Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Programming IT Technology

More Kylix Information 112

Posted by Hemos from the programming-more dept.
A reader wrote to us with a presentation made regarding Inprise's Kylix. More information goooddd.
This discussion has been archived. No new comments can be posted.

More Kylix Information More

More Kylix Information

Comments Filter:

  • Hemos Strikes Again (Score:3)

    by Jacques Chester ( 151652 ) on Sunday September 24, 2000 @06:44PM (#757208)
    A reader wrote to us with a presentation made regarding Inprise's Kylix. More information goooddd.

    Which is why - of course! - Hemos's editorial comment is the shortest all day.

    --
    "Don't declare a revolution unless you are prepared to be guillotined." - Anon.

  • I think they'll have done their home work. dbExpress is probably going to be very cool. I'm wondering how apps will be packaged and distributed though. How would a stock (haha) Linux workstation get dbExpress onto it? Then I see all sorts of lib issues... ack. -Pete
  • Yeah, it has been posted on /., several weeks ago. Just do a search for Mainsoft and you'll find the article
  • Yes, this has been discussed on Slashdot. Check these two discussions: 08/17 [slashdot.org] and 08/16 [slashdot.org].
  • I do. Delphi/Object Pascal can do everything that C++ can do, except macros, multiple inheritence (although, like Java, this can be simulated with Interfaces), and templates. Of course, these are three of the worst misfeatures of the C++ language.

    People complaining that Pascal is restrictive have obviously never worked with a Borland product.

    ----------------------
  • It works like a dream with proper hardware under Linux.

    Translation: It's dog-slow unless you have 192 MB of RAM and at least a 400Mhz processor.

    I've tried using it on a 32MB P5/200 and it literally took 3 minutes to fire up the IDE (this is before doing any work at all). It refused to install with JDK newer than 1.2, and under the latest libc releases (although this is half Sun's fault -- their xfs recognition is *really* bad in jdk 1.3.

    On a good note, Borland's support personnel in thier newsgroups are VERY VERY helpful; JBuilder is actually a good product (although, I like Simplicity better).

    The wheel is turning but the hamster is dead.

  • Heh... my CAD class I took in high school about 5 years ago had predominately 286s running Autocad 9. It's funny to watch it redraw the circle in real time!

    I lucked out and got one of the few 486s. Funny thing was that even that piece of shit was better than the computer I had at the time.

  • First of all, Professional Delphi Developers plan developments just as much as any other professional developer. Code quality is much better in my opinion because you don't need to write screeds of code to achieve simple objectives.

    Programming in Delphi is still an art, only you can concentrate on a higher level of abstration than with lower level languages.

    Unlike VB and other earlier products I have never ever hit a brick wall with Delphi. If you really want you can include Assembler, so there is no real limit on being 'low level' if thats what you need.

    If you are trying to write a device driver, or a operating system, or a flight simulator, perhaps Delphi is not the Language to pick, but for a vast number of apps Delphi really hits the spot.

    Many applications I have developed were stand alone, and didn't need all the 'productivity apps' that go with MS. Apps like Point of Sale Systems, Factory Floor Systems, Stock Control etc, could all easily move to Linux. The Apps could be distributed with Linux tuned for the App, so no more configuring Windows for your App.

    That said, I am also learning Java now. Why? Because Java is cross platform. Server apps I write in Java will be able to run on machines from Linux to huge Unix Servers. Thats not the Market I see Kylix in. I see Kylix making inroads into developing GUI's for Linux, not Middleware as much.

  • Have you tried KDevelop [kdevelop.org]?

    It's not KDE specific, you can write any c/c++ app with it, and it does a lot of cool things like automatically documenting your classes and generating autoconf/automake for you. I like having the class browser as well. It also has a decent debugger as well, though I wish it was more like DDD (Data Display Debugger), which has the capability to graph data.

    Interestingly enough, I hated VC++ 6. I prefer PFE + CygWin for Windows. Fortunately I haven't had to do Windows programming in years...
  • Nope - I think Delphi 4+ does it too...
  • You can find out about Kylix from the official source --

    http://www.borland.com/kylix/ [borland.com]

  • Hummm.... now this might be the reason why critical code (plane computers, medical systems, etc...) is NOT made in C ?

    If the Pascal compilers tells you you shouldn't do something - then it is that what you are trying to do is wrong - PERIOD ! Beside, any language that relies on pointers should be shot immediatly.

  • Re:Kylix - The gun to your head ? (Score:3)

    by A nonymous Coward ( 7548 ) on Sunday September 24, 2000 @10:32PM (#757220)
    Or in other words, what is wrong with more choices? Someone forcing you to use them? Someone confiscating the tools you already have?

    You sound like yet another "I am the expert, do it my way, and no other."

    Especially considering the loudest yell against M$ is that they take away choices, and Linux restores them, this is really hypocritical. And add another shot of hyprocisy for this choice from Borland being a burr under M$'s saddle, which would seem to make them natural partners in crime with the Linux zealots who scream against M$ being the anti-choice devil.

    Gaaakkk.

    --
  • I completely agree with you about properties. They're a handy feature for improved code flexibilty and readability. However, I somewhat disagree with you about operator overloading. Sure, it could be useful if you were doing something like creating the TComplexNumber class. However, I'd be concerned about Object Pascal programmers following their C++ peers in overusing operator overloading. I mean, logic shift for input and output streaming? Class methods (static methods for you Java programmers) are generally a much cleaner solution.

    But who knows. Operator Overloading is a frequently requested feature in the Borland newsgroups.

    ----------------------

  • What makes you think he's a linux / Unix coder? He said he uses notepad and djgpp.

    Personally, I've used VisualC++ and managed to make some pretty cool stuff easily, but now using linux I already have just about everything I need, and anything I need to program myself I can do in sh or perl with vim... but I'm not a professional programmer and my requirements are light.

    I don't know enough about the "freedom" issues to have an opinion on whether this is a Good Thing yet, but an IDE from Borland sounds very useful for those of you who need to produce full-featured and native code GUI applications and don't have the time to learn the ins and outs of emacs/vim and GTK/Qt/Xlib.

    "Free your mind and your ass will follow"

  • You can think of this as being totally in keeping with the java way, what is passed is the thing that the program uses.

    But you're not. You think you're passing in the object, but you are in fact passing a pointer (or reference, if you prefer), which is a different thing entirely.

    By contrast, in C++, if you specify the argument type as Object, you get an Object, and if you specify an Object&, you get a reference to an Object

    --
    "Where, where is the town? Now, it's nothing but flowers!"

  • Yes. I was deep into dBase. Company standards switched to Paradox. Nice QBE. It is ok, but you should like to get into a non desktop database. It will make your applications more robust it is not more difficult to use. I use Interbase. There are other products. However, MS SQL will not run on linux
  • Borland is working to bring probably the best programming language around to you guys, and all you can do is whine about crappy software that's going to get written. AT LEAST some software might FINALLY get written! Good GOD man, are you OUT OF YOUR MIND!?! This is one of those things that might FINALLY bring Linux in to the mainstream. I've about come to the conclusion that none of you deserve to have Delphi ported. It should probably stay a niche OS so people like you have something to complain about.
  • Been there. done that.


    As long as Kylix interfaces with decent SQL databases such as DB2 [which is said to be very good], Oracle, and Advantage, you are gaining NOT to be using Paradox.

  • *grin* Yeah, bad code is caused by the language....I always thought it was VB AND VB Coders that sucked.
  • While I didnt' attend the conference, apparently someone from the IS departement did, and came away quite concerned about Delphi's future (in Windows)

    This is unfortunate. I think what happened at the Conference --- I was there (speaking!) --- is that the emphasis got misplaced. We are not trying to turn ourselves into a Linux shop; we want to be a cross-platform shop, which means that Windows support is just as important as Linux support.
  • Its perfectly consistant IF your midn isn't stuck in C mode.

    try to forget what C taught yo uand coem at this fresh...

    In ALL cases java passes the value of a variable.
    Its just that one of those variable-tyoes is an
    object-reference vairable. You are still passing the VALUE of the object reference.

    Your problem is you are used to thinking in terms of pointers, not references. Every place in memory has a pointer address (even primative types) thus it is soemthing different from a variable value. But references only exist for objects and ARE a primative type on an euqal footing with the other primatives.

  • ...references only exist for objects and ARE a primitive type on an equal footing with the other primitives

    Cogently put, you just have to remember that the ONLY way to access objects is via a reference. I'm not sure why the 'other' primitives aren't accessed via reference (a la wrapper classes) too, but I suspect performance might be a factor.
    ----

  • Certainly the work that our former coleague did may have caused some of our pain, but the language and/or IDE seems suspect in a few areas. It may be that we're still on the steep part of the learning curve (steep, sharp with broken glass and barbed wire in our path) but I've seen some pretty basic things just not work. I really hope Kylix is good, but I'm not buying into the hype (and that's all this is, there isn't really much content in the article) just yet. As to that crack about job security....well, that was a little below the belt. But I do know of a delphi programmer working in web development. He needed to do web pages that hooked up to a database back end. So what did he do? He re-implemented asp in delphi - poorly. He started about a year ago and is still going. His employer has started doing all their work in this beta quality asp workalike language. Hence my quip about job security.
  • I'd say it's more a reflection of individual programming style. Personally, I prefer to start with the easy creation of a visual template for my application, and then proceed with the actual programming of it. Whilst the old methods did indeed force us to put our ideas on paper, the pain of then having to manually create the user interface and then cobble thte code onto it was more of a pain then the benefits that enforcement of paper designs provided.

    I shudder to think back to some of the early Windows 3.1 developments tools... I was never able to finish any serious graphical application until Delphi 1.0 was released. And since then, I've never looked back.

  • A bit off-topic, but I noticed the "42" tattoo on the Linux Penguin in the presentation slides. Is this an existing part of the penguin's "look and feel", or is it a subtle indication on the part of Borland of their hopes for Kylix? ;-)

    PS: 42 being a reference to Douglas Adam's Ultimate Answer, for those who don't get it.

  • After getting into the C++ world I've noticed that IDE's tend to piss me off more than help me out ... right now all I really would need out of an IDE would be that it could show all the sources for a project and could print out color coded to make it easier to read.

    Besides that I really see no points in all the glamour of IDE's ... MS Visual C++ is non-impressive to me and I really dislike using it even though I'm required to at school. I find more functionality and freedom from loading up djgpp which is what I compile on at home ... and what do I use for an IDE? ... yep notepad ... Truly I see no reason for all these fancy IDE's except they put the tools you'd need to save time right there for you.

  • If I am relying on a computer to help decide what code is need in my application, isn't it going to try and error on the safe side, and give me more than I need? Sounds like a fast way to make bloated code again.

    Pardon me for griping. At least one used to be able to play Frogger(tm) in 64 KB of RAM in full color and sound on a C64, or even 16 KB in Black & White without sound on a Timex Sinclair 1000.

  • Programming for hire has always been about churning out something "good enough", borrowing heavily (cutting&pasting) from previous work, and perhaps even plagiarizing others. RADs help by doing some of the grunt work, so you can get to the fun part of designing and implementing algorithms.

    You remind me of someone who would complain that painting was a purer art when each artist pressed his own papyrus, instead of buying it from a local merchant. Sure, you could do that if you wanted, but why would you want to? Get on with the important details.


  • RAD is good. RAD is fab. RAD has all the rage.

    But a RAD such as Kylix may spells trouble for Linux.

    Not that Linux has no bloatwares of its own - netscape/mozilla and staroffices are two perfect examples, - but the easy-to-use RAD thingy like Kylix would inevitably introduces LOTS AND LOTS of fun-time programmers to Linux, and opens up the Pandora Box for Bloatwares that we all know too well how they have done to the Windoze arena.

    My only hope is that the programmers in the Linux Camp are a better quality type, and that may cut down on the unnecessary bloats all of us are so darn tired of.

  • I think my highschool still teaches pascal, and up until last year they were teaching it using turbo pascal 6 on a lab full of 286s, 386s and a few 486s. Now I think they teach it on P133s or something. Right now I'm taking AP Computer Science AB on a lab with everything from 386s to a few P90s. Why the dumb beginers get the better comps I will never understand. All the good comps at our school go to the media production (photoshop) and CAD classes, it's so unfair, but I digress.
  • You're almost right, except that you need an editor.

    Compiler, Linker, Editor. That's the classic programming model (trust me, a full-featured editor with regexps is a godsend.)

    You can get versions of vim or emacs for Win32.Try one. Anyway... Truly I see no reason for all these fancy IDE's except they put the tools you'd need to save time right there for you.

    Hmmm, a software package designed to give you useful tools to save time? Yeah, I don't see the point of that either. Huh?

  • RAD is not a bad reflection of our times. (Score:5)

    by The Optimizer ( 14168 ) on Sunday September 24, 2000 @07:46PM (#757240)
    As someone who has regularrly used both Micrsoft's Visual C++ and Borland C++ Builder for the past 4 years, I can answer in word: NO

    Let me sum it up this way: If it was all about "Programming as a pure art form" we wouldn't ever be using Windows, X Windows or any other layer between us and the computer. We would write our own display drivers, window managers, (G)UI stuff, etc, etc ad nausum...

    It has very little to do with the tools/languages/compilers and an awful lot to do with the programmer and his circumstances. It sound like that in your situation you are being pushed too fast, regardless of the tools you have at your disposal.

    And that's all each RAD product is: Another (alibet very powerful) tool there for you to choose ot not choose.

    In the past 4 years, I've written some code that I believe was fricking awesome code. Some of it was in Assembler, some in straight C++, some using C++ builder. And in the past 4 years I've written a smaller amount of code that would qualify as anti-awsome. Some it was in straight C++, some in C++ builder (None in assembler, ;-) The choice of development tools was a non-factor.

    Borland's development environment and components don't take away my opprotunities for creative expression in code; rather they take away much of the drudgery. I wrote my own window-button-click code back in '89. I got tired of re-writing it by '92. Now I'm glad it's taken care of --- UNLESS- I have a reason to monkey with it - in which case the VCL lets me jump in and do just that. RAD never took away the cool/unique parts of the appiclations I wrote, they just let me spend a larger percentage of my time budget on them.

    As a person experienced with both (Borland's) RAD development, and the more traditional approach I 110% agree with the following statement:

    Apart from the Delphi developers, Kylix also brings hundreds of thousands of applications, built in Delphi today, ported with Kylix tomorrow. On the first night that Kylix ships, we'll probably see more new Linux applications than we've seen in the past few months...

    Kylix will be a huge boon for Linux by providing the reason that will give non-geek users the opprotunity to try it: applications that mean something to them.

  • Of course, these are three of the worst misfeatures of the C++ language.

    You forgot operator overloading. ;-)

  • I love Pascal too. The property keyword is something I really feel missing in C++. But then again C++ has operator overloading. Which is a GOOD thing!
  • That's why the Free (beer/speech) software community is so important. It gives us the opportunity to develop meaningful software as the art it is intended to be.

    Case in point: I just grabbed a copy of Flight Gear. I used to think FGFS was an ambitious attempt that would never get anywhere, but DAMN! It's made an AMAZING amount of progress. It's really not that far from a widely useful and complete flight simulator. It hasn't been the subject of a rushed release schedule or marketing bullshit. It was created by people who think flight simulators are cool. They have applied their talents to something they care about, and the results are beginning to show. I suspect that the FlightGear project will produce a superior simulator most commercial offerings eventually.

    Unfortunately, the keyword is "eventually". Businesses often don't like the unpredictable nature of free software development schedules, even if they can figure out a way to make money from it. That's the reason for RAD tools - they are driven by the companies that need to release products tomorrow, not coders who would rather do things right.

    I'm not involved with the FGFS project; this is just an example (I'm a flight sim enthusiast). It could be equally applied to many free software projects.

    -John
  • I agree that people should be told that they
    can download linux for free. But, I do not agree that the open source movement means you cannot make money off of it. Too often people confuse free software with open source software and vica-versa. The GPL protects software from harm that
    commercialism can cause it. If Redhat wants to sell something that can be obtained for free, I say, let them. It's in everyone's interest that they stay in business, so that they can continue to pay programmers, just like us, to do work on an OS that we all love. Not to mention that by buying a pre-packaged CD from redhat you get a manual, and tech-support, as opposed to a burned CD, and having to bug a friend for tech-support, in their already too busy lives.
  • I've always heard good things about delphi and naturally expected it to be a good programming language/development environment. Recently a member of our team left and the others in the group (non-delphi programmers) have started to have to do maintenance work on the code. The members of our team have a broad range of experience including VB, c++ and Java. Since our delphi programmer left management has resolved never to undertake another project in delphi ever again on pain of death due to the problems it has caused. Some members of our team are convinced that delphi programmers only continue to use it for the job security that it brings.
  • I see absolutely no disadvantage to being able to layout the application GIU visually, rather than creating widgets in code and constantly having to fool with the metrics and go through endless fiddle/compile/run cycles until it looks right. In particular a nice RAD IDE like Delphi makes life very easy, with decent layout control like Align and Anchors, though they could add a few more options such as percent of window.

    I used to be a C++ zealot until about 1997 when I got the first taste of Delphi 2, and I never went back. Sure, I had to switch languages, but I've made bigger sacrifices before. Besides, OP leads to a clarity and readability of code rarely achieved nowadays with C++, expecially when dealing with COM.

    RAD from the perspective of Delphi gives you the best of both worlds: easy GUI design, as well as a great class framework and a powerful language. It ties the two together in such a slick and unobtrusive way; it's the way I'd probably do things even if Delphi didn't exist. Sure it adds some overhead, but then every framework does--even MFC, if you can call it a framework. If you want leaner apps, you pretty much have to drop to the API/message loop level, and we've all been there and didn't like it. Compared to something like VC++ that gives you RAD Light and forces you to keep your GUI and event handlers synchronized manually and always keeps you worrying about resource IDs, Delphi takes you to a whole other level of productivity.

    The project manager at the company I used to work for was a great MS zealot, and his big mission in life was to convert the shop from Delphi to VC++. We were five developers churning out the code of ten and still were constantly behind schedule (talk about over-commitment of sales). We pleaded with him repeately to reconsider, explaining what a great work multiplier Delphi is and how such things are important to a small company. His retort was always that VC++ is the tool of the future and that tons of very smart guys work at MS and they can't all be wrong. He constantly denied that there were any productivity differences between Delphi and VC++ and accused us of just being irrational Delphi zealots (which probably was true, but with good reason we felt).

    Anyway, these kinds of attitudes--influenced more by the great MS marketing machine than reality--are what is keeping Delphi down. I guess all of us Delphi supporters are hoping that Kylix will be the second wind for Delphi. With the momentum Linux has at the moment, and its lack of a serious RAD tool, Delphi should stand a decent second chance. And please don't anyone mention .NET or I'll have to hunt you down and decompose you into little bits.
  • >. How would a stock (haha) Linux workstation get dbExpress onto it?

    For instance by having the driver linked right into the application?

    IOW, *no* issues.

    BTW, libc issues are just great. It is amazing how much breaks inside libc (2.1.3!) once you do slightly more advanced programming. Locale functions, for instance? Oh my god. Try reading the lates (2.1.95 beta) Changelog.
  • Crack open the Champagne folks! No Paradox must surely mark the beginning of the end of a terrible product concept (file server databases). Roll on proper relational databases! Roll on object databases! Hasta La Vista Access! Adios Paradox!
  • Oh, yeah, Linux is the OS of coding gods. Each single line of code is laboriously reviewed by the Open Source community. Each line is carefully crafted, tested, and everything is so amazingly well designed.

    How come I cannot really believe that.

    I have seen enough code (on *any* operating system) that increased my blood pressure. It is no better and no worse on Linux.
  • I _KNOW_ Delphi 5 does - I've used it.
    Not to mention that with a little pointer work - all you C programmers should approve - It can even be done in D1!

    Life's a bitch, but so am I.
  • Well put.
  • _But a RAD such as Kylix may spells trouble for Linux_

    Why? You worry about bloatware and Kylix introducing "fun-time programmers" to Linux. In my experience, fun-time programmers are not the ones that produce bloatware, it's the commercial software companies. Lighten up, anything that encourages ppl to develop Linux apps is unequivocally a *good thing*

  • Apart from the Delphi developers, Kylix also brings hundreds of thousands of applications, built in Delphi today, ported with Kylix tomorrow. On the first night that Kylix ships, we'll probably see more new Linux applications than we've seen in the past few months...

    Gee, thanks! Let's hope these new apps are better tested than that. Even if they are not commercial apps, we can hope that someone will port their $30 shareware file-splitter! Wow, it's about time we got some of those sweet windows apps linux has been lacking!

  • *groan* I've been making a living in the programming world for well over 15 years. If you don't know how to make money in the software industry, then who is it that is misguided? _I_ didn't say Linux programmers were poor - the linux programmer I responded to did. I don't know any. Who said commercial software? I do contract database work and I do very well. HOWEVER, commercial software is what makes an OS successful. It's not a question of open minds. I think we're talking about two different things. In any case, I hope your attitude isn't the norm for the linux programming markets. If it is, then I don't _want_ to get into it.
  • Well, this stuff is about as good if not better, then your typical "GTK based frontend" served daily on freshmeat.
  • Since the site has been slashdotted, can someone enlighten us on what Kylix is? Thank you.
  • You might like to try CodeCommander [sourceforge.net]. It was coming along nicely when I last tried it (a couple of months ago, version 0.9 something), and was fine for php, bit of c or java etc.

    Personally I use JEdit [sourceforge.net] nowadays, but if you don't have a decent amount of ram and a fast machine go for CodeCommander.

  • I do. It's really not that difficult if you know C, though I learned Pascal first (language of choice in the BBS world).

    And anyone who used TurboVision in BP/TP7 will have no problem with Delphi/Object Pascal. It's really powerful.
  • Wait.. something's missing in this article...

    Content?

    --

  • As a long time Delphi developer, I'm practically drooling with anticipation. The only major problem I have is Borland's announcement that Kylix will have no support for Paradox tables. Which of course 99% of my existing Delphi applications are based on. ;-(

  • OK, Kylix will be Qt based, but on which version??? v. 2.x (which I really hope so) or version 1.4x (which I greatly hope NOT)?

    All this talk about KDE integration and such leaves the question in the air, as the current "stable" KDE is Qt 1.4x based, but the new 2.x version is much, much better.


    --
    Marcelo Vanzin

  • The first release of Kylix is the Delphi 5 IDE with a new cross-platform component library called CLX ("clicks"). It is very cool. I've seen non-trivial Windows/VCL apps ported from Delphi 5 to Kylix in very short periods of time (a couple of hours to a couple of days, depending on the app).

    There's no real difference between a programmer with no real Linux experience doing Kylix and a programmer with no real Windows experience doing Delphi. In both cases they're junior programmers who learn from the more experienced. The senior programmers, on the other hand, will have to rely on their C knowledge in order to glean useful information from the vast amount of C/C++ code that's available for Linux. This is no different than the early days of Borland Pascal for Windows and (later) Delphi.

  • The language and/or IDE seems suspect in a few areas? Can't you really be more precise than that?

    I come from a very strong VB background before moving to Delphi. I just couldn't continue with VB and let me tell you a few reasons:

    First, speed. VB apps are dead slow. They may get some impressing numbers in carefully chosen benchmarks but otherwise it's a lot slower than Delphi. In anything heavier I got up to 10 times speedup by moving to Delphi. For example, get the ascii value of a char in a string. Delphi just gets the value with aString[index] while in basic you have to asc(mid$(aString$,index,1)) resulting in creation of temp string.

    Second, braindead decisions. VB stores all strings internally in unicode. Try to send a string to a dll and you'll see how VB first converts the string to ansi. If you send an array, each and every string in that array are converted. That's a huge penalty and there's just no way you can change that.

    Third, you can't do anything even a bit more advanced without kludges while Delphi provides clean ways to do anything. You just get used to working past the problems without understanding how dreadful that makes your code.

    In the end, I hated myself for not changing earlier. Delphi gives you the same variable length strings that VB-programmers love. The syntax is not that much different either so it's easy to start. But when you understand the component model and learn real object oriented programming, Delphi is just wonderful. VB can never get there.

    So, blame that programmer for spaghetti code and don't blame the tool.

  • Re:Slashdotted... (Score:3)

    by CmdrGordita ( 227448 ) on Sunday September 24, 2000 @06:30PM (#757264) Homepage
    Project Kylix - Native Rapid Application Development for Linux.

    What is Project Kylix?

    The goal of Project Kylix is to produce a high-performance Rapid Application Development (RAD) development tool for Linux. RAD here means component-based, two-way visual development of your user-interface (GUI), database, internet, and server applications. Development tool means a high-speed native Delphi/C/C++ compiler for linux. Project Kylix should also simplify the porting of existing Delphi and C++Builder applications between Windows and Linux
  • Borland already has the JBuilder [borland.com] development enviroment out for Linux. Although this is Java based, the environment brings the familiar Borland's power IDE interface. It works like a dream with proper hardware under Linux. The Foundation release is free for use. You need to download a license key from the site though.
  • I disagree on two counts.

    But before I go on, let's distinguish what we're talking about here. The RAD tool Kylix (and others like it) are not the total of RAD methodology. Your argument seems to be that the RAD methodology is bad, and seems to conclude that RAD tools are too. Both, however, have a place in development.

    RAD tools belong in programming for the same reason the STL and design patterns exist; to keep developers from reinventing the wheel. They also allow junior programmers to be productive while their seniors concentrate on problems that cannot be covered by existing tools. I don't know about you, but coding static web pages bores me, and if some tool can competently code what our designers dream up then I'm all for it.

    The RAD methodology also belongs on a programmers resume. Every time a project is desinged you'll find that the functionality your customer wants cannot be crammed into the time you have. So, rather than say "If I can't do it all I'm not going to do it," you compromise. You have your customer prioritize his requirements, fit the highest into your project, and offer to do the rest as a follow-up project.

    Yeah, RAD methodology may cause programmers to rush things and that's bad. However, that's not the fault of the methodology, only it's implementation. If you think it's forced you to cut corners and reduce testing time, you didn't scope the project out properly to begin with. Every methodology will fail under that circumstance.

  • > Are you thinking of method overloading?

    I would imagine I am, as I cannot see a reason for operator overloading, and anyway, I still use Inc and Dec, as that is the reccomended Borland way, so I am unlikely to ever overload an operator.

    One of the major points to using Object Pascal over C++ is readability, so altering the way an operator works would be (IMHO) a retrograde step.

  • Rapid Application Development... (Score:4)

    by Phokus ( 192971 ) on Sunday September 24, 2000 @06:32PM (#757268)
    Does anyone here agree with me that R.A.D. is really a bad reflection of our times? In our fast paced economy, we MUST do everything fast in order to survive. Programming used to be considered an art, we used to actually put our ideas on paper before we even entered a key stroke. Now what do we do? We spew out half baked code in order to meet demand.

    Honestly, this is more of a pity post more than anything else. I work for a company that almost demands that i put speed ahead of quality, and i have friends in the same situation as well. Anyone else have this problem?

  • I hear ya! I just have to convince the people around here of that.
  • I want to write my program this way.

    No, you can't write it like that.

    But really, my way will work.

    No it won't, 'cause I said so.

    But who are you to say what works?

    I'm the compiler. That's why.

    Fuck fuck this. I'll use C instead.

    I don't care. This is academia. I've got tenure.

  • What I'd like to see most particularly from this long-awaited "mass-production" environment, is a drastic reduction of (headaches from) the notorious proliferation of grossly overlapping libraries for Linux. With many more people using what amounts to a "standard" environment, one can foresee the library scene cleaning up these gray, fuzzy borders into more sharply delineated borders in which one library much more clearly does this, and another much more clearly does that, with way less ridiculous overlap from almost identical functions (not even functionality, but functions!)

    Many of the best Linux programmers, seeing an opportunity to actually make some money selling excellent special-purpose libraries (sorry, rabid "free-software" dudes, some people want to make a living at what they do just like everyone else and not live on cheap noodles and dumpster leavings just to appease the Angry Gods of Open Source), will stop re-inventing the damn wheel and start creating original, useful libraries to meet ever-more specialized demands.

    This could be a "very good thing" indeed, if the shipping product is not grossly overpriced. Too many of the best Linux programmers are poor students or very much part-time volunteers working on pet projects, who won't be willing to pay through the nose when they can obtain so many other, totally free, development environments.

    (This has been something of a rant, I guess. Wading through hundreds of, and carefully selecting, Debian packages one by one by one, over weeks on-line, and discovering how very much duplication there is in basic Linux libraries even at Debian, did something weird to my mind).

    Man, I hate this stinking cruft so, so bad.

  • Interesting - I found the Delphi IDE to be easier and more straight forward than VBs. Of course, I looked at VB6 AFTER I had been doing Delphi since D1. Could it be it's just not what you are use to? If you're coming from a VB background, no offense intended, but that's probably the problem. As far as converting ASP to Delphi, that was probably pointless from the start. While Delphi has Web tools, I'd either leave it in ASP or use Cold Fusion and extend it with Delphi. THere is little doubt Delphi is a more efficent language IF you can past the learning curve, which is still a LOT less steep than say C++ (or that C# crap) Is it as easy as VB? Nope. But there is a price for ease.
  • I can certainly see where it would rub you wrong - I have had the same thing happen to me from some of the posts on this thread (bascially insulting windows programmers *grin*). All I can say is I am _sure_ the windows programming market is larger and pays better. And that is precisely because more people use it. And that is because there is more software readily available (and easy to use & install). I really want to see Linux take off. My biggest fear right now is that Linux may have waited to long to get in the game. There is one thing that Micrsoft does right - and that's learn from past mistakes. They WILL revision a piece of software until it surpasses most items on the market. This has happened time and time again. And that's what they are trying to do now with NT. Win 2000 (and I HATE Microsoft) is pretty cool. Installs easy. Runs easy. Have had little or no problems with it. I really think (IF it works) that Delphi coming to Linux can be a very strong positive for it. It scares me even more to see the attutiude of MANY of the posts on this thread. My buddy said to me it reminded him of the attutude a lot of OS/2 coders had. It's like many of them seem to enjoy that Linux is a niche OS and would rather it not go into the mainstream. Some kind of martyr syndrome or something. I guess if your mainstream you can't take as many pot shots as Windows. I dunno.
  • Does anyone else remember a year or so ago when another major company decided that Linux was the future? They talked about how Linux was going to be this, that, and the other thing. And that company was moving their products to Linux, and really staking their future on Linux. Now that company is in the midst of what can only be called its death throes. Sadly, we remember today the company called Corel. And the great promise that we saw in them making things available on Linux.

    I do not wish to seem negative before Kylix is even here, but I really have to wonder. It doesn't seem to me that Borland/Inprise is staking their future on Linux to quite the degree that Corel did, but what if Kylix isn't what's its cracked up to be? How would that affect the future of the Linux community? I approach this matter very seriously, because my livelihood depends on it. I work for a development firm that is strongly grounded in Delphi, and there are more than just rumors about shifting over to developing with Kylix. It concerns me greatly that we will be depending on Kylix on Linux, when nary a Delphi developer in the house has any real world experience with either Linux or any type of Unix.
    I, for one, will have to wait and see whether the buzz about Kylix is really worth checking into...
  • You might be able to use the fact that a Borland R&D guy is saying this publically in a *linux* forum to your advantage. :)
  • You are a knot-head.

    "why is RedHat getting so rich off of a free product"

    Who says they are "getting rich" ? Have you examined their revenue stream, comparing it to, oh, say, Microsoft, or Oracle?

    "I'm sick and tired of people thinking they can make money off the open source movement"

    You should not only be friendly and gracious to these "people", you should be thankful that these companies/people have devoted the resources they have towards bettering the Linux you use. (If you do) Where do you think the almost user-friendly RPM came from?

    Have you in ANY WAY contributed to the "open source movement" other than to threaten to burn CDs?

    Somehow, I doubt it.
  • Turbo Pascal and Turbo Assembler are how I learned to code. I eventually learned Turbo C++ for a C++ class (but never took a class in C). Haven't used pascal in...over a decade...lol, but the enforced structure wasn't that much of a pain. Its an easy language. :-)

  • IDEs (Score:1)

    by ACK!! ( 10229 )
    First place I ever worked as tech developed their app in Delphi of all things. It was not bad(the app that is). Borland had a rep back in the day of making decent products so I hope this works out for them.

    My question is simple though, has anyone tried out Codewarrior for Linux? Is it worth the time?

    A couple of people have bemoaned the lack of good IDEs for Linux and other have all responded with the typical KDevelop replies. I don't like it that much but then again I only do C,C++ coding on small utility style projects so I am not the big project oriented style programmer that this thing was made for.

    I just want a nice, easy to use editor with sensible color coding and easy compiling and debugging tools. I don't need the revision and project tools at all. Is CodeWarrior worth the time? I know a couple of Mac hacks and people stuck programming Windoze stuff that love the thing.

  • Delphi is cool (Score:3)

    by robinjo ( 15698 ) on Sunday September 24, 2000 @08:25PM (#757279)

    I've spent the last 1,5 years programming in Delphi and it rocks. I love the language. It's easy to learn, produces fast and small executables and helps producing high quality apps.

    People coming from a Visual Basic background will notice that that Delphi is just as easy but it's a lot faster and really made the right way. The class libraries make sense and you can create new components really easy. After programming a few years in VB, there's no way I'll go back to the old monster. Delphi is also a lot easier than C. The language helps you avoid errors. I suppose Object Pascal is a good language to learn before C(++)

    But the greatest thing about Delphi will be Kylix. It will be a pleasure to port our software for Linux. I really think Borland did a great decision. Kylix will give them a good start as Linux still lacks Visual Basic. It can really become the state of the art RAD tool for Linux and that will also help them gain markets in the Windows world.

    Oh, you may also want to have a look at Lazarus [freepascal.org], which is open source. It looks like Kylix will be ready first but Lazarus and the Free Pascal Compiler [freepascal.org] look good. I already use FPC for small apps and hope to get a strong alternative for Kylix which would boost competition and quality in both.

  • Yeah, Kdevelop 1.0 didn't impress me, and the 2.0 candidate is a buggy mess right now. I'm writing embedded C, and the C++ browser is not useful.
    I guess what would be really good is an open source Visual Slickedit... I'll take another look when they say it's done.
    (Besides that, I don't like KDE or GNOME - plain X, with icewm for me)

  • C'mon - be fair, it would have been the same if the programmer had been a COBOL (e.g - don't flame) programmer or used any other language - the best tool for the job is the one you know best, or the one most of you know.

    I would dispute what some members of your team are convinced of - I use Delphi because it is a hell of a site easier to debug than C (or C++) and just as powerful IMHO.

    I would say that your management have made a good descision based on their current skillset and hope that they would reconsider, if they found themselves with a different skillset.


    Life's a bitch, but so am I.
  • We've been waiting ages - and still no official release date AFAIK - somebody spill the beans and let us know when to expect it in the shops.
  • >Wait.. something's missing in this article...
    >Content?

    I'm sorry, you've obviously mistaken this site for kuro5hin.

  • Re:Sure, here's how pascal works: (Score:3)

    by Sneakums ( 2534 ) on Monday September 25, 2000 @01:46AM (#757284)
    Beside, any language that relies on pointers should be shot immediatly

    Leaving aside for a moment the question of exactly how one would shoot a language, I must point out that pointers are not bad. What is is bad is the unthinking use of random pointer arithmetic.

    It's worth noting that every language that is any use at all uses pointers: Lisp, Pascal, Java and of course C and C++. They differ in how exposed to the programmer the pointers are. Java, for example, has an annoying impedance mismatch where objects are passed by reference but things like ints and chars are passed by value.

    --
    "Where, where is the town? Now, it's nothing but flowers!"

  • Kylix sounds like Kylie Minogue's ancient gaulish ancestors. discuss.
  • An important point you missed is that
    Kylix will also support C++ like C++ builder
    in the next release. That's when the large
    takeup will happen IMHO.

    Note this will also mean that Inprise would
    sell more C++ builder/Delphi for doze also,
    as this would allow people to write progs
    for both platforms, which will really be needed
    over the next while.

    Most people will not port from doze to Linux just
    because of the enhanced stability etc. They will
    need an easy transition to Linux that Kylix will
    provide.
  • As a long time delphi coder (D1) I have one concern: 'Windows' Delphi was practically non-existent at the last Borland conference. While I am ALL for a Linux version (and very excited) I am very concerned (read: my clients are very concerned) that Borland is bailing out of the Windows market (let's face, the market share is slowly eroding) Scary scary - VB still sucks.
  • let's keep Linux only usable by those few people with too much time on their hands. It's this kind of attitude that has kept Linux in the closet. There are too many tech guys ( and I AM a tech guy) that want to keep their little piece of non-windows in a little shell. It's easier to bang on Windows that way. Look, one of the big reasons I don't bother with Linux is because, AHA!, my clients have ZERO interest in it. Why? Too hard to maintain and NO APPS. That's right, NO APPS. And don't give me that 'oh it has a lot of apps' Compare to windows, it doesn't have CRAP. And THAT is what my clients see. We aren't talking about practical. We aren't even talking about rational. We are talking about the people that PAY THE BILLS. And I make a _lot_ of money coding in Windows. And thanks for the underhanded insult to my Delphi coding ability. Believe it or not, Linux coders are not the only quality coders around (and some would say, given market share, they must not be too great) Catch a clue - Things like Kylix are the only chance Linux ever has of not just being a niche file server/web server OS. So get off your high horse. Linux ain't all that great. And neither is Windows. If you guys keep SCREWING AROUND, Microsoft is going to run you over. And I would LIKE to see Linux take off. Apple manage to elite themselves into the ground - Linux is the next in line. And people with your attitude contribute to the problem.

  • Regarding your .sig... if you are referring to the French Revolution, it was the revolutionaries who were doing the guillotining.

  • And this is why so many Linux programmers are poor. There's a very slim contract market. And this is going to go away until there are better coding tools. It's going to be rough to start off with, but it will pay off in the end. I make a lot of money coding in Windows. I'd love to do the same in Linux. Right now, no way I'm going near it. With Kylix, I may take a look at it. Not trying to say my opinion matters, but I imagine I'm pretty typical of a lot of Windows coders. And the more coders, the bigger the market - I steer most of my clients as far as software and OS that they use - and I have yet to have a compelling reason to use Linux. It would be nice to have one.
  • Comment removed based on user account deletion
  • dude, it is kylix, not kylnx.
  • >>Anyway, these kinds of attitudes--influenced more by the great MS marketing machine than reality--are what is keeping Delphi down. I guess all of us Delphi supporters are hoping that Kylix will be the second wind for Delphi. With the momentum Linux has at the moment, and its lack of a serious RAD tool, Delphi should stand a decent second chance. And please don't anyone mention .NET or I'll have to hunt you down and decompose you into little bits. <<

    Excellent points made. MS wogs have been talking about the demise of Borland for over 12 years. Every time they feel threatened the hype starts again. Just watch when Kylix gets out, every MS loyalist will start talking about how Borland is getting ready to go out of business or is gonna get bought by (insert your favorite company to hate here). And of course they heard from a friend of a friend at Borland.

    Really pathetic.

    Shadrack
  • I've been programming in Pascal in one form or another since Turbo Pascal 3.0 from Borland came out. Recently, I've crossed the line back into C++, for the work I'm doing. And I hate it.

    I was under the assumption that the tight-nit Delphi group (see the Delphi Super Page [borland.com]) existed as a support group for Delphi programmers to be able to keep up with what was going on in the C++ community. I could not have been more wrong.

    Clumsily traipsing into the MSVC++ environment, I thought that the best thing to do would be to use the classes that were available - surely they followed the "standards" of the C++ community, and they would be the best tools to use. So, I innocently wrote all my tools based on CFile, CStdioFile, CString, and the like. What a frickin mistake.

    CString is its own abomination, not available anywhere else on any platform, as far as I can tell. I was astounded to learn that the string library that MSVC++ pushes on people is not portable. Until I remembered that the "MS" is Microsoft. Of course they want you to use classes that can't be ported! Duh!

    CFile crashes in situations that I've never seen fopen crash. How stupid is that?

    Anyway, the thing that, on reflection, impresses me the most about Delphi is that when you pick a tool to use, you're probably using the right one. No, Borland doesn't subscribe to the ANSI Pascal standard - but I've never met anyone who prefered ANSI Pascal over the latest and greatest Borland product.

    Yes, it'd be better if Delphi (read Kylix) were standard, free, portable, and already ported to every system on the planet. But I'll be happy with it, when it comes out. It'll instantly make me a /competent/ Linux application developer. Not necessarily a great one, but I'll be competent, instantly. Unlike my situation in C++. If I took my MSVC++ smarts, and tried to use them to code in gcc, I'd be lost. Completely lost. I'd essentially have to start over, and re-learn all the tools I thought I knew how to use. And I've seen the same thing happen the other way, too. Linux / gcc coders scrambling to get anything to work in MSVC++.

    When a newcomer uses Delphi to code Windows applications, they are a lot more competent a lot more quickly than someone in the same situation, using MSVC++. I hope, hope, hope that the situation is the same with Kylix on Linux.

  • How many people here know pascal?
  • Do they even teach pascal anymore? I heard my old highschool phased it out...
  • 'Windows' Delphi was practically non-existent at the last Borland conference

    Not in the spotlight maybe, but hardly "practically non-existant" (IMO).

    Even though it was officially an "unannounced product" the Delphi 6 beta was used in several sessions. And, of course, there were many of the usual Delphi sessions using D5.

    Can you give more details about what gave you this impression? The lack of a Delphi 6 announcement or something?

    -- CP
  • Bad developer = bad language? And you claim to have experience in VB,C++,Java. Accusations of Job security? How lame. The only reason your employer would feel that way, is because you wanted them to.

    Hmmm... So it's already started. The FUD from the MS wogs is gearing up for the release of Kylix.

    Look out people,the better Kylix is, the worse the lies and inuendos will get. VB programmers will not idly stand by and allow their future to be threatened.

    shadrack

  • The all important screenshots... (Score:4)

    by baywulf ( 214371 ) on Sunday September 24, 2000 @06:38PM (#757299)
    http://www.drbob42.com/kylix/hotshots.htm
  • Near the bottom of the Kylix slide show [drbob42.com], there is a link to Dr.Dobbs Kylix news [drbob42.com]. The news item from 2000/08/18 states the following:
    According to
    reliable sources [wininformant.com], Microsoft is beginning to (hire another company to [wininformant.com]) port some of its applications to Linux. Starting with Internet Explorer and Windows Media Player 6.3 (which is already ported to Sun Solaris).
    The 'other company' is MainSoft [mainsoft.com]. Apparently, they did the IE port to Solaris, and are currently porting Windows Media Player.

    Hmm. Has this been on slashdot before? Has anybody else heard this news? And why would Microsoft do this (if it's true)? What would they have to gain?

    Certainly they do not want to assist in bringing a free OS to mainstream desktops, do they? Is there an evil plan behind this, or do they just need things to spend money on...

  • Mirror over at my web site: http://www.johncglass.com/mirror/kylix.h tm [johncglass.com]

    P.S. Is Hemos drunk? He's been posting some wierd stories today..

  • I see absolutely no disadvantage to being able to layout the application GIU visually, rather than creating widgets in code and constantly having to fool with the metrics and go through endless fiddle/compile/run cycles until it looks right.

    There's one big problem that I have with visual GUI layout: you can't do dynamic interfaces that get layed out at runtime.

    At the place where I work, we have been using a proprietary language for DOS called Clipper that is pretty primitive in the user-interface department, but that actually turned out to be a good thing in the long run, since it got us to write our own libraries to handle things right.

    Unfortunately, we're now getting forced into more and more Windoze stuff, and being good unthinking Clipper loyalists (*sigh* you can tell what I think of this move), we're slowly switching over to the language CA currently is trying to sell: Visual Objects. It's one of these RAD things where you draw your interface forms using a graphical layout tool.

    It sucks. We can't even do some of the same interfaces that we could with Clipper, because everything has to be known at compile time. You can't have the interface depend on the user's data or configuration. (Well, not unless you write your own GUI libraries instead of using the visual crap that comes with it. No one has done that yet.)

    You know what approach works best? It's the one I've seen on the Amiga, with toolkits like MUI, BGUI, and ClassAct. The idea is that you write text (or build arrays) that specifiy a logical structure to the widgets. It's just as fast (maybe even faster) than using visual layout tools, more accurate, doesn't require any repository or non-human-readable "resource files", and the data structures can be built at run time if you need that. It fucking rules! So, naturally, since the Amiga had it in 1992, the Windows world should eventually move up to this step sometime around 2007.

    (BTW, Linux dudes: is this this how toolkits like gtk and qt work?)


    ---
  • 1) Microsoft knows they can't avoid being broken up and the application section is quietly getting ready to slit the throat of the OS section.

    2) Microsoft plans to put out substandard software for Linux and then claim that weaknesses in the OS make it impossible to provide good implementations. More chances to slam Linux while appearing to diversify to keep the DOJ happy.

  • a 400MhzpII with 192 MB is by all means a modest PC if programming is your profession. Really, a programmer earning between 50 and 100K annually (pessimistic for good programmers) should have some decent equipment to do his/her work. Contrary to the common belief, memory is much more important than processor speed.

    I think it is very brave of you to even try running java 1.2 on a 32 MB PC. I recall running jdk 1.2 beta1 on a 133 Mhz machine (64MB, it doubled as a server machine). It worked fine, I could even launch the swing demo (for non Java people, this loads just about all available swing components). Of course the application response was terrible, but hey what do you expect?

    However, with sufficient memory (192 MB is about the mininimum for JBuilder) it should work.

    JBuilder refusing to install with a newer JDK is entirely borlands fault. Apparently they have some dependencies on non standardized parts. The windows version of JBuilder 3.0 came with its own JDK. Replacing it with the superior 1.3 beta from sun (there was no final version yet) wouldn't work and was not supported by borland. Even the released 1.2 did not work. You had to use the borland JDK. This was one of the reasons I refuse to use IDE's such as those provided by borland or IBM. You always get locked into obsolete software at some point. Don't misunderstand me, both deliver excellent products. But if you want to use the latest and greatest, you'll run into problems. Visual age for instance used jdk 1.1 long after 1.2 was released. Apparently it is now possible to use 1.3 with it.

    So I have to agree with your last line: simplicity is better. Don't get locked into some IDE. Always keep the way open to use something else.
  • This is great - but what would be even better is a MS Visual C++ killer for the Linux desktop. The article looks very promising, and I hope they do it right. A truly awesome IDE is more important (and contributes more to programmer productivity, and programmer enjoyment) than just about anything else.

    I'm very happy to see them basing some of their product on Qt - in my (brief) exerience it's great.

    I've been looking at all the Linux IDEs over the past week and they all have their problems. A lot just aren't available for PowerPC. Most are ugly, clumsy, or are simply a poor wrapper around GCC/GDB. The biggest problem is that integration of the debugger justs isn't as good as DevStudios. Maybe the best of all the ones I have used is Code Crusader/Code Medic - but at best it's 75% of what Microsoft offers.

    The lack of a great IDE is a significant disincentive for people to switch platforms. Sure, we can all do makefiles and command-line gdb, but I don't know many professional programmers who don't acknowledge DevStudio as one of the best environments they have ever worked in... and once you've had the best it's hard to give it up, even if you do prefer X-windows to MS-windows.
  • I am very concerned (read: my clients are very concerned) that Borland is bailing out of the Windows market

    I spent most of the spring and summer working on features for a Windows product; we are not bailing out of the Windows market. However --- the goal is that the VCL which uses QT widgets will work on both platforms, and ship in both the Windows and Linux products, which has obvious implications for scheduling.

    Robert West
    Delphi R&D

  • Kylix is... (Score:3)

    by Dacta ( 24628 ) on Sunday September 24, 2000 @06:43PM (#757318)

    For all those asking (because the site is /.ed) Kylix is Inprise/Borland's version of Delphi (and later C++ Builder) for Linux.

    It includeds a (Qt based) visual form designer, an OO-language (Object Pascal), great database tools and a cross platform OO class library called CLX (which supplements the Windows only VCL).

    It will support application developement for KDE, GNOME (or at least GNOME aware, but non-GTK apps) and Apache modules.

    More info at http://www.borland.com/kylix/ [borland.com]

    I believe that Kylix is going to be a breakthough app for the Linux desktop. Many large companies have many apps written in Delphi (and VB). Rewritting these apps in C/C++ isn't an option, and Scripting languages like Python/Perl don't have the tool support for writing these GUI database applications. Kylix is going after that market, and knowing Borland's reputation for writing good programming tools, I'm looking forward to it.

Slashdot Top Deals

Only God can make random selections.

Close