Kylix in Limbo 443
IgD writes "Kylix, Borland's Linux port of their popular Delphi compiler has been covered on Slashdot before. LinuxWorld is reporting that Kylix development is in limbo. Many speculate this is a politically correct way of saying the project has been abandoned. There hasn't been any updates to Kylix 3.0 in well over a year. One user who attended BorCon this year wrote in his blog that Borland didn't have any updates to Kylix planned for 2004. This is really disheartening news. Why didn't Kylix sell? Does this say something about the application or about the difficulties of marketing a commercial Linux application?"
hey borland... (Score:5, Interesting)
Not the right product for Linux (Score:5, Interesting)
Unfortunately, Delphi is a marginal product on Windows (for various reasons), and Windows is the platform most software development efforts target. Move it to Linux, even if you can capture the same percentage of the development market on Linux, you now have a marginal product on a marginal operating system. Not gonna work.
An additional problem is: Linux runs on a myriad of platforms, x86, PowerPC, unix workstations, you name it. Kylix/Delphi work on x86 ONLY, so although code will be portable between windows and linux, it will never be portable to any other platform. This is a problem that would be very difficult to fix, if you look at the VCL much of it is written in x86 assembler, it will take a long time, and require much effort to port it to another platform. This portability problem further reduces the market share that Kylix could ever achieve.
And then there is the problem of price, enough other people have pointed this out, so I won't repeat them. But yeah, expensive.
Just my 2 cents.
-Spyky
Came too late (Score:3, Interesting)
It even came too late to have the Neverwinter Nights Toolset ported and usable in Linux.
StarTux
Re:The Problem (Score:3, Interesting)
> but the problem may lie with Delphi, dontcha think?
Elaborate please. It's still the best tool for whipping out large native Win32 apps. Sure, it's dwarfed by the number of users of MS development tools, but then which other development system isn't? The very fact that Borland survived the development tool shakeout and is still around is pretty amazing. Just because MS languages have such an overwhelming market share says nothing about the (lack of) quality of other tools.
Re:seen the price of VS.NET? (Score:5, Interesting)
I don't think the marketing and development departments at Borland have ever met. They've had some of the best developers over the years, yet especially in the last few years their marketing and PR was filled with arrogant know-it-alls. And the hordes of apologists for whom Borland could do no wrong don't help. Microsoft may have done VB first, but Borland did it right, yet ironically it's Microsoft reaping the benefits of much of that hard work at Borland.
they didn't treat their users very well (Score:3, Interesting)
I've used Turbo Pascal/Delphi since 1984 (Score:2, Interesting)
Re:We had been thinking about using kylix (Score:5, Interesting)
For native Win32 apps I still think Delphi is best, even in arrested development. But for cross-platform apps I'm very intrigued by Python and wxWindows (or wxPython). The apps seem to be truly portable, and wxWindows has such good binding to native widgets that you can create truly nice-looking and seamless GUIs. For most business-type applications I think it's a really viable option.
Stop your bitching. (Score:4, Interesting)
Delphi vs Kylix (Score:2, Interesting)
AFAIC Delphi is a great product. As said before, the only problems with it were: It's not Microsoft. It's not C++. And it's not VB. I've worked on government tenders that had a 3rd party company endorse our design and product recommendations. Then the customer's IT department ignored the product recommendation for using Delphi and Oracle and demanded it was done in VB and SQL Server. Why? Because they were products they knew (even though VB apps caused most of their support problems).
Delphi is a great product for producing gui and database apps. The language has a lot of power and flexibility. Kylix was designed to produce similar apps on Linux, but nobody wants them. Linux is a server OS, not a desktop OS. So customers want server applications, not gui applications. I've also written server-side stuff in Delphi (using Corba), and from that experience it's not suited to that sort of work either.
And since Kylix hasn't been as extensively used as Delphi, it also hasn't been as extensively tested. That's why the compatibility isn't there and the stability isn't there.
In summary, I think there are three reasons for the non-take up of Kylix. Price, stability and suitability. Personally, I'd like to see them do what they did with Interbase. But I doubt they'll do that.
Kylix isn't all bad (Score:2, Interesting)
I'm working on a MMPORPG, mostly working in Delphi and OpenGL, but the server runs on Linux. The complex data structure libary that represnes the Player's Data was written in Delphi and when moved over to Kylix to build the server it compiled without needed a single line of code.
I also used it to write a Apache Runtime Module so I could link the same data structure, account information, etc. inside Apache without needing to make Database calls in another language. Check it out for yourself @ www.vq-2.com [vq2.com]
I've even have the OpenGL engine (not yet released) working just fine without requiring too many changes.
I'd hate to see Kylix go. Maybe they just need something high quality/profile written in Delphi ported via Kylix to show people the real power. Or, maybe they were hoping they could catch the "Linux on the Desktop" wave and make some quick cash on developers wanting to move aps to Linux.
Re:Kylix/CLX has too many problems (Score:5, Interesting)
I hacked it for about 5 months working on a cross-platform idea I had (2 months on K2, 3 on K3). The interface to QT was to shallow and they installed an older patched QT version to link statically. The C++ learning curve is smaller than the K3 bug stomping exercise. There were too many features that you just couldn't use because of minor bugs or incomplete interfaces. Just try manipulating fonts on a TPrinter canvas and you'll see what I mean.
The concept was great. I drooled for the chance to get my hands on it. I would have gladly paid the $1K if the test/GPL version had proven a little more robust. I eventually had to abandon it too. If the finished shipping product had that many problems, I wasn't going to wait for the fixes. Now, I'm glad I didn't.
It's really a shame. Borland used to be the best there was on compiler/IDE usability. Their vision wasn't lacking on Kylix, just their engineering. Oh well. Back to the fish tank.
Re:Delphi? (Score:3, Interesting)
I come from a background of using emacs and the command line for building my code, and for me, Delphi sucks big time for being very restrictive and rather stupidly focused on the graphical layout of your project as it's primary organising principle. To me, code has it's own structure and very rarely does it revolve around the look of the application - yet this is the single organisational principle offered to the developer my Delphi. I very strongly think that this is a mistake.
I'm not anti IDE - I've used Visual C++ for C/C++ and Eclipse for Java and I really like both of those environments, because they are based around your code (a class browser being the standard main view). But in the case of Delphi I really think it gives the wrong idea about software development, and the code I've seen produced with it is pretty when it runs, but ugly and brittle on the inside.
IMO the reason that Kylix hasn't sold well is that it is a tool for people who don't understand that code needs to be elegant on the inside, not just flash on the outside.
Cross-platform compatibility was the problem (Score:2, Interesting)
I Tried it, and It Stinks (Score:2, Interesting)
Actually it reminds me of the line from Crocodile Dundee when he sees a TV set and says something like "Oh yea, I saw one of these in a storefront in Sydney once
Sorry Borland... maybe you should merge with Greenhills Software, and release a new version of "Multi 2004" Between both programs, maybe you could compile BOTH halves of the program.
Competition is just too hard (Score:4, Interesting)
Because noone shells out that kind of money for something that's arguably the _least_ mature Linux development environment. I've done lots of Delphi development and love it - elegant language, good extensions, garbage collection, nice IDE, good 3rd party components. A shame the MS tools have an unfair advantage, but that's how it goes.
I had great hope and expectations when Kylix was announced, and the good fortune to get Kylix 3 with my Delphi 7 package. But that didn't go very far. Delphi is a mature and feature-rich environment while Kylix feels dummied-down. Partly because the CLX is a subset of VCL, partly because hardly any 3rd party components exist, and partly because it's closed - that doesn't go down well on Linux. Kylix has a huge uphill battle to win - against tools that are FREE (GPL), are being developed rapidly, and have large communities around them. Alternatives like KDevelop and Qt Designer are hard to beat on their home turf - and an order of magnitude harder, the gcc.
Kylix is dependent on a revenue stream to afford future improvements, while the competition does fine without, and you start fearing that Kylix might not be around for long - another reason to stay away until it's proven itself. A chicken-egg circle.
Does this say something about the application or about the difficulties of marketing a commercial Linux application?"
Well, both. They've entered the market late with an overpriced and immature product. That's the application side of it.
The other side is that competing with mature GPL'ed products is very difficult. You're not going to win over many of the existing Linux developers, they'd have to rely on hordes of Windows people moving. That just didn't happen.
Kylix was a neat concept, but closed source development tools are (IMO) a dead end on Linux. I'm headed off to learn Qt.
Re:I own a copy (Score:2, Interesting)
Both have support for Delphi? That is simply not true. GNU Pascal doesn't support any of Delphi's nicer features (function overloading, dynamic strings and arrays, even classes). Free Pascal does better, but it looks like it still lacks things like dynamic arrays.
Not to mention, of course, that Delphi's IDE is one of the main reasons anyone uses it.
Delphi is up to version 8, and neither of these alternatives offers the features of Delphi 4. They're getting there, but they aren't acceptible alternatives yet.
Re:The Problem (Score:3, Interesting)
Well, the Kylix 1.0 install I had was quite unstable. It was at that point that I discovered that i could get 80% RAD using KDevelop and QT designer, rock-solid, and free (beer and speech). The only thing KDevelop lacks IMO is good code completion.
It makes you wonder if Borland is migrating from a tools vendor to simply an IDE vendor.
No, this is MS pulling the Windows paradigm shift v2.0. Borland survived the last one in the early 90's, so they're mopping up. Soon there will be only one significant development tool for Windows.
Try KDevelop. (Score:5, Interesting)
I tried Kylix before, and seriously, KDevelop royally kicks its arse. I don't know why anyone wanting a graphical linux IDE would use anything else.
Of course, there's always VI for the non-grpahical peeps.
It's all about the timing (Score:1, Interesting)
Too little, too late... and not for long enough to get adopted.
The biggest Kylix market is people who are existing Delphi-Windows developers who want to move to Linux... and this isn't a very large group.
The second biggest market is those who have already moved to Linux, but are having trouble deciding what set of GUI-builder tools to use, and how to use them. Do they use freely available tools, or pay a nice chunk of cash to use Kylix -- a barely supported product that only recently got out of beta testing, only to be left in limbo.
Borland gave up on Linux far too soon. If Kylix becomes complete abandonware, then Borland will miss the boat and lose their investment.
Linux on the desktop is still in its infancy in many ways. The best is yet to come.
Re:What about... (Score:3, Interesting)
Sadly, I must admit I'm one of the few people whose usage is specific in that way that Lazarus is a very good and sufficient tool
Here are the reasons why Lazarus does not belong in Delphi league
1. It's GTK1, so... no international fonts support, but as I see GTK2 is progressing very well
2. Some fancy features are yet to come in freepascal, but again progressing nicely
3. debugger is weird, it doubles watches, and watches crash IDE a lot, but still a lot better with every new version
4. no integrated help, and I don't need it
5. No database components, at least not finished, then again progressing very nicely
6. No reporting components
Ask that question in few months again, probably response would be a lot different, at least based on reading mailing list archives. Bug fixing and efforts put in Lazarus are enourmous, probably one of the best record of bug fixing ever. Sometimes not even 15 minutes, but almost every bug is fixed in a lot-less-than-daily matter in cvs. Kudos to the Lazarus team, and I really wish them well on replacing Kylix and Delphi as alternative
Re:Delphi? (Score:1, Interesting)
1) Hire really good people
2) Extremely aggressive scope management - including intentionally letting a whole lot of known bugs just sit there.