This entry was posted
on Thursday, December 28th, 2006 at 8:11 am and is filed under Uncategorized.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
probably not the slashdot you would have hoped for. The slashdot itself was aimmed negatively and the article it linked to was pretty negative as well.
On the whole, were I to read the article with no Delphi experience, it would do nothing but discourage me from even bothering to look at any of the codegear products.
2. Nick, in the spirit of "extending and innovating on top of Open Source": CodeGear should consider offering "innovations" on top of the FreePascal compiler and possibly even the Lazarus IDE. In addition, it seems to me that many of the non-visual Delphi components and classes could also be used to build outstanding extentions to the dynamic languages you are targeting, similiar (what about ECO?) to what this guy is doing => http://members.chello.be/ws36637/php4delphi.html Supporting the FreePascal compiler would make that a cross-platform reality. If your going "Active" languages, you’ve got to be cross-platform or you will be the laughing stock of the development world.
I have to admit the stance of embracing open source efforts while simulaneously quashing the one project open source enthusiasts most want is something of a confused stance.
That said, I do COMPLETELY understand how a business wants to see money for its efforts, and I understand why they like free work done by others.
Clearly, however, Kylix shows us how shortsighted and doomed for failure that mix can be. Makes me wonder how the same company can say that Kylix is a financial failure and then try to embrace PHP - that thought pattern is just too Borland for my liking.
Novell put 800 engineer to support Suse Linux 10. Suse Linux 10 is simple, cheaper, powerfull and very attracktive for corporates.
Even MS considers Linux ( http://www.novell.com/linux/microsoft/allsystemsgo.html ), how about you? Come on Nick! I totaly believe that CG want to make correct product. But if you continue in this way, developers will begin to think CodeGear’s thought is "Where Money Matters!"..
Why does everybody always keep screaming for cross platform support? The strength of Delphi _is_ GUI Database apps (at least imho ). These apps are never (ok, never say never ) going to be big on nix platforms. I have seen only very few big companies where nix is used as a desktop environment. So why invest a hell of a lot of money in something you know will probably never hit it off? I think killing Kylix (come on, is this really news?) is a good decision.
CodeGear should take Delphi further in the direction of things it’s good at. That goes for anyone by the way: do what you’re good at, don’t do something you know you have a big disadvantage in ever getting good at.
Nick, is the sacking of 120 people true? could you clear that up for us? those tiny comments are what makes me worry and it would be good to have someone trustworthy comment on that, not some random uncheckable comment from an anonymous guy…
Crossplatform support is important for those of us who write server apps.
Delphi/Objectpascal has just too many advantages over other langugages (fast compile times, easy to read, no runtime) to let that opportunity pass by.
I’m not saying we need another Kylix, but some kind of "native" support for Linux would be great. For example, it would be great if we could write Freepascal Linux server apps with BDS. Isn’t it possible to add a Freepascal "personality"?
I think adding a FreePascal personality would be a phenomenal move, exactly in line with the "embracing the open source community" thing. Also it would be embracing without potential loss of customers, since it would open BDS (or is it CDS? ) up to a whole lot more devvers out there. Just depends on how hard it is to get it done, and how high it is on their wish-list
I find Ben Smith’s comments concerning his reasoning for killing Kylix particularly troubling.
Historically when a software company is seeking a foothold in the marketplace it performs SERIOUS market research to find the growth market(s). Then it positions itself to ‘ride the wave’ of the growth market(s).
While M$’s growth is slowing, Linux growth continues to push the envelope. To ignore the enormous markets that Linux has developed and is continuing to develop because of revenue figures from - what, more than 4 years ago? - is idiotic.
No, this isn’t good business sense. It’s the fear that is born out of ignorance concerning a particular market segment. It’s a reflection of a lack of market research performed by Mr. Smith and company.
It also reflects the same executive decision making pattern perfected by Borland’s brilliant executive management team - Windows is the ‘hype machine’ of the day, so let’s just hang off of their coattails, eating the ‘crumb’s from the master’s table’.
Please Ben, call up Larry on the phone and tell him what a bad investment risk Linux is for the database/software development market. I fear you may get an earful, once the laughter and derision decides.
Perhaps you’d like to explain to RedHat or Novell why they shouldn’t be successful at what they do either.
It is unfortunate that the only thing that holds back Linux from grabbing major desktop market share is a lack of a good dev tools platform, and the company most optimally positioned for filling that void has descended into the depths of cowardice and fear.
I believe that we will see the emergence of an excellent - or at least very good - dev tools platform for Linux. Unfortunately it probably won’t be coming from Scotts Valley.
Once that occurs, Ben, what will you excuse be then??
IMHO it’s an error to look at Kylix only for Linux applications, you will not find the revenue you expect from there. But you can think on Kylix as a nix/OSX dev platform, now you have a much larger market, OSX market will spend more than the *nix one.
, which looks pretty great. But guess what. It’s available for C++, Java, Ruby, PHP, Python, C# and Visual Basic. No Pascal to be seen. Delphi needs to get out of its ever shrinking Windows niche, otherwise things will not be looking to good in the near future, IMHO.
@Delphi Bigot: face it, *selling* developmet tools to Linux developers is an hard task. Even "Larry" released JDeveloper as a free tool.
He can sell Oracle simply because no OS database available on Linux can compete in the high-end segment with it. But in the low-end segment (covered by MySQL and Postgres) they had to relase Oracle Express, and tried to acquire or acquired some OS stuff.
Oracle is so expensive that it makes Linux an interesting cheap option as the OS for IT departments, although RedHat prices are not very cheap, especially for the AS version. And Oracle is certified to run on RedHat and Suse - try to run it on a Debian - it could be even difficult to install and you get an unsupported installation.
For a development environment like Kylix, it was and it is very different. Most Linux developers are C++ guys, Java guys or PHP/Python/Perl guys. They already have a lot of free tools. They are used to them, and many make a religion of using those tools. They are used to download source code and compile it locally - and often under Linux is the only way to have a working app due to dependencies - sometimes it looks to be far worse than "DLL hell" under Windows Do they need a RAD tool, especially one using Pascal as its language? Right now, it looks they don’t. I never see a real demand for Linux desktop applications. The lack of a common widget set and desktop does not help the common user and then demand. Also, Linux is still more difficult to install and to manage especially, and you can’t use any hardware you want. Windows grew *before* RAD tools were made available.
IMHO, there could have been a market to move some server side applications to cheaper Linux boxes. Borland could have made available a way to Windows developers to move not their client applications, but their server modules to Linux - as long as they could, something Kylix didn’t deliver.
Kylix should have been an option for Windows developers, trying to sell it to the gcc people it’s just a waste of time.
Kylix was a failure, for several reasons, noone can deny it. It can’t be resurrected as it was - CodeGear needs a different approach to other platforms.
Let me reply to your comments below, by stating that in order to reply I have to make assumptions about your experience with Linux/Kylix/Oracle since I don’t know what your experience level actually is. However I can assure you that I am sufficiently authoritative in a technical and business sense in all 3 of these areas, as well as in other areas relevant to this discussion.
>> face it, *selling* developmet tools to Linux developers is an hard task.
>> Even "Larry" released JDeveloper as a free tool.
I would have to agree with this statement on it’s face. However this statement leads me to conclude that you either don’t understand business, don’t understand the application of Linux technologies to business, or both. This would seem to be consistent with the larger Delphi programming community. However this most definitely does *not* apply to the whole of the Delphi community. Some of the most capable people in ‘our’ community are quite adept at cross-platform development.
The goal of being successful at selling development tools to developers is applicable only to the hobby lobby, or those programmers who program with tools as a hobby. The vast majority of programmers who program in the various toolsets and languages that they use do so because the businesses that pay their salaries/consulting wages want them to. In other words, most businesses dictate what languages/tools they are willing to invest in. So the goal in ’selling’ development tools - as opposed to ‘giving them away’ - is to sell these tools to businesses, not hobbyists. I do not cease to be amazed at how isolated and ignorant the average developer is from the considerations and contingencies of business planning.
As for "Larry" releasing JDeveloper as a free tool, Larry’s prime revenue stream is database software, not development tools. Larry can afford to give away development tools; Larry’s database licensing and renewal revenue is in the billions. That’s billions with a ‘b’. Approximately 16 ‘b’ in the last year alone. Add this to the fact that the vast majority of major database installations - and the premier, preferred platform for database installations in business - is UNIX; since Linux came on the scene the mass exodus to Linux as the new database platform has made Larry the number one product purveyor of any type on the Linux platform.
>>He can sell Oracle simply because no OS database available on Linux can compete in the high-end
>>segment with it. But in the low-end segment (covered by MySQL and Postgres) they had to relase
>>Oracle Express, and tried to acquire or acquired some OS stuff.
Again, you need to do your research. Larry’s ‘acquisition’ is not about acquisition, it’s about indemnification and support quality. As opposed to the M$ FUD machine, which is getting it’s lunch eaten on the server side by Linux. M$ can’t compete with Linux on a technical, price nor performance level, so it has to resort to the same old, tired template of utilizing it’s primary, core product: FUD generation through it’s massive marketing machine. M$ - following in the footsteps of another business entity that has failed to compete, SCO - is attempting to scare business into not using Linux by hanging the threat of intellectual property lawsuits over their heads. The words ‘Microsoft’ and ‘intellectual property’ used in conjunction is the great oxymoron of our time. Oracle is simply seeking to counter the M$ FUD nonsense for their HUGE Linux customer base.
As for the reason that Larry can sell Oracle on Linux, please note the statement above; UNIX was - and Linux now is - the premier database platform in the world. It’s not about a lack of competition, it’s all about quality/performance/scalability/stability/price. In other words, no non-UNIX platform - other than some of IBM’s mainframe offerings - even comes close to offering the *capability* that UNIX/Linux offers the world as a database platform.
Finally, the release of OracleXE is all about EXPANDING market share into a market that Oracle has never been in; the embedded or small database market segment. Larry is crazy like a fox; he knows very well that if he can get customers to learn to use his database as a free tool they will be much more willing to invest in the larger commercial product. This also provides excellent marketing for Oracle; OracleXE is truly free as in ‘free beer’, unlike MySQL which is actually a commercial product.
>>Oracle is so expensive that it makes Linux an interesting cheap option as the OS for IT departments,
>>although RedHat prices are not very cheap, especially for the AS version. And Oracle is certified to run on
>>RedHat and Suse - try to run it on a Debian - it could be even difficult to install and you get an unsupported installation.
Wow. You need to find a different news source. At one time Oracle was considered expensive in the small database segment. However in early ‘04 Oracle lowered their prices to be competitive with M$ SQL Server at a product - product level. However Oracle is actually LESS expensive than M$ SQL Server because with Linux the OS is free as in ‘free beer’.
Since the GPL forbids selling open source software businesses don’t pay a penny for Linux; they only buy a service contract if they so desire. Businesses can code proprietary software that incorporate elements of open source software, but they can’t sell software built on/based on open source software. GPL is designed to take the sting out of implementing software, instead focusing the cost of using software on the service segment. Many businesses who are savvy have spent the time and money to train up good Linux people; with such IT personnel there is little need to pay for an OS support contract at all.
Having implemented Oracle on RedHat, Mandake/Mandriva, Windows, and Debian I can assure you that it is feasible to run Oracle on any Linux platform that any business would ever want to use. The premier Linux business platforms run Oracle quite capably. Yes it is more challenging to install Oracle on Debian that it is on RedHat. It is also significantly more difficult to install Oracle on RedHat that it is on Mandriva. Since the cost of implementation is exactly -$nothing- in comparison to the cost of maintenance and the ROI for successful apps, this is a ridiculous argument that only Windows weanies use to find some justifcation for not upgrading their skills to more advanced enterprise platforms and technologies.
>>For a development environment like Kylix, it was and it is very different. Most Linux developers are
>>C++ guys, Java guys or PHP/Python/Perl guys. They already have a lot of free tools. They are
>>used to them, and many make a religion of using those tools. They are used to download
>>source code and compile it locally - and often under Linux is the only way to have a working
>>app due to dependencies - sometimes it looks to be far worse than "DLL hell" under Windows
This argument is yet another example of the blind leading the blind. Try to think of Linux as a frontier that is waiting to be explored, staked out and taken ownership of. Linux is the growth market, Windows is the shrinking market. These shallow arguments are the very reason why CodeGear will ultimately be the same business failure that Borland has been in the developer tools segment. Great technology, poor business management combined with no vision whatsoever.
If someone wants to they have free license to implement any sort of architecture they wish to implement on Linux; Oracle is an excellent example of this. When installing Oracle 10g on Mandriva Linux the only dependancy that I had to concern myself with was OpenMotif - and that only for the installation utility; all of the default OS libraries that Oracle relied upon were already in place as a by-product of the default Mandriva installation. In the open source world - as in the Windows world - once can use any approach to core application support that they choose. This is why KDE apps tend to run fine on GNOME desktops and vice-versa; the underpinning of X-Windows and QT is the core of both systems, so if you have both GUI platforms installed apps are interchangeable across both.
>>Do they need a RAD tool, especially one using Pascal as its language? Right now, it looks they
>>don’t. I never see a real demand for Linux desktop applications. The lack of a common widget
>>set and desktop does not help the common user and then demand. Also, Linux is still more
>>difficult to install and to manage especially, and you can’t use any hardware you want.
I have to assume that by ‘they’ you again are talking about hobbyists or open source developers. The business community absolutely needs a RAD tool for Linux; it is the one concept that can catapault the Linux operating system into the realm of a viable desktop applications platform for the business community. It is a money tree that is ripe for the picking; if you had any idea how much businesses who properly manage multi-thousand Windows desktop installations pay for licensing, maintenance and support you would realize how great an opportunity the Linux Desktop will be once good application development support is available for the platform.
As for the ‘lack of a common widget set and desktop’, please check your sources. QT is the primary common platform for GUI development. BusinessObjects, one of the largest consulting/VAR firms in the world, uses QT as a cross-platform development widget set for Windows and several UNIX platforms. Please step outside of the Windows FUD zone and take a look around; you’ll find the world to be a very different place than it has been represented to you as being.
Linux is still more difficult to install and to manager? Hardly. Having implemented networks from vendors such as IBM, Novell, Banyan Vines, Microsoft & Linux I find the facts do not at all support this argument. Again, a Windows weanie false argument designed to assauge the ego of those unwilling to upgrade their skillsets to the more advanced enterprise platforms.
>>Windows grew *before* RAD tools were made available.
>>IMHO, there could have been a market to move some server side applications
>>to cheaper Linux boxes. Borland could have made available a way to Windows
>>developers to move not their client applications, but their server modules to
>>Linux - as long as they could, something Kylix didn’t deliver.
See my statements below; Kylix very much delivered this. Those who have actually used the product can testify to the fact that Kylix as of it’s last revision in ‘02 is STILL to this DAY the best server-side application development platform available to the Linux developer. The amount of effort to produce complex server-side applications in JAVA/C/C++/scripting languages is exponentially greater than it is with Kylix. Check your facts with valid sources please.
>>Kylix should have been an option for Windows developers,
>>trying to sell it to the gcc people it’s just a waste of time.
Agreed. GCC people will never pay the bills. Business people do. Sell to business.
>>Kylix was a failure, for several reasons, noone can deny it. It can’t be
>>resurrected as it was - CodeGear needs a different approach to other platforms.
The only failure on the part of Kylix was a failure of follow-through on the part of Borland. Kylix is to this day used in some of the largest and most complex systems in DoD and commerce. Talk to Simon Kissel, the developer who has kept Kylix functional on current kernels and has about 400 customers who use his Kylix-based products.He even created CrossKylix - and is currently working on CrossFPC(FreePascal) - so that Delphi developers can code in BDS/D7 etc. and output their results to Linux. I’d list a DoD resource if it wasn’t classified; suffice it to say the government has found Kylix to be a great success and is developing the largest system in it’s category in the world using Kylix. Not to mention a comprehensive PostgreSQL GUI database managment product that runs natively on Linux. I have used Kylix to create both native and cross-platform solutions that utilize advance technologies such as n-tier, apache & desktop gui applications. Database, file I/O, socket’s, web etc.
You appear to be accepting the ideas that Borland has put forward as an excuse for not following through with Kylix at face value. Get out of the FUD-zone; talk to the people who have invested serious effort in working with it and you’ll get a very different story. For a product that was less than 2 years old when Borland stopped supporting it, it quickly became one of the top 2 or 3 Linux development products on the market. Had Borland properly marketed the product to business and continued to develop the product - even with a small team of 2 or 3 developers - Kylix would easily be the number one development product on Linux today. And the Linux desktop market would have much greater traction that it currently does.
@Delphi Bigot: face it, Delphi/Kylix has been successful where the *developers* can choose their tools. Where they can’t, Visual Studio, Java and the like rule. It’s not me saying it, it’s numbers. Look at job offers. What the "business community" is looking for? Pascal developers?
They were successful at selling JBuilder because that was what the market wanted. It looked the market didn’t want Kylix. Why didn’t the business community buy lots of Kylix licenses? I guess Kylix would have been alive and kicking now.
Linux is no longer "a frontier". It’s no longer something new. It has shown its strenght and weakness already. Maybe it’s growing, but slowly, and the lack of some common standards and hardware support does not help. It’s not FUD: we *use* Linux and *sell* Linux-based applications, and we know what we talk about.
Just like Oracle is so much more difficult to install, tune and administer than almost any other RDBMS.
Our customers and installations are so large - sorry, I never talk about hobbyists - that we cannot simply install Oracle on any unsupported platform. The related support costs will be very high, even if the OS is free - I am not going to certify and support someone else application on an unsupported Linux, or any other OSes, sorry.
And Oracle *is* expensive. The Standard edition is not more expensive than others RDBMS, but as soon as you need the Enterprise and some options (because of data guard, partitioning, RAC, ecc.) prices get high very quickly.
You’re right, Larry make so much money with it they can develop something for free. CodeGear can’t do it. It has no golden egg chicken - they need to sell anything they code.
Perception is different. "Businesses" are ready to spend $$$$$$$$ on databases, but not so keen on development tools. The former is an "investment", the latter are "a cost". Look at Eclipse. Was it better than JBuilder? I don’t think so. But it is free. It looks "businneses" - not hobbyists - dumped it.
"Kylix as of it’s last revision in ‘02 is STILL to this DAY the best server-side application development platform available to the Linux developer."
C’mon - if a old, buggy, feature-lacking, x86-32 only development tool is the best, that would mean only the average quality is very, very low - and it is not true. There are some good Linux tools - nothing on par with Delphi - but good enough, otherwise how could all those "gcc hobbyists" code such a great OS and applications??? How could Oracle develop its RDBMS without Kylix?
And believe me, I’d really welcome a great RAD development tool on Linux - but Kylix wasn’t. And we *bought* all three releases. We just got a half-backed, hobbyist-level tool sold as an enterprise one.
Anyway, I really wonder anyone but an hobbyist - and especially a government - could start developing anything with an unsupported product - no matter how good it was. I’d fire whoever decided it.
Things are simple. Kylix makes no sense for the native linux crowd - they have gcc. However, Kylix was a motive for windows developers to stick to delphi and port some of their apps to linux. So, while Borland certainly cannot make money from Kylix THEY COULD HAVE GIVEN Windows programmers a reason to stick with Delphi. This reason exists no more. Why should an fresh (windows) programmer choose Delphi? He will be stuck with a single platform and single processor architecture. That makes no sense.
Personally, I see no reason to stick with Borland. Some 8 (eight) years ago I was considering switching to c++. Borland announced Kylix. So I stuck with them. Now I need to port some code to linux. I don’t even care for visual apps, I want a 64 bit object pascal compiler (at the level of delphi 2/3) and a reasonable IDE with debugger. Such do not exists. FPC is a trashy compiler - wierd compatibility issues, bizarre bugs which reapper from one version to another. So, I spend the last few months collecting old versions of Delphi (must be in original box) on ebay. I am only missing Delphi 3. After I complete my archive collection (and finish my current project) I will change to C++.
probably not the slashdot you would have hoped for. The slashdot itself was aimmed negatively and the article it linked to was pretty negative as well.
On the whole, were I to read the article with no Delphi experience, it would do nothing but discourage me from even bothering to look at any of the codegear products.
Fences to mend indeed.
December 28th, 2006 at 9:37 amThats not what slashdotted means Nick.
http://www.urbandictionary.com/define.php?term=slashdotted
As C said, not the most positive article … But I’m sure if CodeGear kills off the QC bugs someone will submit a positive article to slashdot about it
December 28th, 2006 at 3:02 pm1. I would generally agree with C Johnson here.
2. Nick, in the spirit of "extending and innovating on top of Open Source": CodeGear should consider offering "innovations" on top of the FreePascal compiler and possibly even the Lazarus IDE. In addition, it seems to me that many of the non-visual Delphi components and classes could also be used to build outstanding extentions to the dynamic languages you are targeting, similiar (what about ECO?) to what this guy is doing => http://members.chello.be/ws36637/php4delphi.html Supporting the FreePascal compiler would make that a cross-platform reality. If your going "Active" languages, you’ve got to be cross-platform or you will be the laughing stock of the development world.
December 28th, 2006 at 3:19 pmI have to admit the stance of embracing open source efforts while simulaneously quashing the one project open source enthusiasts most want is something of a confused stance.
That said, I do COMPLETELY understand how a business wants to see money for its efforts, and I understand why they like free work done by others.
Clearly, however, Kylix shows us how shortsighted and doomed for failure that mix can be. Makes me wonder how the same company can say that Kylix is a financial failure and then try to embrace PHP - that thought pattern is just too Borland for my liking.
December 28th, 2006 at 3:30 pmI agree with C Jhonson.
Linux native compiler support must have…
Novell put 800 engineer to support Suse Linux 10. Suse Linux 10 is simple, cheaper, powerfull and very attracktive for corporates.
Even MS considers Linux ( http://www.novell.com/linux/microsoft/allsystemsgo.html ), how about you? Come on Nick! I totaly believe that CG want to make correct product. But if you continue in this way, developers will begin to think CodeGear’s thought is "Where Money Matters!"..
December 28th, 2006 at 7:48 pmWhy does everybody always keep screaming for cross platform support? The strength of Delphi _is_ GUI Database apps (at least imho
). These apps are never (ok, never say never 
) going to be big on nix platforms. I have seen only very few big companies where nix is used as a desktop environment. So why invest a hell of a lot of money in something you know will probably never hit it off? I think killing Kylix (come on, is this really news?) is a good decision.
CodeGear should take Delphi further in the direction of things it’s good at. That goes for anyone by the way: do what you’re good at, don’t do something you know you have a big disadvantage in ever getting good at.
Then there is one more thing: in the replies to the /. article, there is someone replying: http://developers.slashdot.org/comments.pl?sid=213920&cid=17394354
December 28th, 2006 at 9:21 pmNick, is the sacking of 120 people true? could you clear that up for us? those tiny comments are what makes me worry and it would be good to have someone trustworthy comment on that, not some random uncheckable comment from an anonymous guy…
Crossplatform support is important for those of us who write server apps.
Delphi/Objectpascal has just too many advantages over other langugages (fast compile times, easy to read, no runtime) to let that opportunity pass by.
I’m not saying we need another Kylix, but some kind of "native" support for Linux would be great. For example, it would be great if we could write Freepascal Linux server apps with BDS. Isn’t it possible to add a Freepascal "personality"?
December 28th, 2006 at 10:38 pm@Fritz Huber:
) up to a whole lot more devvers out there. Just depends on how hard it is to get it done, and how high it is on their wish-list 
December 29th, 2006 at 12:25 amI think adding a FreePascal personality would be a phenomenal move, exactly in line with the "embracing the open source community" thing. Also it would be embracing without potential loss of customers, since it would open BDS (or is it CDS?
Forget Linux.
December 29th, 2006 at 12:57 amWhat I would really like to see.. is a Delphi and/or CBuilder for the Apple systems.
I find Ben Smith’s comments concerning his reasoning for killing Kylix particularly troubling.
Historically when a software company is seeking a foothold in the marketplace it performs SERIOUS market research to find the growth market(s). Then it positions itself to ‘ride the wave’ of the growth market(s).
While M$’s growth is slowing, Linux growth continues to push the envelope. To ignore the enormous markets that Linux has developed and is continuing to develop because of revenue figures from - what, more than 4 years ago? - is idiotic.
No, this isn’t good business sense. It’s the fear that is born out of ignorance concerning a particular market segment. It’s a reflection of a lack of market research performed by Mr. Smith and company.
It also reflects the same executive decision making pattern perfected by Borland’s brilliant executive management team - Windows is the ‘hype machine’ of the day, so let’s just hang off of their coattails, eating the ‘crumb’s from the master’s table’.
Please Ben, call up Larry on the phone and tell him what a bad investment risk Linux is for the database/software development market. I fear you may get an earful, once the laughter and derision decides.
Perhaps you’d like to explain to RedHat or Novell why they shouldn’t be successful at what they do either.
It is unfortunate that the only thing that holds back Linux from grabbing major desktop market share is a lack of a good dev tools platform, and the company most optimally positioned for filling that void has descended into the depths of cowardice and fear.
I believe that we will see the emergence of an excellent - or at least very good - dev tools platform for Linux. Unfortunately it probably won’t be coming from Scotts Valley.
Once that occurs, Ben, what will you excuse be then??
December 29th, 2006 at 5:26 amIMHO it’s an error to look at Kylix only for Linux applications, you will not find the revenue you expect from there. But you can think on Kylix as a nix/OSX dev platform, now you have a much larger market, OSX market will spend more than the *nix one.
December 29th, 2006 at 6:37 amSpeaking of Corba, I’ve stumbled over http://www.zeroc.com/ice.html
December 29th, 2006 at 7:32 am, which looks pretty great. But guess what. It’s available for C++, Java, Ruby, PHP, Python, C# and Visual Basic. No Pascal to be seen. Delphi needs to get out of its ever shrinking Windows niche, otherwise things will not be looking to good in the near future, IMHO.
@Delphi Bigot: face it, *selling* developmet tools to Linux developers is an hard task. Even "Larry" released JDeveloper as a free tool.
December 30th, 2006 at 9:52 amHe can sell Oracle simply because no OS database available on Linux can compete in the high-end segment with it. But in the low-end segment (covered by MySQL and Postgres) they had to relase Oracle Express, and tried to acquire or acquired some OS stuff.
Oracle is so expensive that it makes Linux an interesting cheap option as the OS for IT departments, although RedHat prices are not very cheap, especially for the AS version. And Oracle is certified to run on RedHat and Suse - try to run it on a Debian - it could be even difficult to install and you get an unsupported installation.
For a development environment like Kylix, it was and it is very different. Most Linux developers are C++ guys, Java guys or PHP/Python/Perl guys. They already have a lot of free tools. They are used to them, and many make a religion of using those tools. They are used to download source code and compile it locally - and often under Linux is the only way to have a working app due to dependencies - sometimes it looks to be far worse than "DLL hell" under Windows
Do they need a RAD tool, especially one using Pascal as its language? Right now, it looks they don’t. I never see a real demand for Linux desktop applications. The lack of a common widget set and desktop does not help the common user and then demand. Also, Linux is still more difficult to install and to manage especially, and you can’t use any hardware you want. Windows grew *before* RAD tools were made available.
IMHO, there could have been a market to move some server side applications to cheaper Linux boxes. Borland could have made available a way to Windows developers to move not their client applications, but their server modules to Linux - as long as they could, something Kylix didn’t deliver.
Kylix should have been an option for Windows developers, trying to sell it to the gcc people it’s just a waste of time.
Kylix was a failure, for several reasons, noone can deny it. It can’t be resurrected as it was - CodeGear needs a different approach to other platforms.
@Kent Morwath -
Let me reply to your comments below, by stating that in order to reply I have to make assumptions about your experience with Linux/Kylix/Oracle since I don’t know what your experience level actually is. However I can assure you that I am sufficiently authoritative in a technical and business sense in all 3 of these areas, as well as in other areas relevant to this discussion.
>> face it, *selling* developmet tools to Linux developers is an hard task.
>> Even "Larry" released JDeveloper as a free tool.
I would have to agree with this statement on it’s face. However this statement leads me to conclude that you either don’t understand business, don’t understand the application of Linux technologies to business, or both. This would seem to be consistent with the larger Delphi programming community. However this most definitely does *not* apply to the whole of the Delphi community. Some of the most capable people in ‘our’ community are quite adept at cross-platform development.
The goal of being successful at selling development tools to developers is applicable only to the hobby lobby, or those programmers who program with tools as a hobby. The vast majority of programmers who program in the various toolsets and languages that they use do so because the businesses that pay their salaries/consulting wages want them to. In other words, most businesses dictate what languages/tools they are willing to invest in. So the goal in ’selling’ development tools - as opposed to ‘giving them away’ - is to sell these tools to businesses, not hobbyists. I do not cease to be amazed at how isolated and ignorant the average developer is from the considerations and contingencies of business planning.
As for "Larry" releasing JDeveloper as a free tool, Larry’s prime revenue stream is database software, not development tools. Larry can afford to give away development tools; Larry’s database licensing and renewal revenue is in the billions. That’s billions with a ‘b’. Approximately 16 ‘b’ in the last year alone. Add this to the fact that the vast majority of major database installations - and the premier, preferred platform for database installations in business - is UNIX; since Linux came on the scene the mass exodus to Linux as the new database platform has made Larry the number one product purveyor of any type on the Linux platform.
>>He can sell Oracle simply because no OS database available on Linux can compete in the high-end
>>segment with it. But in the low-end segment (covered by MySQL and Postgres) they had to relase
>>Oracle Express, and tried to acquire or acquired some OS stuff.
Again, you need to do your research. Larry’s ‘acquisition’ is not about acquisition, it’s about indemnification and support quality. As opposed to the M$ FUD machine, which is getting it’s lunch eaten on the server side by Linux. M$ can’t compete with Linux on a technical, price nor performance level, so it has to resort to the same old, tired template of utilizing it’s primary, core product: FUD generation through it’s massive marketing machine. M$ - following in the footsteps of another business entity that has failed to compete, SCO - is attempting to scare business into not using Linux by hanging the threat of intellectual property lawsuits over their heads. The words ‘Microsoft’ and ‘intellectual property’ used in conjunction is the great oxymoron of our time. Oracle is simply seeking to counter the M$ FUD nonsense for their HUGE Linux customer base.
As for the reason that Larry can sell Oracle on Linux, please note the statement above; UNIX was - and Linux now is - the premier database platform in the world. It’s not about a lack of competition, it’s all about quality/performance/scalability/stability/price. In other words, no non-UNIX platform - other than some of IBM’s mainframe offerings - even comes close to offering the *capability* that UNIX/Linux offers the world as a database platform.
Finally, the release of OracleXE is all about EXPANDING market share into a market that Oracle has never been in; the embedded or small database market segment. Larry is crazy like a fox; he knows very well that if he can get customers to learn to use his database as a free tool they will be much more willing to invest in the larger commercial product. This also provides excellent marketing for Oracle; OracleXE is truly free as in ‘free beer’, unlike MySQL which is actually a commercial product.
>>Oracle is so expensive that it makes Linux an interesting cheap option as the OS for IT departments,
>>although RedHat prices are not very cheap, especially for the AS version. And Oracle is certified to run on
>>RedHat and Suse - try to run it on a Debian - it could be even difficult to install and you get an unsupported installation.
Wow. You need to find a different news source. At one time Oracle was considered expensive in the small database segment. However in early ‘04 Oracle lowered their prices to be competitive with M$ SQL Server at a product - product level. However Oracle is actually LESS expensive than M$ SQL Server because with Linux the OS is free as in ‘free beer’.
Since the GPL forbids selling open source software businesses don’t pay a penny for Linux; they only buy a service contract if they so desire. Businesses can code proprietary software that incorporate elements of open source software, but they can’t sell software built on/based on open source software. GPL is designed to take the sting out of implementing software, instead focusing the cost of using software on the service segment. Many businesses who are savvy have spent the time and money to train up good Linux people; with such IT personnel there is little need to pay for an OS support contract at all.
Having implemented Oracle on RedHat, Mandake/Mandriva, Windows, and Debian I can assure you that it is feasible to run Oracle on any Linux platform that any business would ever want to use. The premier Linux business platforms run Oracle quite capably. Yes it is more challenging to install Oracle on Debian that it is on RedHat. It is also significantly more difficult to install Oracle on RedHat that it is on Mandriva. Since the cost of implementation is exactly -$nothing- in comparison to the cost of maintenance and the ROI for successful apps, this is a ridiculous argument that only Windows weanies use to find some justifcation for not upgrading their skills to more advanced enterprise platforms and technologies.
>>For a development environment like Kylix, it was and it is very different. Most Linux developers are
>>C++ guys, Java guys or PHP/Python/Perl guys. They already have a lot of free tools. They are
>>used to them, and many make a religion of using those tools. They are used to download
>>source code and compile it locally - and often under Linux is the only way to have a working
>>app due to dependencies - sometimes it looks to be far worse than "DLL hell" under Windows
This argument is yet another example of the blind leading the blind. Try to think of Linux as a frontier that is waiting to be explored, staked out and taken ownership of. Linux is the growth market, Windows is the shrinking market. These shallow arguments are the very reason why CodeGear will ultimately be the same business failure that Borland has been in the developer tools segment. Great technology, poor business management combined with no vision whatsoever.
If someone wants to they have free license to implement any sort of architecture they wish to implement on Linux; Oracle is an excellent example of this. When installing Oracle 10g on Mandriva Linux the only dependancy that I had to concern myself with was OpenMotif - and that only for the installation utility; all of the default OS libraries that Oracle relied upon were already in place as a by-product of the default Mandriva installation. In the open source world - as in the Windows world - once can use any approach to core application support that they choose. This is why KDE apps tend to run fine on GNOME desktops and vice-versa; the underpinning of X-Windows and QT is the core of both systems, so if you have both GUI platforms installed apps are interchangeable across both.
>>Do they need a RAD tool, especially one using Pascal as its language? Right now, it looks they
>>don’t. I never see a real demand for Linux desktop applications. The lack of a common widget
>>set and desktop does not help the common user and then demand. Also, Linux is still more
>>difficult to install and to manage especially, and you can’t use any hardware you want.
I have to assume that by ‘they’ you again are talking about hobbyists or open source developers. The business community absolutely needs a RAD tool for Linux; it is the one concept that can catapault the Linux operating system into the realm of a viable desktop applications platform for the business community. It is a money tree that is ripe for the picking; if you had any idea how much businesses who properly manage multi-thousand Windows desktop installations pay for licensing, maintenance and support you would realize how great an opportunity the Linux Desktop will be once good application development support is available for the platform.
As for the ‘lack of a common widget set and desktop’, please check your sources. QT is the primary common platform for GUI development. BusinessObjects, one of the largest consulting/VAR firms in the world, uses QT as a cross-platform development widget set for Windows and several UNIX platforms. Please step outside of the Windows FUD zone and take a look around; you’ll find the world to be a very different place than it has been represented to you as being.
Linux is still more difficult to install and to manager? Hardly. Having implemented networks from vendors such as IBM, Novell, Banyan Vines, Microsoft & Linux I find the facts do not at all support this argument. Again, a Windows weanie false argument designed to assauge the ego of those unwilling to upgrade their skillsets to the more advanced enterprise platforms.
>>Windows grew *before* RAD tools were made available.
>>IMHO, there could have been a market to move some server side applications
>>to cheaper Linux boxes. Borland could have made available a way to Windows
>>developers to move not their client applications, but their server modules to
>>Linux - as long as they could, something Kylix didn’t deliver.
See my statements below; Kylix very much delivered this. Those who have actually used the product can testify to the fact that Kylix as of it’s last revision in ‘02 is STILL to this DAY the best server-side application development platform available to the Linux developer. The amount of effort to produce complex server-side applications in JAVA/C/C++/scripting languages is exponentially greater than it is with Kylix. Check your facts with valid sources please.
>>Kylix should have been an option for Windows developers,
>>trying to sell it to the gcc people it’s just a waste of time.
Agreed. GCC people will never pay the bills. Business people do. Sell to business.
>>Kylix was a failure, for several reasons, noone can deny it. It can’t be
>>resurrected as it was - CodeGear needs a different approach to other platforms.
The only failure on the part of Kylix was a failure of follow-through on the part of Borland. Kylix is to this day used in some of the largest and most complex systems in DoD and commerce. Talk to Simon Kissel, the developer who has kept Kylix functional on current kernels and has about 400 customers who use his Kylix-based products.He even created CrossKylix - and is currently working on CrossFPC(FreePascal) - so that Delphi developers can code in BDS/D7 etc. and output their results to Linux. I’d list a DoD resource if it wasn’t classified; suffice it to say the government has found Kylix to be a great success and is developing the largest system in it’s category in the world using Kylix. Not to mention a comprehensive PostgreSQL GUI database managment product that runs natively on Linux. I have used Kylix to create both native and cross-platform solutions that utilize advance technologies such as n-tier, apache & desktop gui applications. Database, file I/O, socket’s, web etc.
You appear to be accepting the ideas that Borland has put forward as an excuse for not following through with Kylix at face value. Get out of the FUD-zone; talk to the people who have invested serious effort in working with it and you’ll get a very different story. For a product that was less than 2 years old when Borland stopped supporting it, it quickly became one of the top 2 or 3 Linux development products on the market. Had Borland properly marketed the product to business and continued to develop the product - even with a small team of 2 or 3 developers - Kylix would easily be the number one development product on Linux today. And the Linux desktop market would have much greater traction that it currently does.
January 1st, 2007 at 8:30 amI completely agree with DelphiBigot. And thanks for the up to date software market.
I did not start a linux project yet. I plan seriously expand our bussines software support to Linux.
And a news says: Half of businesses will switch to Linux : http://www.theinquirer.net/default.aspx?article=36678
I hope CodeGear do not miss the target.
January 3rd, 2007 at 10:24 amhops.. And thanks to DelphiBigot for the up to date software market information..
January 3rd, 2007 at 10:29 am@Delphi Bigot: face it, Delphi/Kylix has been successful where the *developers* can choose their tools. Where they can’t, Visual Studio, Java and the like rule. It’s not me saying it, it’s numbers. Look at job offers. What the "business community" is looking for? Pascal developers?
They were successful at selling JBuilder because that was what the market wanted. It looked the market didn’t want Kylix. Why didn’t the business community buy lots of Kylix licenses? I guess Kylix would have been alive and kicking now.
Linux is no longer "a frontier". It’s no longer something new. It has shown its strenght and weakness already. Maybe it’s growing, but slowly, and the lack of some common standards and hardware support does not help. It’s not FUD: we *use* Linux and *sell* Linux-based applications, and we know what we talk about.
Just like Oracle is so much more difficult to install, tune and administer than almost any other RDBMS.
Our customers and installations are so large - sorry, I never talk about hobbyists - that we cannot simply install Oracle on any unsupported platform. The related support costs will be very high, even if the OS is free - I am not going to certify and support someone else application on an unsupported Linux, or any other OSes, sorry.
And Oracle *is* expensive. The Standard edition is not more expensive than others RDBMS, but as soon as you need the Enterprise and some options (because of data guard, partitioning, RAC, ecc.) prices get high very quickly.
You’re right, Larry make so much money with it they can develop something for free. CodeGear can’t do it. It has no golden egg chicken - they need to sell anything they code.
Perception is different. "Businesses" are ready to spend $$$$$$$$ on databases, but not so keen on development tools. The former is an "investment", the latter are "a cost". Look at Eclipse. Was it better than JBuilder? I don’t think so. But it is free. It looks "businneses" - not hobbyists - dumped it.
"Kylix as of it’s last revision in ‘02 is STILL to this DAY the best server-side application development platform available to the Linux developer."
C’mon - if a old, buggy, feature-lacking, x86-32 only development tool is the best, that would mean only the average quality is very, very low - and it is not true. There are some good Linux tools - nothing on par with Delphi - but good enough, otherwise how could all those "gcc hobbyists" code such a great OS and applications??? How could Oracle develop its RDBMS without Kylix?
And believe me, I’d really welcome a great RAD development tool on Linux - but Kylix wasn’t. And we *bought* all three releases. We just got a half-backed, hobbyist-level tool sold as an enterprise one.
Anyway, I really wonder anyone but an hobbyist - and especially a government - could start developing anything with an unsupported product - no matter how good it was. I’d fire whoever decided it.
January 3rd, 2007 at 3:42 pmThings are simple. Kylix makes no sense for the native linux crowd - they have gcc. However, Kylix was a motive for windows developers to stick to delphi and port some of their apps to linux. So, while Borland certainly cannot make money from Kylix THEY COULD HAVE GIVEN Windows programmers a reason to stick with Delphi. This reason exists no more. Why should an fresh (windows) programmer choose Delphi? He will be stuck with a single platform and single processor architecture. That makes no sense.
Personally, I see no reason to stick with Borland. Some 8 (eight) years ago I was considering switching to c++. Borland announced Kylix. So I stuck with them. Now I need to port some code to linux. I don’t even care for visual apps, I want a 64 bit object pascal compiler (at the level of delphi 2/3) and a reasonable IDE with debugger. Such do not exists. FPC is a trashy compiler - wierd compatibility issues, bizarre bugs which reapper from one version to another. So, I spend the last few months collecting old versions of Delphi (must be in original box) on ebay. I am only missing Delphi 3. After I complete my archive collection (and finish my current project) I will change to C++.
March 20th, 2007 at 10:23 pm