Re: Why aren't you upgrading?


Things we need right now and don't have with BDS 2006.

1. Stable IDE - this is first and foremost before anything else. I
don't want to see a single feature added to the IDE until it is super
stable.

2. Fast IDE - It needs to meet the speed and response of D7. If this
cannot be achieved with .NET then I suggest they switch to native only
code for IDE. Obviously to support .NET this is probably not an
option. But your customer base is claiming they don't want .NET they
want native code. And those that do want .NET 2.0 have already moved
onto C# or Chrome/VS combo.

3. MS alternate choice - As I see it the reason Delphi became so
popular is because it offered a different and better choice then MS
products.

- Easier and more powerful implementation of Win32 API and GUI via VCL
- Fast native code compiler vs a slow VB
- better database support
- better language features than VB (True OOP)
- 3rd party vendors releasing components with source (unheard of in the
VB world)
- no DLL hell, single EXE for delivery

Fast forward to today and it almost looks like:

..NET + Chrome + VS + ReSharper or CodeRush

would be the better choice than what Delphi is offering. If Delphi was
to provide a comparable framework to .NET that would be native compiled
code and in future support 64-bit, that would be a different and
innovative choice. It also might just bring back those Delphi people
who decided to move to .NET and VS 2005. If it built a better GUI
windowing system vs WinForms/WPF/XAML that was cross platform (Linux,
OSX) and rendered fast then that would be a different choice.

I think it has already lost the .NET race and is too far behind to try
and compete with C#/VS combination. It would make more sense to offer
package with a plug-in for VS IDE and a new Delphi language like Chrome
for .NET than to try and play catchup. Most people who want .NET 2.0
have already moved to VS and C# or Chrome.

Key problem is that Borland was known for true innovation in the past
and offering different choices. Then due to mismanagement they lost
alot of key people to MS, specifically the designer of Delphi/VCL. And
we can see from the attractiveness of .NET and C# that he did bring his
influence to MS in its design. So the choice Borland management is
left with is to get onboard with MS since they did not want to fund the
resources to do anything innovative with Delphi and the VCL. That
money went to the Core SDP and Inprise crap.

Today we have things like AOP in Java, Generics, Reflection, true full
Serialization of objects, and huge open source projects supporting both
Java (Hibernate, Spring, Struts) and .NET (NHibernate, dotNetNuke,
Atlas Control Toolkit, Piccolo, Mono).

But if you look at the VCL it is just not up to snuff when compared to
todays modern frameworks. And the choice Borland/DevCo leaves you with
is to move to .NET for these new things.

Delphi offers ECO, but look at the options .NET has:
http://csharp-source.net/open-source/persistence

DevCo, I think needs to stop thinking like Borland and do some radical
and innovative things. But, first deliver what their customers need
today, right now.

- Stable and Fast IDE
- Good Help system in IDE
- Generics and a better Collections Framework in Win32
- More RTL functions from FastCode project in Win32
- Unicode in Win32
- RTTI expanded to offer full Reflection like modern languages
- streaming and serialization of objects (both binary and XML) in Win32
- first class open Tools API and fully documented, so that 3rd Party
can focus on IDE enhancements (i.e. JetBrains, DevExpress)
- mobile device support (Palm, CE)
- version control support for Subversion, PerForce, etc.
- new VCL and language features that make it compelling to upgrade

Then announce that 64-bit native compiler is in progress and started.

And eventually announce what new features are planned for the native
VCL/RTL:

- ECO for Win32 or a simplier persistence framework, maybe one based on
Hibernate/NHibernate
or release the Bold Source to opensource
- MVC framework making use of tech. like AOP and XML based GUIs (i.e.
XUL, XAML)
- Cross platform support for Linux and OSX in the native compiler
- easy to implement multithreading programming support for dual cores
- native code interoperatibility with .NET code, I know Java can
iteract with .NET
- new TWebBrowser built around Mozilla XUL or ActiveX XUL control

Java Spring is a good example of the innovation we need to see in the
VCL. Spring allows developers to add cross cutting features to all
their code with having to change any code using AOP programming. So I
can add persistence, logging, transactions, security, MVC
implementation, etc without changing my code. And it makes it so
flexible that I can choose what persistence or MVC implementation I
want to use and can switch it out with no changes to code or classes.
All this is done via the technology of AOP. However, native Delphi
would need language improvements in RTTI (full reflection) and more
powerful interfaces to be able to accomplish anything like this.

I think Borland needs to get back to its roots of offering developers
different choices than Microsoft. The classic Borland market and
customer base was built on their customer's love of fast compiled
native code and easy to use VCL. And I think DevCo should focus on
just this right now.

Maybe DevCo could still use .NET at its core but have Delphi compile it
down to native code. And wrapper the WPF GUI framework into its VCL
and have it generate native implementations that use fast hardware
accelerated features. Maybe it could even generate native GUIs for
Linux and OSX in future versions from the same code. It could then
market as, Hey, use .NET but make sure you use Delphi so it compiles to
a fast native code application for any platform. Wow, would that be
impressive. However, I am not even sure if one can take a VM language
and make it compiled. But I thought I have seen this done before for
Java.
.

Relevant Pages

  • Re: Oh man!
    ... a native code DELPHI language IDE. ... platform. ... I dont even want the IDE - just a cross compiler would be ... You could write the code using Delphi on windows and then cross compile it to any platforms supported by FPC/Lazarus. ...
    (borland.public.delphi.non-technical)
  • Re: Chrome - competition for Borland?
    ... Ultimately the goal would be to have everything in a single language, ... in a single IDE, and Chrome would be a stepping stone to that goal ... Delphi would be for legacy, but for that task, there is already D7. ... basic Windowing support and native compilation in some areas (latest batch ...
    (borland.public.delphi.non-technical)
  • Re: Opportunity passed by
    ... Delphi for .NET on the ... hand has its own compiler that has to be modified to use the latest things ... has folded into .NET, ditto for a special version of the VCL, and the IDE ... C# developers, as they want/have moved to .NET 2.0, and Delphi ...
    (borland.public.delphi.non-technical)
  • Re: CodeGearT Introduces DelphiR for Win32R, Providing RADDevelopment for Windows VistaR and AJAX
    ... So a project would be say a "Delphi 6.0 project" but hosted in the newest IDE. ... The IDE would come with all the old versions of the compiler, and a project option would indicate which version to use at compile time. ... my projects use a lot of stuff, including custom packages, and custom visual controls build with the CDK, and /those/ are what really keep me from "simply" getting the projects into the newer IDE - not the compiler and core RTL itself. ...
    (borland.public.delphi.non-technical)
  • Re: Moving from Delphi to C++
    ... >> Borland develope the native Delphi compiler. ... > Yep, and just like MS, should also include 64 bits support. ... Just one language for most. ...
    (borland.public.delphi.non-technical)
Loading