Board index » delphi » usual nontechnical issues

2003-11-25 06:59:50 PM
delphi5
Hello.
Although I am not usually initiating "future of Delphi" threads, I have need to share some thoughts and hear oppinions.
1. Kylix is dead. You can still write applications with it, and there will be poeople willing to share experience, but if current bugs are your problem, you are stuck.
2. There is no way to justify Kylix usage with enterprise applications.
3. VCL is about to be dead. There will be port to .net, so people can do transition, but in long term, it will be droped.
Borland said that, and I agree, there is no need to maintain VCL in future, because .net framework is there.
4. VCL can be used for few years for win32 applications, or to maintain existing VCL apps.
5. There is no way to justify VCL usage with enterprise applications.
6. Delphi.NET is actually Delphi IDE, with Delphi language on top of VLC port to .net, and .net framework. When Borland drops VCL, it will be IDE and language that complies to MSIL on top of net.
7. C#Builder is safer to go with than Delphi, because you learn C#, and therefore are in position to get VS.net job also. And because of beforementioned Delphi situation, its better not to do VCL.net anyway, because app would need to be rewritten anyway.
8. CBX is promising, but for now it is for tools and components makers.
9. Borland is gradually droping compilers, focusing on IDEs and tools.
10. Java is best choice for enterprise app now, especially if you need multiplatform.
11. C like languages are winners over more human languages. :))Awful syntax is cushioned with class names, method calls, naming conventions, so its easier to tolerate less recognizible syntax enveloping it.
11.1 Pascal is about to be dead, without Delphi.
12. What people like about Delphi is IDE and components.
13. Why not drop Delphi now, and focus on C#Builder, JBuilder, and CBX. I can see same IDE backed up with UML designer. Bundle C#, Java and C++ under same IDE and tools envelope, and drop compilers. Sun is doing Java, Ms is doing net, and Intel, Microsoft and open source community are doing C++.
14. Do substitute or addition for SQL. I really dislike bussines rules spread over client, middle tier and server. Perhaps pascal
could do the job, its more alike to SQL than C.
Thanks
 

Hi,
doubt that it will die, as it has, in my experience, been one of those things that made selling Delphi copies easier.
you.
MS has been pushing VB(Visual Basic) for enterprise apps for 10 years now... and you say Kylix should be worse? ;)
be cross platform compatible (What??? I hear someone scream.... Mono is supposed to do that for dotNet developers???
Well, firstly mono v. 1.0 is not going to be ready until earliest a year from now according to Novell who is now in
charge of Mono, and secondly MS has publically stated that Mono will never be fully dotNet compatible. Where the lines
are going to go, I dont know. Big parts of dotNet is actually protected by MS under patents etc. and they are now
showing that they will try to protect that)
I kind of anticipated this long time ago. I think the future of Mono vs MS dotNet is going to be seriously unsure now
that Novell is trying to gain foothold in the OS business (Linux) and as always is a competitor to MS in the
file/print/colaboration area.
will even under Longhorn. Think that its safe to say that if MS dont make Longhorn backwards compatible, Longhorn will
die.
Most EA's dont survive for more than 4-5 years anyway on Windows platforms before they are rewritten or s{*word*99}ped.
Im sure that VCL apps have a longer timespan than VB(Visual Basic) apps.
I dont know Borlands plans, but I dont think they will put all their eggs in one basket.
I still think Borland has a plan of cross platform portability in which case we will see support of other compilers in
addition to MSIL
rewritten anyway.
Today C#'ers can get jobs pretty easily... 3 years ago it was Java folks that boomed. In 3 years time it will be
something else again.
Delphi jobs have always been less and harder to find. Its been like this since D1. It may actually change to the better
now with Borland trying to position itself in the dotNet world as the first and only full alternative to VS.Net.
But ofcourse most managers are choosing MS only products which from some viewpoints are a good decision, but from others
a bad one (being dependant of not only a development tool, but also the OS it runs on).
And its been stated several times by Borland that Delphi for Win32 most likely will have compiler improvements.
But is it _REALLY_ the best choice? In my opinion no, but its one that has been getting lots of hype.
And hype and name often controls what managers choose to use.
Actually I know of several _large_ enterprise Java apps with lots of very talented people involved which have had their
fair share of problems due to lousy performance, crazy requirements to the hardware on which the app. server is running
etc.
The biggest selling point of Java imo is the cross platform capabilities.
However its my experience now that even if it was one of the key arguments used by managers some years ago to choose
Java, the same managers are now choosing dotNet although its far from cross platform. Which just means that many
managers
dont even have a clue of what they in reality want, but merely base their purchases on the amount of hyped around the
products.
C has been a {*word*109} language since 1968 or so. So this is not really anything new.
What is interesting is that even though C has been {*word*109}, lots of alternative languages have been invented.
Some more popular than others but never the less there is a market for non C type languages too.
wasnt going to support it. But thats exactly what they are doing by now _additionally_ adding Pascal to the dotNet
platform. So Delphi isnt dead, and Pascal still lives on.
net, and Intel, Microsoft and open source community are doing C++.
Because there are atleast a couple of million Delphi developers. Why should Borland not earn money on them? Borland
almost have a monopoly on the Pascal/Delphi niche. Seems to be a stupid to leave one lucrative niche to join a place
where the posibility of competing is even much harder.
placed on a middletier and nowhere else. But that discussion we can continue on the kbmMW newsgroups at
--
best regards
Kim Madsen
XXXX@XXXXX.COM
www.components4developers.com
The best components for the best developers
kbmMW - RAD n-tier application server framework
kbmMemTable - High performance memory table
kbmWABD - RAD web development
kbmX10 - RAD house automation
 

I'm not going to reply to your unqualified assertions, but there are
some errors of fact I will correct:
Buch writes:
download the preview for several months now, and the full product was
just announced. And frankly if you don't want to use a .NET visual
framework which is going to be deprecated, then stay away from WinForms
-- its successor was announced one month ago (Avalon).
C++ open letter, but 1) this isn't what they said, and 2) C++ is
handled by a different business unit at Borland with different goals
than the business unit which produces Delphi.
-Craig
--
Craig Stuntz [TeamB] . Vertex Systems Corp. . Columbus, OH
Delphi/InterBase Weblog : delphi.weblogs.com
InterBase Performance Monitor -- Analyze and control your IB7
server: delphi.weblogs.com/IBPerformanceMonitor
 

"C4D - Kim Madsen" <XXXX@XXXXX.COM
(kbmMW/kbmMemTable/kbmWABD/kbmX10)>writes
I don't know how many companies are going to base their future on a product
that is not going to have patches for 2004 and is known to have some
problems *now*. Even if it is not officially dead, I personally believe it
would be pretty silly for people to commit to it given the official
statements of Borland. So, in the end, it is most likely gonna die than be
"cured".
Regardless of the definition of "Enterprise application", one thing common
to most companies developing "Enterprise" software is they want a
committment from software vendors about future upgrades, supoprt, etc.
So take that as a "it's VERY hard to justify Kylix usage"
Where was that "publically stated"?
Anyhow, it is irrelevant. If your current plans are for cross-platform
development, Java is the way.
now should use it or will <G>
Technology evolves, and so do consumers PCs.
If backwards compatibility was such a big problem, then you've automatically
said Linux is never going to go anywhere, regardless of server or desktop.
That, obviously, didn't happen, regardless of it is no compatibility with
existing DOS or Windows apps.
Borland in regards to Delphi should be devoted to making better IDEs and
frameworks, not writing 12 compilers. History should teach lessons.
The question is are those people looking for alternatives?
now.
"Gradually dropping" is the key. What do you think will be next other than
better IDE and tools (which, again, it is what they should do IMO)?
[..]
[..]
Parts of business logic should be automatically uploaded from middle-tier to
the clients, maximixing use of both tiers and still maintaining centralized
control.
--
Best regards,
Alessandro Federici
RemObjects Software, Inc.
www.remobjects.com
 

Craig Stuntz [TeamB] writes:
source from which you could more or less draw that conclusion, namely
Tim Anderson's interview of Dale Fuller and Blake Stone at
www.itwriting.com/borlandtop2.php
where Blake Stone has this to say about the future of VCL:
VCL to the same league with the old object model and such. I suppose
it will work for years to come (we have as much time as we need) but
one shouldn't rely on it.
If I am wrong here I will be more than happy to be corrected.
--
ilkka hyvärinen
 

Hi,
Apart from that, I agree in your logic, but the same logic applies to VB(Visual Basic) etc. which are still used for 'Enterprise'
apps.
www.pcworld.dk/default.asp
'Men en fuldstændig kompatibilitet kan der ikke blive tale om. Det fastslog Microsoft overfor PC World p?den nyligt
afholdte konference IT Forum. De øvre lag i .Net-arkitekturen er ikke åbne standarder, men ejes af Microsoft og kan ikke
kopieres.'
Freely translated:
'But a complete compatibility will not happen. Microsoft established that fact to PC World on the newly helt IT Forum
conference. The top layers in the .Net architecture are not open standards, but are owned by Microsoft and cant be
copied.'
If you need something new, then you use something new.
Lots of tools for building houses has been invented the last 50 years, but the good old hammer still do the job nicely
specially in the hands of an experienced user.
But one of the reasons why it has still not taken 100% off is exactly that the good level hasnt been good enough.
Hence users are reluctant to use it because of the existing codebase.
niche areas that may be more lucrative and easier to control.
of a bunch of 3rdparty compilers. That wont make crossplatform very well, as each compiler has its own incompabitlities.
Currently there is one more than there have been before... so I'd rather call it to extend its compiler base.
developer is totally convinced that is fine and wont ever change.
--
best regards
Kim Madsen
XXXX@XXXXX.COM
www.components4developers.com
The best components for the best developers
kbmMW - RAD n-tier application server framework
kbmMemTable - High performance memory table
kbmWABD - RAD web development
kbmX10 - RAD house automation
 

Ilkka Hyvärinen writes:
VCL in about the same timeframe as Borland dropped the old object model
(object was deprecated in Delphi 1 in 1995 and the first version of
Delphi which won't support it is Delphi for .NET in 2003) then I don't
have a problem with that interpretation. They may drop it sooner if
Microsoft completely deprecates the Win32 API in less than 8 years.
But (and I have the same issue with Blake's comments) it is IMHO
disingenuous to suggest using WinForms instead of the VCL for .NET on
the basis that the VCL for .NET will be deprecated, since the same is
true of WinForms.
-Craig
--
Craig Stuntz [TeamB] ?Vertex Systems Corp. ?Columbus, OH
Delphi/InterBase Weblog : delphi.weblogs.com
How to ask questions the smart way:
www.catb.org/~esr/faqs/smart-questions.html
 

"C4D - Kim Madsen" <XXXX@XXXXX.COM
(kbmMW/kbmMemTable/kbmWABD/kbmX10)>writes
[..]
[..]
what it looks.
Where's the "official statement"? (this doesn't mean I think there's any
interest from MS in making compatibility complete).
better tools (i.e. VB6 users <G>).
[..]
????
[..]
before we came along and MS provides you with all you need to develop nTier
apps without spending an extra penny than Delphi Standard doesn't seem to
hurt any of us, does it? <G>Markets that are only in the hands of big
players now might change when smaller players offer more cost-effective
solutions which might also provide better value (this is a general
statement). Or, to loko at the Linux example, even when quality is not so
good but good-enough.
My real-life experience tells me the exact opposite.
[..]
I didn't mean it is a good alternative to Java or has a sure future. I only
said it is the only other realistic and proved (C++) alternative.
[..]
I missread it. It sill has to be seen if there will be more. IMO it will
either stabilize or shrink after DFDN.
that, 3rd parties solve most issues ;-)
That's another domain of "rules".
 

"Craig Stuntz [TeamB]" <XXXX@XXXXX.COM [a.k.a. vertexsoftware.com]>
writes news:XXXX@XXXXX.COM...
Maybe he's a VB(Visual Basic) user trying to test alternatives or a poor D5/D6 user <G>
 

Mike Swaim writes:
WinForms in that timeframe.
-Craig
--
Craig Stuntz [TeamB] . Vertex Systems Corp. . Columbus, OH
Delphi/InterBase Weblog : delphi.weblogs.com
InterBase 7 supports long-running transactions; read more here:
delphi.weblogs.com/stories/storyReader$379
 

Craig Stuntz [TeamB] writes:
So which to use? Guess I will just use the better one till the next
miracle comes along (be it Avalon or something else). <g>
--
ilkka hyvärinen
 

Ilkka Hyvärinen writes:
don't see either choice as being clearly better.
If you have legacy code in the Win32 VCL then use the VCL for .NET,
since it may well port with nothing more than a recompile.
For a brand new app I would seriously consider WinForms, since its finaly
class heirarchy is likely to be more similar to Avalon than VCL for
.NET -- but only if the limitations of WinForms today (smallish number
of components available, performance) are not showstoppers. I don't
think that either the VCL for .NET or WinForms are clearly superior for
new apps.
-Craig
--
Craig Stuntz [TeamB] ?Vertex Systems Corp. ?Columbus, OH
Delphi/InterBase Weblog : delphi.weblogs.com
Useful articles about InterBase and Delphi development:
delphi.weblogs.com/articles
 

Buch writes:
be snappier than WinForms apps. Yeah, the situation might reverse itself on
new machines three years from now, but that is not a given.
(such ar Ruby or Python) that much like C?
reasonably intelligent client, they have to go in the client. Other times,
for performance, they need to go in the server.
 

Craig Stuntz [TeamB] writes:
As I see it now, it will a case-by-case evaluation, the deciding
factors
will vary project-by-project and over time.
--
ilkka hyvärinen
 

Hi,
thats the one I read today.
Further even the founder of Mono themself are aware about there can be potential troubles:
If Microsoft decided to make our life really hard in terms of compatibility, it would also hurt its own customers. If
it changes the APIs, that affects its customers as well. So I think the APIs will remain fairly stable, and I hope that
Microsoft won't go into proprietary protocols or protocols that would make it really hard for us to implement Mono.
There's is always the possibility it will do so. Microsoft has some strange patterns in terms of how it competes. I
really hope it will "behave like a good citizen," as Steve Ballmer said recently it would.
www.ftponline.com/wss/2002_07/online/netalternatives/page4.aspx
Thus putting ones eggs in the dotNet basket only could potentially lead to yet another MS monopoly. And I'd be
seriously surprised if that wouldnt happen. MS is in the world to earn money, not to make developers life easy.
purposes, from a cost/benefit point of view, be better to stick with what you have than with something new.
Surprisingly many people are still coding Fortran and Cobol even though 'better' or more advanced languages has existed
for decades.
are going to work on the Longhorn platform for many years to come. If not, I predict Longhorn a terrible death, and
Linux its total victory.
fight more fiecely against eachother in the n-tier arena? If we would, then it would be because we find that more fun.
It wouldnt be because it would be economically wise.
I know of several VB'ers that got rather tired of VB(Visual Basic) when they discovered that they had to rewrite large portions of
their code to work under dotNet.
--
best regards
Kim Madsen
XXXX@XXXXX.COM
www.components4developers.com
The best components for the best developers
kbmMW - kbmMemTable - kbmWABD - kbmX10