Re: Delphi and smart-devices
From: Danny Thorpe (dthorpe_at_bozofilter.borland.com)
Date: 02/24/05
- Next message: Dennis Landi: "Re: interview with danny"
- Previous message: Danny Thorpe: "Re: Thank you Borland!"
- In reply to: Peter Overweel: "Delphi and smart-devices"
- Next in thread: John Jacobson: "Re: Delphi and smart-devices"
- Reply: John Jacobson: "Re: Delphi and smart-devices"
- Reply: Rick Beerendonk: "Re: Delphi and smart-devices"
- Reply: Peter Overweel: "Re: Delphi and smart-devices"
- Reply: Nick Hodges [TeamB]: "Re: Delphi and smart-devices"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 23 Feb 2005 16:20:25 -0800
Peter Overweel wrote:
> Hi,
>
> The article "How to Persuade Borland to Do Anything" from 14-04-2004
> emphasises on presenting the problem, rather than presenting a
> possible solution.
> (http://blogs.borland.com/dcc/archive/2004/04/14/2388.aspx)
>
> So here's my problem:
> I want to develop applications for (off-line) devices like
> smart-phones and PDA's that can synchronise with my server when they
> get online again. Preferably with some advanced
> conflict-manager-tool-stuff-component-thingy. I know and like
> Delphi, so it would be nice if this can be done RAD with Delphi.
>
> ...and...
>
> ..arghhh... ..Sorry... can't help it....
> ...Must...
> ...suggest...
> ..solution....
>
> In Ms Vi$ual Studio, developing for smart-devices is possible thanks
> to the Compact Framework. And if dotNet is still the future (which I
> assume), Delphi should also. But-correct-me-if-I-am-wrong, Borland
> hasn't got the license to integrate the CF with Delphi???
>
> Borland wake up! Your losing a customer here!!
Thanks for the feedback. Borland has been in negotiations with
Microsoft for the past three years on how Microsoft can support third
party tools vendors to produce tools for the .NET Compact Framework.
To summarize the discussions and the lack of forward progress, it
mostly boils down to tribal politics and engineering shortsightedness
within Microsoft.
The Microsoft team responsible for building the .NET CF runtime has
been very supportive of Borland and done everything they can to pave
the way for Borland tools to support the CF platform. The CF runtime
team, however, does not have ownership or control of the CF design
surfaces.
In the desktop .NET Framework, the WinForms form designer is actually
part of the framework. Third parties such as Borland can instantiate
the form designer and provide it with hosting interfaces and away you
go - you have a design environment.
In .NET CF, the designers are part of Visual Studio. Since small
footprint was the highest priority for CF, design-time support was the
first thing cut from the runtime footprint. This is understandable.
What makes things difficult is that the CF designers were not designed
to support third party tools. It's not just a matter of Microsoft not
wanting to license the CF designer to third parties such as Borland,
but also a matter of baked-in design limitations - the CF designer is
so interdependent upon the Visual Studio architecture that a non-VS
tool vendor would not be able to do anything with it even if the CF
desginer were licensed.
This is truely sad, because we know by the WinForms example that it
doesn't have to be this way. The WinForms designers were built by the
"new-school, welcome everyone to our platform" Microsoft. The CF
designers were built by the "old-school Visual Studio, third party
tools must die" part of Microsoft.
If you think I'm being melodramatic here, check this out: Even if a
third party were to license the Visual Studio architecture to build
their own IDE using VS bits, those VS bits do not include CF designers.
You can't get the CF designer *without* VS, and you can't get the CF
designer *with* VS. So, no matter which way you slice it, CF is
screwed. By Microsoft.
Borland's options for CF support are:
1. Wait for Microsoft to support third party development on the CF
platform.
2. Reverse engineer the VS CF designer architecture to build our own
designer against the CF framework. Reverse engineering a black box as
big as the .NET framework is a very expensive proposition. This route
runs a very high risk of litigation and exposure to unknowns hidden
inside the black box.
3. Build something else that runs on the CF IL platform but not using
Microsoft proprietary bits. This gives us the greatest freedom and
flexibility, but also the greatest cost to develop. This option will
always be playing catch-up. This option is also the least attractive
to Microsoft, since it threatens to split the CF platform into
incompatible camps just like Win CE's fatal fragmentation.
It's a question of business, not technology. Borland can certainly
build CF tools without Microsoft's blessing or assistance, but it will
be much harder and much more expensive than if we had the cooperation
of the platform vendor. "More expensive" means we have to see an even
larger financial return before starting down that path. CF development
would be a great feature to round out our tool suite and would drive
additional sales of the tool suites, but it is not hugely lucrative
market by itself - certainly not as lucrative as Win32 or .NET
development as a whole.
I am hopeful that we will be able to break this impasse with Microsoft.
My impression is that most of the tribes within Microsoft recognize the
value of having independent third party tools supporting Microsoft
platforms and indirectly driving sales of Microsoft platforms. There
are only a few groups in Microsoft that appear to still place greater
priority on protecting sales of a specific product (VS) against
competition over enabling and driving third parties to promote sales of
the underlying Microsoft platform. Unfortunately, CF design support is
owned by one of the latter groups.
The good news is that Avalon appears to be leaning toward the WinForms
designer mentality rather than the VS "all your base are belong to us"
model that created this CF designer blockade.
Stay tuned.
-Danny
-- Delphi Compiler Core: http://blogs.borland.com/dcc
- Next message: Dennis Landi: "Re: interview with danny"
- Previous message: Danny Thorpe: "Re: Thank you Borland!"
- In reply to: Peter Overweel: "Delphi and smart-devices"
- Next in thread: John Jacobson: "Re: Delphi and smart-devices"
- Reply: John Jacobson: "Re: Delphi and smart-devices"
- Reply: Rick Beerendonk: "Re: Delphi and smart-devices"
- Reply: Peter Overweel: "Re: Delphi and smart-devices"
- Reply: Nick Hodges [TeamB]: "Re: Delphi and smart-devices"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|