borland was good. Borland begat InterBase and Paradox and Quattro Pro and Turbo Pascal and Delphi and C++ Builder and J Builder. And the Win32 developers did rejoice.

Behold, a curse came upon the Borlanders and that curse was Del Yocam. This evil leader, full of guile, changed the name to Inprise to capitalize on the buzzword "Enterprise". There was much weeping and wailing and gnashing of teeth in all the land.

The fairy stockholders were displeased with these actions and did cause that the stock price should plummet. After many fortnights, Borland rose again from the ashes of Inprise as a beacon of hope for all to follow. The evil Del was banished and all rejoiced.

From the vast hinterlands to the north came courting a certain Micheal Cowpland of Corel. This Cowpland did offer bounteous riches that he might own Borland and the Borlanders were beguiled by his cunning. Thus, Borland became united with Corel and Inprise was banished entirely.

ATTN: All Linux Zealots

[cwhicks]

This is a must read for all Linux zealots. I tried to say something along these lines in a post to yesterdays Delphi article, but I did not put the same thought and effort into it as this guy did.
This article is important for Linux freaks to read because this is how the rest of the world outside of the community looks at Linux. They community has trouble understanding that the world doesn't care about open/closed, MS/RH, or any other opposed forces. If you want Linux to be big, this article shows the path to it being a huge success with the other 99.9997% of the world.

Windows =! Innovation ?

[SnatMandu]

$foo = Linux is very innovative - how else could you account for its success?

$foo=~s/Linux/Windows/

Success is not a measure of innovation by any means. There are certainly some innovative things about linux, and the community behind it. But to call Linux on the whole "innovative" is going a bit far. "Cool", "useful", "Good", and "Robust" are better adjectives.

Good for business.

[B.D.Mills]

Delphi for Linux will be good for business, because it gives business more choice. One of our major products is software applications written in Delphi, which means our customers have to use Windows NT. Once we have Delphi available for Linux, we'll be able to port these applications to Linux, and our customers would have a choice of Windows NT or Linux.

Borland/Inprise say in the article that they would be hitching a ride on the Linux growth curve to make some money. What they haven't mentioned is the effect of a Delphi for Linux release on that growth curve. Having a product of the good reputation of Delphi released for Linux will mean that more businesses will be able to choose Linux where Windows NT was the only choice before. This will help Linux take market share from Windows NT. This effect is what will help Delphi for Linux become a major player in the market of Linux software development, and as a result, the Borland/Inprise management will be rewarded for their decision.

--

Answers

[Dacta]

(I program in Delphi, and I've been following Kylix pretty closely)

On Win32, Delphi supports COM to a great extent, put it is not overly dependant on it - apart from MIDAS (more later on MIDAS)

Borland have publicly stated that Delphi will not support COM on Linux. Delphi will continue to support Interfaces which (under Win32) are reasonably closely tied to COM (at least everything that supports IUnknown is).

Delphi supports CORBA on Win32 already, so yes, it will under Linux, too.

It will "play nice" with GNOME/KDE, but I believe that it is going to use Qt for its native toolkit. (That is from assorted developer briefings I've come across)

MIDAS is Borland's three-tiered data access technology. I'm programming with it at the moment (in Delphi), and it's pretty nice. It does depend on COM under Delphi, though - but there is a Java implementation, and you can run it over sockets, so I guess a Linux version won't be too hard.

I look forward to seeing what they are going to use for the actual DB access under Linux. Delphi comes with something called the BDE, which supports access to lots of databases - kind of like ODBC, but Borland has stated that developement of that has stopped. Delphi 5 (the latest version) introduced a number of Non-BDE data access components (which have previously been supplied only by third parties).

Re:Linux != Innovation

[El Volio]

Here's the thing: most of those successes you list (I'm not familiar with JServ, since I don't work with Java at all) aren't Linux-specific. The most outstanding example in your list is Apache, which was succeeding quite well before Linux. Our favorite OS may have contributed to Apache's continued success (and much of the interest in GNOME/KDE/Enlightenment/a thousand other packages), but Linux itself did not innovate those things.

And there are other ways to account for Linux' success other than innovation. Linux has given a relatively powerful and stable environment to the low end user, and despite what the FSF may say, many users are more attracted to the "free beer" than the "free speech". In my case, I think that the "free speech" is cool and brings a lot to the table, but I originally became a Linux user because I could get it for free without violating any license agreements or IP laws.

One could make the case that open source development is innovative, but let's be very honest. OS did a lot before Linus came up with this project. And someday, there will be a new darling of the open source world.

Lots of things succeed without being innovative. In fact, many times, you don't want to be on the "bleeding edge", but able to take advantage of new developments when they do in fact pan out. Doing something well is often better than doing it first.

We've already got RAD

[psp]

Absolutely no disrespect for Insight's efforts with Delphi.

One can get an awful lot done using PERL and GTK+-bindings. This combined with UI-designer programs like Glade is pretty powerful. After some practice I can fire up a working prototype of an application in about an hour. Perl and friends is already installed on most recent Linux systems.

Re:We've already got RAD

[mikera]

Have you actually used Delphi seriously?

I guess not, because if you had then you wouldn't be putting Perl/GTK/Glade in the RAD category.

This issue isn't so much the power or flexibilty, but the speed and convenience with which you can develop.

Don't get me wrong, I love all the free software tools. But they just can't compare with an application as refined and powerful as Delphi right now.

Re:Uh... OK?

[cwhicks]

There needs to be a place besides slashdot for this stuff. It's funny, but the furthest off topic I've seen yet on /.. Dude, you need to put your creative energy to work someplace. You're wasting it trolling here. Someone needs to start a slash style creative writing page. That could actually be pretty cool.
Thats your assignment for tomorrow class...
P.S. It's hard to pick your favorite episode, isn't it?

Whatever..

[Blue Lang]

A lot of people have said that this was a really 'articulate' article.. I thought it was pretty condescending. It filled me with a definite sense of 'well, duh.' I think Danny does an outstanding job of stating the blatantly obvious - which makes me really, really wonder why so many people are impressed with that pseudo-rant.

Are we really so used to poor grammar and hot grits that we'll take that kind of shit from a career money-hacker and like it?

I hope not. Linux might need applications, but I hesitate to agree that Yet Another "Rapid Application" tool is the proper panacea to engender them. I'd rather have 'lots of educated programmers.' ;)

Mebbe I'm wrong.

--
blue

Hooray!

[mikera]

No sooner do I make the transition from Windows to Linux then my favourite development tool follows me. There is a god after all.

Seriously, Delphi is a great product and will really speed the development of applications for Linux. Think of the simplicity of Visual Basic with the power and flexibility of C++ and you are somewhere close.

But the real advantage is the great environment that Delphi provides for GUI design work. Linux applications often seem to lack polish in the user interface. I am sure that this is in part due to developers not wanting to spend loads of time tweaking convoluted front-end code.

This is why Linux needs a good visual IDE. It's extremely difficult and time consuming to design a decent user interface without one.

In Delphi, you can build, compile and test a GUI for your application without even touching the keyboard. You rarely need to consult any documentation because most features are just a click away. All the chores of programming are basically handled for you, although you can still dive down into the nitty gritty if you feel the need. It's a nicely designed environment that helps you out with all the tedious tasks but doesn't restrict you in any way.

Basically, the sooner I can start coding with Delphi/Kylix on Linux, the better.

Big mouth

[aav]

Personally I hated a bit the way he put the things. Not because I cannot stand what he calls a "critical point of view" but because his ideas, apparently correct, hid something that I don't like.

Wouldn't it be cool if the VCL core packages were distributed with the OS? No it wouldn't. This only means that we would be somewhat forced to use the VCL for any application we have. And unless it's released under GPL I wouldn't agree with this. Why do you think Microsoft didn't agree to include those packages on its own OS ? Because it would have ment to give Borland the control over a part of their OS. Of course, they weren't that stupid. Should we be instead ?
As a possible solution : unless they release the packages under GPL (or some license that would not allow them to control the OS) they should stick with the installation kit solution

Another question : what are they going to do about the non-ANSI extensions in the BCB compiler (__property and so on) ? I agree they are actually extremely useful, but they are not ANSI. If we are going to use them, at least let's standardize them.

Third question : what about Qt - on which is based KDE ? Will the VCL based on that ? Will it try to replace it ? In what conditions ?

Ok. I asked a bunch of questions that may seem that I see this article and the whole Delphi story aas an evil thing. I don't. I worked for a couple of years with BCB/Delphi and I consider them the best RAD tools available now. Still, I wouldn't like Borland to become something it always wanted to be : some sort of Microsoft. Not because I have something with them - it's just the idea of monopoly that I don't consider viable.

Re:Big mouth

[Arandir]

QT wasn't GPL'ed through 1.4, either

Qt hasn't been GPLd at all! It's under the QPL license instead. I hope that this was a mistype on your part, and not an extremely erroneous assumption that all free software is under the GPL.

What about C++ builder?

[banky]

We're hearing about Delphi a lot, but everyone here is clamoring BIG TIME for C++ Builder. They're all crazy for it. I heard a while back that it was up and running on Linux... When will we get big news about that?

I'm all for Delphi, but we have a project that we're working on in MSVC6 that we will have to port to Linux sometime this year, and porting to to Borland is an interim decision that would, in theory, make the Linux port much easier (if the compiler is upwards of 95% compatible between Linux and Win32).

I agree, but dont care so much about porting if...

[torpor]

... the GUI tools for Linux are as easy to use as they are for Delphi/Win.

If Delphi/Linux requires a shift of paradigm in the way the GUI stuff works/integrates with code, I won't mind that so much - as long as I get the same sort of environment.

Porting most of my apps from Windows to Linux would be *good*, but it wouldn't disappoint me if I wasn't able to do this as a result of some design decision Borland/Inprise had to make in order to make Delphi/Linux work well, such as a difference in the way the various GUI toolkit/libs work compared to Win32 API's...

RAD to Linux, if anyone can do it, Borland can.

[Effendi13]

A big point to remember here is that Borland has had a long and rich history of bad business management. The reason the company still exists today within that 10% growth curve Danny was talking about is that their development tools are extrordinary! They far exceed the MS tools as far as ease of use and grasp of good technology. I've never heard of a developer using Borland tools going to Visual Basic or somesuch because they liked it better. You notice that Danny is Senior R&D, this means you weren't getting a mouthfull or "synergy" in that article, and reguardless of how the business is carried out, eventually you will end up with a fantastic, multi-language, GUI, X application developing, enterprise grade package that will allow all those people born to computers after the command line to produce quality applications for Linux... in droves of thousands! The number one complaint I hear about Linux is, "It doesn't have all of the applications I need." Here and now, Borland and Corel are going to solve that. I say, Welcome back Borland! -Effendi

A Good Thing

[Mr T]

He's right, we don't need to destroy windows to succeed. Our success will just be that much sweeter if we do though...

This is a good thing though. Firstly, I've got a thing for Wirthian languages and Borland has been keeping the flame going for quite a while. I hope they continue to do so. Pascal is wonderful.

Second, competition is good. VisualAge for Java and Delphi fill an important void in the application development market. No matter how badass a hacker you are, there are a ton of people who spend a ton of money not to be. We need to continue to embrace those people. There are movements to build similar products out of Python and then there is squeak and a good solid Delphi will spur development and provide a ready and working solution for RAD.

Lastly, and perhaps one of the more important reasons is that Delphi has been deployed. There are businesses with substantial amounts of Delphi code in production and they are tied to Windows until there is a Delphi somewhere else. This enables them to move to Linux. Along the same lines, we need to rumble and get IBM to revive "Bart" and port it to Linux. If we can soften the blow of porting your custom apps then we make Linux that much more desirable, particularly in corporate America where it is already infiltrating as a web, print and file server.

Proposed inclusion of VCL stuff

[Esperandi]

In the article, the guy mention that Microsoft wouldn't let Borland/Inprise package the Delphi VCLs into Windows itself, but he also says that Linux isn't Windows.

Is this speculation on his part? Does he realize that the Linux community will most likely not simply say "no, thanks" but instead firebomb their headquarters if they even think of integrating something that seems so close to Visual Basic?

Esperandi

Re:Proposed inclusion of VCL stuff

[Esperandi]

Just wait until they actually DO package the VCLs with a distro and Delphi programmers from Windows start coming over to Linux... I'll be surprised if something as extreme as a firebombing doesn't happen.

Linux zealots are just that, zealots.

Esperandi
Not a Linux zealot, but a very good predictor of them ;) (that statement offended all Linux zealots more than suggesting they might firebomb a place, I guarantee)

Why are some Linux users so damn greedy?

[EvlG]

This is a bit of a rant addressed to the small (but vocal) sect of Linux users that whine and scream and cry whenever a company announces a product that is not free (in all senses of the word.)

In two words: GROW UP.

Companies exist for a number of purposes; chief among them are to offer products and services to customers willing to purchase them. To do this requires that talented people be employed to develop and maintain those products and services. How can the company stick around, to continue to provide those products, as well as offer new ones? Sticking around requires a keen sense of the market, to continue to innovate and lead the pack in ways your competition isn't doing. And sticking around also requires that you continue to make money so you can pay those talented developers.

This is exactly what Borland is doing; they are among the few (but growing number) of companies that have the foresight to know that Linux is going to be a huge market, and soon. They are simply providing a product that many developers (myself included) are clamoring for: RAD tools.

Borland's RAD tools are among the easiest to use and most productive I have had experience with. Delphi and BCB give you all the benefits of a powerful language, without all the hassle and trouble of a "traditional" solution laden with mindless, cumbersome details.

It's really quite simple, folks. If you like the idea of developing a fully-functional, complete product in a couple weeks, without having to care about all the silly details that make your life miserable, then you'll love Borland's RAD tools. Afterall, isn't that why we all love Perl? It makes easy things easy, and hard things possible. However, if you'd rather spend all your time under the hood and get to know the application intimately, you're probably already quite comfortable with your current options.

What gets me is that amount of bickering and ugly fighting that rears its head every time a company offers a new product for the Linux market that isn't free. Why try and discourage companies from offering products that many of us would like, just because it doesn't satisfy your personal (albeit warped) agenda? Why not vote with your dollars, and simply not purchase the product that you don't want, rather that making a big scene every time it happens?

It's time many of the kiddies in this community grew up and got a clue. Linux IS all about choice, and you too have the freedom of choice to develop without whatever you'd like. There are many of us who want and/or need the benefits a RAD tool can provide. Why rain on everyone's parade?

Re:Why are some Linux users so damn greedy?

[Arandir]

It doesn't matter if the softwware is free beer, free speech or free verse, YOU are the one that is free. No one is holding a gun to your head forcing you to use Delphi. The original post still applies: don't piss on someone else's parade just because you don't have access to Borland's source code.

Re:Uh... OK? HAW HAW

[deltavivis]

Every time I've pointed out how i thought an 'offtopic' troll comment was actually original and funny (its rare, but it happens) i've always gotten a -1 for paying attention to the expressions of free speech delivered by the 40 year old virgins posting crap here. But now, score:3 for pointing out an onion article.
HAW HAW! (my best Nelson Muntz impression in ASCII)
This PROVES anyone with enough spare time to moderate has just got to be on the pipe. I'm only killing time now to write this because the #@!! server for /. keeps hanging and this is as far down the page as i can see...

Oh yeah--I'll submit another "slash style creative writing page" someone has started, i think its called freshmeat or something like that.

Re:You don't need C++ to work in Linux

[elflord]

There are three fairly easy scripting languages for Linux -- Tcl, Python and Perl. These will feel much friendlier and less foreign to someone coming from your background. Personally, I'd recommend you try Python. This is probably the best choice for GUI stuff. perl is very good at manipulating strings ( say for automatically writing html and stuff like that ).

IMO, I don't think we need BASIC, because we already have languages that are much better. However, development tools ( such as GUI buiders and IDEs ) to go on top of those languages would be cool.
Cheers,

Re:Delphi != VisualBasic

[Esperandi]

I'm completely aware of this. I am also aware that other posters have noticed his statements and, as I predicted when I wrote my original post before even reding the others, that there are a faction of Linux users vehemntly against packaging these things into Linux. I say Delphi is like Visual Basic because they serve the same purpose, to create an easy object-oriented RAD environment (and yes, VB IS object oriented, at least as much as C++ is, go read up about it if you're still arguin against VB 3).

Mark my words, saying you are a Delphi programmer on Linux will make you an outcast in the future. Linux users will laugh at you for using such a "stupid" language, and Windows Delphi programmers will insult you for using such a "stupid" OS.

Esperandi
Personaly, I use what works for my purpose and don't pay heed to what other people say, but I'm good at predicting it.

Not lame at all. Bring on the RAD tools!

[Ian Schmidt]

Linux currently lacks a way to make quick one-off GUI apps, which is the sort of thing Visual Basic excels at. By the time you toss something together in any existing GUI language for Linux (including PyGtk and other script bindings) it's become a Real Project.

Given that most open source programmers are "vi+gcc" gurus that would consider a VB-alike beneath them, I think it's terrific that companies like Borland/Inprise are filling in the gap. To use an ESR-ism, Delphi and other RAD tools will lower the barriers for scratching your own itch. That'll result in more Linux software and more World Domination. Bring it on!

What you say is not lame. However, VB is lame.

[torpor]

I don't think anyone would have a fair argument against having a tool that, as you say, allows you to do a bloody good job at building apps with a tool you are comfortable with.

One of the reasons everyone is so excited about Delphi on Linux is because Delphi is such an excellent rapid application tool, with a very good easy to use and easy to understand programming language, with great GUI capabilities.

Most technical/programmer people, in the Linux world at least, like well designed software that does a good job, and which fundamentally makes sense most of the time. The reason I personally find Delphi to be enjoyable as an environment, and which many people agree with me on, is that it makes programming *FUN* again - and the way that happens has a lot to do with the sensible design of the Delphi environment, and the Object Pascal language.

There is a lot of nonsensical design in Visual Basic however, which does *not* make sense. If you want to read a good article that describes a lot of problems with the way that Microsoft has designed Visual Basic, I suggest you read this article:

Thirteen Ways to Loath Visual Basic

I point this out not to be inflammatory, but to show you one of the reasons why you get so much flack as a Visual Basic programmer, and to point out that Delphi on Linux is a really good thing because a) it provides the ease of use that you're used to on Visual Basic, without any of the utter stupidities of VB (see article for details), and b) it's a very powerful environment for creating world class apps in its own right.

And, lastly, the design of Object Pascal and the Delphi environment is so good that, dare I say it, it will actually c) make you a better programmer.

And that's why its exciting that Delphi is coming to Linux.

Re:What you say is not lame. However, VB is lame.

[Zigg]

Thirteen Ways to Loath Visual Basic

Hmmm.

I started reading the article and couldn't help but notice that some of the syntax complaints that the author has for VB are also at least similar to the ``accepted'' way to code in Perl, if you've spent any time reading the documentation.

Thankfully, Perl doesn't (usually) force you to do it that way; but it sure could do with fewer multisymantic expressions (how many things do parentheses do in your language?) :-)

Re:I think you're misreading the comment

[divec]

Hence, the advocates of an operating system that claims to be all about choice, have in fact, restricted the choice of others.

The operating system doesn't "claim" to be about anything, it's a program.

Some people may believe that "Linux is about having as many software options as possible, regardless of the licensing".

But this is just one opinion. Other people believe that non-DFSG-free licenses are bad because they are socially divisive and infringe on people's right to share information.
If they think it is bad for Delphi to be sold under a non-free license, then that's not hypocritical, it's just their view. They never claimed that "we should allow as much `choice' as possible, regardless of how damaging we think the licensing is".

Re:RAD to Linux

[divec]

Even if those of us that support the open source model don't use it. The more RAD development tools we have, the better the community will be.

I'm not convinced that closed-source programming environments will *neccessarily* help free software. Free software is not that free if you have to buy a certain compiler to compile the source code. If lots of developers who would previously have used gcc switch to Borland then we could end up with a whole load of "free" software which won't compile on a free compiler. One of the current strengths of the community is that anyone on the Internet can potentially become a developer for any piece of free software. If lots of stuff won't compile on a free compiler, then it means millions of potential developers become second-class citizens who can't work on a whole host of exciting projects.

Re:Question 3:

[Arandir]

We'll if you want to get that picky about it, if a GPL application won't let me write a Qt front end, I'd say that it's the GPL that stifling my free speech. After all, it will be the GPL author who will sue me, and not Troll Tech.

Oh, by the way, Qt is 100% free, and actually has far fewer restrictions than GPL.

QPL incompatible with itself

[divec]

if a GPL application won't let me write a Qt front end, I'd say it's the GPL that stifling my free speech.

Actually, if you have an application QPL'd by anyone other than Troll Tech, then you can't combine it with QT. This is because you'd have to give the application developer the right to create non-free derivatives of the combined app, which you can't do.

The GPL may be incompatible with the QPL, but at least it's compatible with *itself*.

Re:QPL incompatible with itself

[divec]

Sorry, I'm probably not being clear (it's 5AM here!).

The QPL contains the following clause:

When modifications to the Software are released under this license, a non-exclusive royalty-free right is granted to the initial developer of the Software to distribute your modification in future versions of the Software provided such versions remain available under these terms in addition to any other license(s) of the initial developer. (emphasis mine)

This means the initial developer is allowed to make a non-free version of your modification, provided he makes it available under the QPL as well.

But say you've got a QPLed window manager, and you modify it, using some code from QT. To legally distribute this, you'd have to give the initial developer of the window manager the right to make a non-free version of your modification. But you can't do that because the QT part of the modification belongs to Troll Tech, and only they have the right to make non-free versions of this bit. So you can't distribute your modification at all.

In other words, the QPL is incompatible with itself.

Did I make more sense that time? Or am I still rambling in 5AM mode?

My Hats Off To Them

[Anonymous Coward]

As a originator of the lazarus project I tip my hats off to them. To convert the vcl to something that runs on linux is a totally daunting task. It has taken us a year to do it but we are now down to basic component finishing and ide codeing. We started out by having to totally abstract the widget toolkits from the vcl to allow it to run on anything what a bitch. We had to totally recreate our own vcl using no borland source which is useless anyhow (to much api junk). The cool thing about this is that we have a abstract vcl that we can plug in any kind of widget set into by just creating a interface unit. We had to use GTK for our first set but that is another story in itself. GTK is a bitch to write object based wrappers around since it is written so flat. Well anyhow I wish them the best of luck it is a hell of a job. Cliff Baeseman Lazarus Project http://lazarus.freepascal.org "GPL / LGPL"

Re:Delphi != VisualBasic

[divec]

Templates in C++ are not OO approach

What is a class template but a declaration of an unboundedly large family of objects all at once? Ok so you might say there's a difference between "object-enabled" and "object-oriented". BTW I was posing this as a question - can this be done in the latest VB?

Couldn't agree more

[guran]

Delphi is my platform of choice for Win development.

Sure, there are other languages better suited for different jobs. For an allrounder, though, Delphi rules.

Why? Combine a low lerarning threshold with decent power and you have a winner. Sure VB is easier, but you end up wanting more juice. Sure C++ is a lot more powerful, but debugging (esp other peoples code) is a nightmare. Sure perl is nice, up to 20 lines of code...

What Linux is lacking is RAD tools to quickly put a dummy-proof layer between the code and the user.

Re:Delphi != VisualBasic

[guran]

Linux users will laugh at you for using such a "stupid" language, and Windows Delphi programmers will insult you for using such a "stupid" OS.

No
Stupid Linux users will laugh at you for using such a "stupid" language, and stupid Windows Delphi programmers will insult you for using such a "stupid" OS.

Users and programmers with any clue will use whats best for the job.

Those who discard a good tool as "Stupid" will soon have to learn that their new boss won't let them call customers "stupid" because they "don't want fries with that"

XBasic

[AftanGustur]

Since some people have been inquiring about if there already existed a VB tool for Linux, it's here. XBasic. Runs on both Windows(95|98|NT) and Linux. Click here for screenshots.

--
Why pay for drugs when you can get Linux for free ?

AMEN!

[Pfhreakaz0id]

This is what I do too. As I speak, I am debugging a scheduler program that uses 700 + lines to code to 1: Define a new stock item (that really has very few characteristics) and 2: add that Item to a schedule for manufacturing! It's a mess of code with no rhyme or reason. I've done a bunch of stuff for this company (at $75 an hour, in a state in bottom five of cost per living)all like this. A while back I did a whole new program from scratch. The guy was amazed how fast it came together, how fast it was (even against a crappy Access databsae). I just said "see what happens when you have a REAL programmer, who actually thinks through the design, do something?" VB is my tool. I'm getting to know it inside and out. I can adapt to different databases (not really all that different) for a back-end. But I have no desire to learn another language when VB can do 95% of what I want and my firm can always outsource a COM object to a C++ programmer if we have to (we haven't yet). I won't recommend linux desktops to a client until they have VB for it. Period. That's not to say it's a bad tool or anything, but it's not what I know. I'm trying to learn some linux, but I have a list of things a mile long to learn and since I've yet to have a SINGLE client who has even heard of Linux, let alone brought it up!, it's pretty far down on my list.
---

Re:Linux =! Innovation ?

[Junks Jerzey]

Linux is successful because it went with the tried and true solutions rather than attempting to be innovative. If Linus had done something experimental and wacky, then he wouldn't have been able to leverage off of twenty years of UNIX tool development. Other people have written their own operating systems, or are attempting to, but in general you never hear about them because, while they may have technically succeeded, they don't have gcc and vi and emacs and Ghostscript and LaTeX and Xwindows and Mozilla and such.

Most high profile Linux-based open source projects are attempts to bring something from Windows or the Macintosh over to Linux. KDE and Gnome are heavily influenced by Windows. The Free Pascal and Lazarus projects have been attempting to clone Delphi. The Gimp is an attempt to outdo Photoshop. There are several projects trying to clone Microsoft Office. Code Crusader is cloning the CodeWarrior-style IDE. There are attempts to clone games like Civilization and old retro classics. None of this is bad! But almost never do you run across an open source developer with his own concept of what the future will be like. In some ways I find that fascinating, that the opportunities of a free operating system and open source haven't brought forth any Alan Kays or Charles Moores or Jef Raskins. I give Stallman credit for following his own vision with Emacs, but admittedly that project was started well over a decade before the Linux kernel was written.

Re:Splitting hairs

[Arandir]

This is splitting hairs too fine to be seen. A modification to the application IS NOT a modification to Qt! Troll's license only covers Qt, while the application's license only covers the application. Unlike how some view the GPL, copyright law does not grant the author rights to the works of third parties. Troll has no claims to any modifications of mine unless they they are modifications to Qt, not nothing else.

Hypothetically speaking, it might be possible to have a single code fragment that modified both the application and the library at the same time, but it would be extremely unlikely. I can't imagine this situation in any realistic code.