Delphi - does catastrophe loom ? (long)

g'day everyone,

I've been buying and using Delphi since Turbo Pascal 5, and think that it is the best general-purpose application programming language bar none.

However, I am beginning to think that Borland has made a major mistake going down the .NET path, and that Microsoft will be the beneficiary.

The reasons for this are as follows.

I believe that .NET has been grossly over-hyped, in typical Microsoft fashion. While the original intent seemed for it to be a Java replacement, it is now pushed as the general programming language for future versions of Windows. However, recent posts on its performance have confirmed its Java-like performance (memory x 5, speed x 0.02-0.2), which makes it unsuitable for general desktop applications. On the other hand, how good a competitor will it be for Java given the amount of baggage needed to provide a desktop API ? It reminds me of a classic warplane of the 1960's, the F-111: a fighter-bomber that was not as fast and manouverable as a fighter, without the range and payload of a bomber.

The standard response to this is "it will be fixed in the next version", or "next year's faster CPU will fix any performance problems". Whether or not it will be fixed could take up an entire newsgroup, but next years CPUs will not be faster - they will be _SLOWER_, and their speed will increase very slowly in the next few years, if at all.

The CPU paradigm has switched from "more megahertz" to parallelism and 64-bits. Anyone who doubts this just have a read of March's Dr Dobbs:

http://www.ddj.com/articles/2005/0503/

Unfortunately .NET (especially VCL.NET) is a 32-bit technology that was not written to exploit parallelism to the fullest.

So .NET applications that are built with Delphi 200x will perform poorly compared to their 32-bit equivalents.

However, 64-bit applications can show speed improvements of up to 50% or so, particularly on AMD hardware, over their 32-bit versions, and only require about 50% more memory (not 500%).

Guess who has released a 64-bit compiler ? Our good friends at Microsoft - who have been ramming .NET down everyone else's throat.

It's a bit like the old OS/2 debacle: Microsoft conned lots of people into investing in one technology (OS/2), then brought out a different one (Win95 + NT).

It has been said that Borland invested several million dollars in research into 64-bit technology. however, when a 64-bit compiler was first mentioned at Borcon 2002 in Australia, it was targetted solely at Itanium (or "Itanic" as the Register dubbed it). There was a question from the audience about the recently announced AMD 64-bit project, but the offical response was that it was not on the agenda.

Well, it still seems to be off Borland's agenda. What Borland Marketing does not seem to comprehend is that while the market for an IA-64 compiler was probably several tens or even hundred of units, the market for an AMD64 (alias EMT64) compiler is in the range 10^4-10^5 units. What they also do not seem to comprehend is that an AMD64 compiler is a much simpler beast than an IA64 compiler, and hence much cheaper to "produce".

Prediction:
***********
If Borland sticks to just .NET its market share will plummet because .NET applications have severe performance problems that future hardware developments will not solve. Delphi will become a niche product then die out. Microsoft will clean up the compiler market with a 64-bit product that also does .NET if you really insist on it.

Gets down on knees and begs:
****************************
That Borland's Marketing people pull their finger's out: the market is _HUGE_. Ten years ago Borland was there for the 16- to 32-bit revolution. Please be there for the 32- to 64-bit one. It's here. Now.

At first we don't need a 64-bit Delphi IDE - we just need a compiler that can produce 64-bit images (just like D7 could produce .NET assemblies with the preview CIL compiler).

Let your technical people show that they are the best !
.

Relevant Pages

  • Re: Intel announces 64-bit extensions
    ... "When market demand for blue Martian tools is large enough to sustain ... Borland financial objectives, that is when Borland should deliver blue ... > That's all we, Win64 weenies, are asking for: a stated future for Delphi ... > native compilation, so we can postpone, with some backing, tool migration ...
    (borland.public.delphi.non-technical)
  • Re: Cat among the pigeons......
    ... You know how attack if know what need/expect the market. ... I'm one of this from years and now strat a small shop... ... This is something that not MS not Borland not appreciate enough, ... dream a deal of delphi + DevExpress or TMS on this. ...
    (borland.public.delphi.non-technical)
  • Re: Death of Kylix and migrating towards Java
    ... But it can't beat Delphi's compiler speed! ... app developers on non-Windows platforms. ... It is NOT Delphi. ... > I do agree that Kylix was a very good promise and concept, that Borland ...
    (borland.public.delphi.non-technical)
  • Re: Delphi job market - observation
    ... Ah, the fallacy of the unused market, a very common economic fallacy. ... development tools would always have an advantage over Borland's because MSFT ... than Delphi, but because you can find C# programmers cheaply who will work ... I bet this is a field that Borland thought it had covered with Java and C++. ...
    (borland.public.delphi.non-technical)
  • Re: Borlands linux strategy ?
    ... What would be so awful for Borland, as a company, from a strategic perspective, to start a new product line that targets non-asp.net, non-windows development? ... My point is that if you try to port Delphi, and what makes Delphi Delphi, to another development platform, it will fail. ... This is the main point of my argument for why Delphi is and should remain a Windows development tool. ... It's hard for a software vendor to make any money from dev tools alone, especially in the ISV market. ...
    (borland.public.delphi.non-technical)