Slashdot Mirror
Kylix in Limbo
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?"
I don't know much about Delphi, but I know a good deal of Java, and it seems that the "new thing" for Linux is Java.
I'm sure there's a market for Delphi, but why not just use C or C++?
Does this say something about the application or about the difficulties of marketing a commercial Linux application?
Not to put too fine a point on it, but the problem may lie with Delphi, dontcha think?
By comparison, even if you could port a Windows app over to Linux tomorrow, that doesn't mean that every company out there is going to do it.
The opposite of progress is congress
Maybe it had something to do with the 1000+ price you had to pay for the full developer version? You think?
... hello to increasing "Shareholder Value".
Oh yes, Borland has come a long way since Phillipe's idea of a full blown compiler as good (if not better) than anything on the market for 99 bucks. Gone are the days of Turbo Pascal and Turbo C
And Helloooo to you too linux you cutie...you're looking better by the minute!
----- In Your Cubicle No One Can Hear You Scream...
When you have a multitude of free (beer), easy to use dev tools already out there for a platform, it's gonna be tough to push a product such as this. The biggest use I can see for it is to port apps, and even that doesn't seem to be quite popular.
I got a +5, Troll
The problem with Kylix was it's price. Borland was charing a ridiculous price for a product that albeit good wasn't worth the price. It's also hard to convince your boss (atleast in my situation) that Linux was free and came with C/C++ compilers but I had to pay for Kylix.
If they had a reasonable price perhaps it wouldn't flown but lets be realistic, it's not going to get a lot of support without having a cheap price or an open source version available.
Kylix didn't sell because it was a pile of crap. I used to do a lot of stuff with Delphi (paid lots of money to Borland too), but when I ditched Windows I felt no incentive to carry on with Kylix. I tried the Open Edition, and it wasn't a patch on Delphi. Klunky, buggy, lousy unportable code. Not worth it.
On Linux, there is a cornucopia of free programming languages and tool boxes ready to use. Why then should I use a commercial closed implementation of a proprietary non-standard language with non-standard libraries, not portable beyond merely Linux and Windows, and then only some versions of those?
I don't mind spending big bucks on good tools. After all, it is magnitudes more expensive to familairize oneself with new tools than actually buying them. But I do mind when my favourite tools suddenly become deprecated at the mere whim of a corporate - and Borland has a poor track record here.
Thus, no matter how good the performance of Kylix, and no matter how excellent and slick the IDE and libraries, I would not touch it with a ten foot pole unless I have some guarantee that I will be able to access the full source when I really need to.
Most people knowledgeable enough to develop on Linux have been burnt in the past by proprietary tools, have learnt expensive and painful lessons that way. Never more! Our freedom is too precious to sell out ever again.
I bought Kylix 3.0, and my biggest complaint with it is that it feels like a Windows program forced to run on Linux. Not just the IDE (which uses WINE to run), but the language implementation itself.
:-)
It feels like the developers have hardly used it itself, and I guess that's why it just isn't as much of a pleasure to use as (say) Turbo Pascal was.
I love having a decent pascal compiler for Linux, and I like the fact that I can recompile my code on Windows, but I keep bumping into things that just shouldn't be the way they are.
For example: I have triggered segfaults when exceeding boundaries on arrays. Excuse me? I'm using a typesafe language with bounds checking specifically enabled. I expect the program to halt on the line of code that is attempting to access an out of bounds address BEFORE said access happens. I expect all variables to be current and correct. I expect to be able to see exactly what went wrong exactly as it happened. That's one of the reasons to use pascal. I'm paying 5% overhead for that luxury, now hand it over!
The other reason to use pascal is the fast compile times, which is great.
I'm happy to have a pascal compiler with a nice IDE and neat rapid application development stuff for applications, and I use it by preference. It just feels unpolished and rough.
Oh, yeah, shipping apps sucks too - they require you to make wrappers and point LD_* things to shared libraries that you have to identify yourself. VERY MESSY and STUPID. Let me make static apps if I have to, but I get pissed off when the recommended solution for messiness is to wrap every executable I make in a script. Yuk. Not likely.
*sigh* So I guess Love/Hate it is.
Love pascal. Loved Turbo Pascal. Like Kylix. Hate icky stupid bits in Kylix.
Kylix devs should be forced to eat their dogfood. When they release a fully functional IDE written in Kylix, I will be willing to believe they have actually used it. Until then, I'll use it anyway, and occasionally rant in public.
---
When it comes to selling stuff, my old man always says "there's a lot more people with 5 bob in their pocket than 5 quid."
I like Delphi, but having to spend $1,500+ to buy it means I tend to skip versions nowadays.
I remember buying Turbo Pascal for about $90.
Perhaps Borland would sell a lot more copies of Delphi and Kylix if it was $150 instead of $1,500.
Linux users by their nature are averse to paying for software. I would rather roll my own stuff using Java, Tcl, PHP, etc. and then not be dependent on a company like Borland.
I looked at Kylix as it looked cool but now it appears I was correct in avoiding it. I pity companies who try to sell software to people like me who are addicted to free (as in beer) software.
----- Refactoring is the reason why man does not mistake himself for a god.
Once Visual Basic came along, it really stole a lot of their thunder in terms of making it easy to write windows programs.
Actually, Visual Basic came out before Delphi did. Delphi was designed later and was for many years a better product than VB, but:
- It was based on Pascal, not C, so lots of people thought it was a toy.
- It wasn't standard Pascal, so Pascal bigots didn't like it either.
- It wasn't a Microsoft product, so people didn't think it would stay around a long time.
There were lots of other problems too: Borland financial mismanagement, MS hiring away designers, etc., but I think "Not C" and "Not Microsoft" were the big ones.
Who needs Kylix when you can write your GUIs in Python using wxWindows, GTK, or QT for FREE?
Who needs Kylix when you can write your GUIs in Perl using wxWindows, GTK, or QT for FREE?
Who needs Kylix when you can write your GUIs in C/C++ using wxWindows, GTK, or QT for FREE?
Notice a trend here? Oh, but there's more...
Linux is found on Open Source software. Why on earth would I write a program in a propietary language than costs $$$ that would be pointless to distribute to the rest of the Linux community because only *I* could compile it? Quite simply, I wouldn't. I'd write it in Java or Python because I know other Linux developers would have Java or Python.
I do not know a *SINGLE* developer who has Kylix, and I only know of one application our company uses that was written in Delphi. That application is a very specialized mortgage application and is not usefull to anybody outside the mortgage industry (and I even question it's usability inside the industry). To add insult to industry, they're planning a complete rewrite in C# for 2005.
Finally, we all know that Borland has been wishy-washy at best when it comes to their support of the Linux environment. The Interbase/Firebird fiasco is proof enough.
I wouldn't trust my money with them. They've been made irrelevant by Microsoft, SUN, *AND* Linux. They consistently and stubbornly refuse to get with the program. That's why nobody users their software anymore.
And Turbo Pascal used to be a really really damn good product. It's sad, really.
Bryan
Considering Delphi is still alive and Kylix is built on mostly shared code from Delphi, I doubt they'd GPL kylix.
No, maybe it's a sign when you take a Windows program and make a half-assed attempt to 'port' it using Wine, it doesn't work right, you slap broken registration code on top of that, and the bosses shout "SHIP!", hopefully over the objections of the engineers.
The failure of Kylix is just another example of the free market working, and in this case the value of Kylix to the consumer is less than zero. That's right, Borland would have to pay me quite a bit to 'switch' to Kylix for anything. And it still might not be enough, if it kept crashing unexpectedly.
But hey, YMMV; that was just my experience with it. And note that I managed to restrain myself to the point that phrases like 'flaming piece of festering monkey shit' never escaped my lips!
pb Reply or e-mail; don't vaguely moderate.
The shame of it was that, as a RAD environment, it was incredible. The best thing was the masses of third party components you could get for free. The source code was available, you could tweak your own little components, and you didn't have to pay for them like you did for most of the VB components.
Things went to the shitter when Borland missed the managed execution environment train. Delphi got to a point and didn't get any better, and lost heap of the market share it once had.
> If you want to make a difference in the roads Borland takes, I can't
> stress enough that you *must* tell them what you think and what you want.
But we HAVE, time and again, for years. I stopped believing in them a while ago. Their "corporate" focus is so entrenched by now that they're completely losing sight of the developer community that brought them here. I really don't think there's any hope for Borland. It seems that when companies go public and join the "big league" their ability to interface to any non-corporate entity in any meaningful way is lost. I'm looking at their recent product developments and announcements, and I have no clue where they're planning on going. Some of the products are just plain wrong-headed (why bother competing with a C# IDE when every corporate IS shop pretty much gets all their tools through the MSDN?), and others are shrouded in such marketing mumbo-jumbo that I have no clue what they're about. Besides, the field is getting flooded with good RAD tools in every conceivable language nowadays, many of them free and cross-platform, I think Borland's chances of reviving Delphi (i.e. gaining significant market share) are pretty slim. I guess they're still going strong with JBuilder, but I'm sure they'll find ways of screwing that up, too.
So, *why*, exactly, did Borland think Pascal would be a good language to sell to Linux users?
.NET work. But I'm not sure I'd like C/C++ work under Linux much better.
Because those of us who earn a living doing Delphi work asked them to. It gave us a choice of platforms in the future. I certainly don't want to do
I think we should force every C programmer who can't be bothered to do bounds checking to use Delphi. We'd definately spend less time patching our systems.
FWIW, the limitations you mention might be a concern for standard pascal, but for Object Pascal they are generally myths. The only thing OP forces you to do is think about what you are doing.
It is amazing what you can accomplish if you do not care who gets the credit. -- Harry Truman
"Does this say something about the application or about the difficulties of marketing a commercial Linux application?"
First of those, I believe.
If there already are GTK and QT, why anyone would want to pull in another "GUI platform" without any historical roots in Linux, coming from alien environment?
I very much like Delphi and use it a lot, but I would really appreciate Borland making RAD tool set based on GTK or QT, instead of trying to mechanically bring over their Windows-oriented toolkit.
It just does not work. It is not serios...
6. pascal sucks
again...enaugh said
Oh, you were doing so well until then, and then you turned into a standard troll. Oh well.
The original Pascal was a clunky, limited language, sure. The famous "why Pascal is not my favorite language" covers all that. But Delphi bears about the same relationship to it as perl to sh.
Example: the original Pascal only had fixed-length strings; you had to pad out your strings with spaces to the size of the buffer. And Delphi? Dynamic heap-allocated garbage-collected buzzword-compliant strings. They can even contain null characters, which is one up on C.
And I have yet to find another object system as good as Delphi's, although the VCL lets the side down a bit there, declaring all sorts of things private that should have been protected and the like.
Incidentally, Kylix does offer something over pyqt, pykde, pygtk, and even perl-gtk: speed. Pity about everything else.
These glitches aren't all Borland's faults. For instance, it seems CLX runs much, much better on Windows than on Linux.
IMHO Borland's worst move was to depend on Qt, instead of porting the VCL to a 'C' (not C++) toolkit. That f*cking glue library is the root of all problems.
I like Qt very much and it's design looks a lot like VCL's own design, but the version of Qt available at the time - which is 2.3 and is used up today! - is leaps behind the current Qt3.2. But the glue library idea doesn't seem to facilite upgrades...
Out of the technical field, Borland's change of pricing was really dumb. There _needs_ to be a basic version which costs no more than US$200; it's what moves small companies in the "accouting software" market.
(oh, and that "open" edition is a fiasco)
Last: I hope Lazarus gets "ready" someday. There've been great improvementes lately!
We do a lot of work with Delphi. We are accustomed to the prices of their "Pro" editions, and they've always done what we needed to do... "Enterprise" was a waste of money for us. Now, though, to do the same development work under Linux as we do in Windows, the cost would be doubled, JUST because we're developing database applications for a web site, instead of a Win32 desktop.
So, we'll continue to use Delphi for Windows work, and PHP for Linux/apache. It's worked well. Kylix became a non-issue, because essential-to-us features were picked by some marketing type to be "Enterprise" level.
I think part of the problem was that Delphi=Pascal=educational-language *therefore* not professional. Thats from a professionals perspective.
From a noob perspective, Delphi would've been okay except that the learning curve was much much steeper than for VB. Delphi has very strong type checking, which is hard for people new to coding to grasp, so getting things to compile is difficult, so there's no immediate feedback, so the user is discouraged. VB doesn't suffer in this respect since by default you didn't have to declare anything and pretty much everything is a variant, so anything the user writes will probably work... albeit it may be buggy and is very unlikely to scale, but they don't care. They wrote a program, it runs, they've got the feedback of seeing it run, so they like it.
Its a huge shame that Delphi wasn't more successful. It was the correct language for writting business apps in terms of performance vs. skills required to use it. In a strange way, I do wish Microsoft had released it since at least it would've been more popular..... but then I guess that is C#!