Board index » delphi » Re: How would you steer Delphi if you were Nick?

2006-09-20 02:49:24 PM
delphi157
Brian Twinings writes:
parts of the .NET framework, than to shell out for 3rd party libraries
that might not perform any better.
--
Dave Nottage [TeamB]
 

Jolyon Smith <XXXX@XXXXX.COM>writes:
for it, and some have preferred to go down a different path. Big deal, you can't
please everyone, but just because some can not (or are not prepared to) see or use the
advantages doesn't mean that others won't. I will just focus on those that do, and
I'm not going to lose any sleep if you don't.
--
Dan
 

Hi!
Ever since JAVA was introduced that was also advertised, but
it never happened. There is one thing if you write software
that makes app user life easier. But it is vastly different thing,
if you concentrate on having the programmers life easier.
In my view, buisness apps and small utility apps that dont require
distribution beyond company borders will become more
.NET based, but for the large part not because people would
make the choice, but because MS left them no choice.
Delphi is an alternative, but since it is not made by MS, it is
not a choice either for many....
the main focus of Delphi to be a .NET development platform but
rather a .NET supporting platform. Instead of offering 1:1 features
against VS.NET I'd concentrate on a few o them and do those
well. There are very few people that use more than 5% of the VS
or BDS IDE and their tools power.
Delphi would be dead and the future would be bright for all.
because whatever you port, it wont work as well anymore. .NET
has performance, (memory/code speed) and distribution problems.
Its only when you use features of .NET CLR that are not available
in Delphi is when you gain something. But.... That is not the case
for existing apps, simply because they already exist :)
used for W32/W64 development with its .NET features used for .NET
interop and support. That is the place where Delphi can compete...
platforms. It is time to focus on the one and single platform: Windows.
Regards!
Atmapuri
 

important. By that, I mean can DevCo build their future on it?
I don't think so. It is valuable for some developers, undoubtedly, but this
is just a transitional thing. Moving apps back and forth between W32 and
.NET is technically very clever, but how many people *really* need to do it?
And for how long?
Surely the great majority of W32 apps will stay as W32 for as long as
feasible, and new apps will be built using *either* .NET or W32, but not
both.
Thack
 

I'm not familiar with Chrome at all, but it sounds fascinating.
Trouble is, hasn't it rather stolen DevCo's thunder in terms of a revamped
Object Pascal for the .NET environment? Perhaps DevCo should consider an
acquisition, here...
Thack
 

What I personally would like to do is the following (in order of
preference), however it might not make any business sense whatsoever.
1) I'd make ECO a VS plugin. I think it is more advanced than anything
out there, but if DevCo continue to implement it only in Delphi then they
are selling to a much smaller market than they could. This would create a
good income stream to help fund the following items.
2) I'd revolutionise "writing applications" in the same way that visual
components revolutionised forms development. To do this I'd expect the
following:
A) A business objects layer (ECO for .net, something also for Win32)
B) The same kind of thing (modelled) for the application logic layer
C) Excellent support for patterns in the BO, AL, and in code. I think
that business apps should be built on patterns and not inheritance. Believe
me, pattern catalogues are the future!!!! :-)
With ECO I find the business objects *very* easy to implement, I think that
native windows users would enjoy benefiting from the same kind of thing.
However I still find that I spend far too much time writing the app to
support the business objects in a GUI. I would like to see an application
framework which is so separated that
A) The app layer can run on another PC.
B) The app could be presented as either a WinForm/Windows/Web
application, or consumed from a webservice.
Again this could be made available for VS customers too.
3) I'd want to see a Win64 compiler.
4) I'd like to see garbage collection and reflection.
I would like to see DevCo become a "developer tools provider" rather than
only a "provider of an IDE with development tools in it". I want it to
become a company full of people doing great new + fun things. The sort of
things that coders *want* to do, the sort of thing they'd like to write if
they were working at home from their garage. Great new ideas!
--
Pete
Blessed are the geek, for they shall public class GeekEarth : Earth {}
====
Audio compression components, DIB graphics controls, ECO extensions,
FastStrings
www.droopyeyes.com
====
 

Staiger writes:
never move back and forth to Win32, but some of which is now being used
in .NET, and the rest potentially so.
.NET code.
--
Dave Nottage [TeamB]
 

"Brett Watters" writes:
assume when you say 25MB you are resorting to the old canard of adding the
size of the .NET runtime to the size of the app and then throwing up your
hands and saying OMG it is so huge as to be impossible.
I have not one but two .NET sys-tray apps running all the time on my system.
One is .NET 2.0 SQL Server Service utility to turn on/off SQL Server 2005
and the other is the new New York Times NewsReader that uses WPF, one of the
first commercial .NET 3.0 apps I might add, read about it here:
firstlook.nytimes.com/
www.wired.com/news/technology/0,71740-0.html?tw=wn_index_2
When I was deciding to leave the NewsReader running all the time I was
concerned about how much memory it was allocated according the the Task
Manager, namely about 50 MB when maximizied but that reduces to 6MB when
minimized to the sys-tray. By way of contrast I also keep FeedDemon, a
Delphi Win32 app, running almost all the time and it currently has 40MB of
memory allocated when maximized and it remains at 40MB when minimized to the
sys-tray.
 

"Atmapuri" writes:
you on the subject of why it is not wise to port Delphi Win32 apps to .NET.
;-)
 

"Staiger" writes:
2) Talk about it.
Best Regards,
Danijel Tkalcec
 

Hi!
I think the problem started in 2002 when Delphi planners went to
the Microsoft .NET conference. .NET was presented as a platform
although it is really not. It has some properties of a platform, but...
So, If you already tried to support another platform: Linux,
the simplest thing to do would seem to take the recepie from
Kylix and apply it to .NET. The funny thing is, it did not work
out for Kylix and for a different reason it did not work out for .NET.
The Kylix was used mostly as a recompile tool and not as the tool
in which actual development was being done.
So, to actually address Linux, Kylix should have been a
simple recompile deployment tool and not a complete
development IDE. But when looking at the license count
that you can sell, you would either have to scratch the project
or think of a different licensing model.
And as for .NET which is not really a platform, they should have taken
the same recepie as for supporting ActiveX instead although in
much more involved way (mixed code compiler).
Regards!
Atmapuri
 

JED writes:
The CF designers seem to be pretty much hardcoded for C# and that ugly V-thing.
Integrating with VS lets you stay on the edge and have MS to do the hard part, I guess.
Chrome was sitting in VS05 *way* before the 2.0 RTM, it is/was still sitting in VS03 , too.
 

Size".
This will add a memory size column that doesn't suffer from the "magic
memory usage reduction" phenomenon, aka. "how to hide some .Net memory
usage from the unwary windows user".
Eric
 

Usually you want to make something new and exciting, not writing similar
software again and again which is merely patterns put together.
Focusing on other things than business apps and logic would help differenciate.
AI is a great domain, embedded systems, intelligent applications, etc. That can
be a revolution if you make apps more user friendly and understand what you
want, help you formulate things shorter, tell you possible contradictions or
omissions.
Another thing is proposing a programming environment for things like LEGO
Mindstorms (Roboters and co), including simple AI capabilities. Making it easy
enough for young people to play with.
 

Mark A. Andrews writes:
If you target business users I am sure you can make a profit. Linux has
made and is constantly making ground as a server OS platform and even on
the Desktop. Going for the home users would be very difficult market to
crack.
Siegfried