New Development

New Development

Post by David Ra » Thu, 02 Dec 2004 01:50:35


I've been out of development for a while, and have a small (3/4-months)
project I'm considering doing in BCB, in spite of its apparent impending
demise.

I'm having a difficult time seeing what I need .net for. Is my application
likely to become unusable when the .net platform is ubiquitous?

TIA
 
 
 

New Development

Post by Alisdair M » Thu, 02 Dec 2004 01:58:53


No, but they might start looking old fashioned, in the same way that
software only supporting Win3.1 GUIs looks old fashioned today. As new
kinds of features become standard with OS support, you will have a
harder time delivering them. I can't tell you what 'must have'
features you might miss, as MS might not have imagined them yet! But as
the .NET matures, you will notice more and more.

For instance, a 'well behaved' WinXP application will have theme
support. That is already hard to support in BCB6. Who knows how much
further theming will go in .NET? Different transitions/animations as
the user moves the mouse across the screen? Sound effects on cue? 3D
GUIs based on DirectX? All these and more have been rumoured at
different times.

AlisdairM(TeamB)

 
 
 

New Development

Post by Andrue Cop » Thu, 02 Dec 2004 02:04:03


Probably but I don't think that point will be reached for many years.

--
Andrue Cope [TeamB]
[Bicester, Uk]
http://www.yqcomputer.com/
 
 
 

New Development

Post by Mike Marge » Thu, 02 Dec 2004 05:50:57

> I'm having a difficult time seeing what I need .net for. Is my application
For now unless you are doing asp web stuff, .net doesn't really offer
very much. Winforms is a step back from the the VCL.

Win32 will be around for a very long time. You can still run DOS apps
right?
 
 
 

New Development

Post by Andrue Cop » Thu, 02 Dec 2004 18:04:30


You can but Win32 DOS boxes have never been particularly good. Dodgy
mouse support, poor performance. The reason I kept my OS/2 box for so
long was that it had far superior DOS boxes. Only Microsoft Virtual
Machine finally solved that problem.

--
Andrue Cope [TeamB]
[Bicester, Uk]
http://www.yqcomputer.com/
 
 
 

New Development

Post by Graeme Pre » Thu, 02 Dec 2004 21:04:04


What does MVM have to do with DOS boxes ??
 
 
 

New Development

Post by Andrue Cop » Thu, 02 Dec 2004 22:12:42


I actually meant Microsoft Virtual PC. It lets you create a virtual
machine and then install any OS you want. I created a VPC that has DOS
6.22 and Win3.1 on it. It doesn't help completely with the performance
issue but does remove the issues with mouse support and the other areas
where the Windows DOS emulator gets it wrong(*).

Running full screen you'd hardly know you weren't using the real thing.

For performance OS/2 VDM is still better but Virtual PC is good enough
in the other areas to allow me to finally pension off my OS/2 box.

Video pallettes are one area. Using a Windows VDM you can never be sure
quite how BC3.1 will appear. For some reason the correct pallette is
the one it is least likely to pick.
--
Andrue Cope [TeamB]
[Bicester, Uk]
http://www.yqcomputer.com/
 
 
 

New Development

Post by Mark Jacob » Sat, 04 Dec 2004 09:20:56

So might it not be fortuitous to perchance, wait until the .NET toolsets are
more evolved and easy to port BCB projects over to? Say, 5 years hence?
--
Mark Jacobs

"Alisdair Meredith [TeamB]" < XXXX@XXXXX.COM >


|
| > I'm having a difficult time seeing what I need .net for. Is my
| > application likely to become unusable when the .net platform is
| > ubiquitous?
|
| No, but they might start looking old fashioned, in the same way that
| software only supporting Win3.1 GUIs looks old fashioned today. As new
| kinds of features become standard with OS support, you will have a
| harder time delivering them. I can't tell you what 'must have'
| features you might miss, as MS might not have imagined them yet! But as
| the .NET matures, you will notice more and more.
|
| For instance, a 'well behaved' WinXP application will have theme
| support. That is already hard to support in BCB6. Who knows how much
| further theming will go in .NET? Different transitions/animations as
| the user moves the mouse across the screen? Sound effects on cue? 3D
| GUIs based on DirectX? All these and more have been rumoured at
| different times.
|
| AlisdairM(TeamB)
 
 
 

New Development

Post by Alisdair M » Sat, 04 Dec 2004 17:30:24


Yes and no. Imagine how much more functional your software will become
in 5 years? What you gain while waiting for the better development
environment to appear is offset by the extra work you are adding all
the time!

That is where attention to classic design details, such as separation
of UI / app logic / database access pays off. If the core of your
system is portable C++, you only have to port the environment specific
parts (GUI and Database-layer) If you can find a portable solution for
these parts, even better!

This is also where something like VCL.NET is extremely useful. It
might not be the identical library to Win32, it may not show off all
the whizzy new .NET GoodnessTM, but it is the 90% solution that gets
you up and running on the platform in next to no time. Then it is your
choice whether you need to investigate a different long-term solution,
but at least you have arrived.

AlisdairM(TeamB)
 
 
 

New Development

Post by Mark Jacob » Sat, 04 Dec 2004 19:00:39

But the issue you seem to be side-stepping with this .NET raving you're so
keen on, is that the runtime environment is too demanding for all but the most
powerful PCs at the moment. The runtime needs about 128MB RAM for starters,
and at least a P3-800 to run reasonable well. It won't go on Win9x, and now
there are 2 very different versions out 1.1 and 2.0 which even have different
language constructs, so compatibility between these 2 versions is already an
issue. Microsoft have done this in true "SP2" style - an absolutely shoddy
show, with lots of user fallout. I'm sorry, but the best advice anyone can
give about .NET (especially if you are *NOT* doing anything ASP) is to wait
and see.
--
Mark Jacobs
DK Computing
http://www.yqcomputer.com/

"Alisdair Meredith [TeamB]"



|
| > So might it not be fortuitous to perchance, wait until the .NET
| > toolsets are more evolved and easy to port BCB projects over to? Say,
| > 5 years hence? --
|
| Yes and no. Imagine how much more functional your software will become
| in 5 years? What you gain while waiting for the better development
| environment to appear is offset by the extra work you are adding all
| the time!
|
| That is where attention to classic design details, such as separation
| of UI / app logic / database access pays off. If the core of your
| system is portable C++, you only have to port the environment specific
| parts (GUI and Database-layer) If you can find a portable solution for
| these parts, even better!
|
| This is also where something like VCL.NET is extremely useful. It
| might not be the identical library to Win32, it may not show off all
| the whizzy new .NET GoodnessTM, but it is the 90% solution that gets
| you up and running on the platform in next to no time. Then it is your
| choice whether you need to investigate a different long-term solution,
| but at least you have arrived.
|
| AlisdairM(TeamB)
 
 
 

New Development

Post by Alisdair M » Sat, 04 Dec 2004 19:36:32


I don't know about raving, I know we have no plans to move to the .NET
platform any time soon. But the platform *is* there, already offers
advantages for some shops and is Microsoft's announced direction for
the future. The latter means it is something no Windows developer can
ignore forever - and BCB users by definition are Windows developers <g>

Hardare issues don't concern me at the moment, a bargain ba *** t PC I
picked up last week for <$500 is easily capable of running .NET. By
the time it *is* MS mainstream offering, enough users will have
upgraded that the majority of us can bite the bullet and move on. That
is unfortunately a requirement of MS shifting platforms.
(Yes, I understand you want to carry on supporting existing customers,
we still have code to handle bugs in Vanilla Win-95 in our library, pre
OSR/2, although I think we finally replaced that hardware 18 months
back)


And at the moment, I think that is great advice <g>

I have a suspicion VS.NET 2005 (or whatever they call it) will really
get the .NET ball rolling, and allow another 12 months for that ball to
pick up momentum. On that basis, .NET would be emerging as a viable
maintstream platform in around 18 months, which probably ties in well
with Longhorn for good measure. But again, that is my gut talking
rather than informed speculation ;?)

AlisdairM(TeamB)
 
 
 

New Development

Post by Mark Jacob » Sat, 04 Dec 2004 22:15:35

Like a belly-laugh, letting your gut do the talking can be very therapeutic
(unless you've eaten too much curried egg)!
--
Mark Jacobs
DK Computing
http://www.yqcomputer.com/

"Alisdair Meredith [TeamB]"


| ... which probably ties in well
| with Longhorn for good measure. But again, that is my gut talking
| rather than informed speculation ;?)
 
 
 

New Development

Post by Chris Uzda » Sat, 04 Dec 2004 22:34:21

"Alisdair Meredith [TeamB]" < XXXX@XXXXX.COM > writes:


<cough>


This is a good point. I remember one game that by the time it was
ready to deploy, their entire graphics subsystem was so outdated that
they had to rip it out and update it. And again it was behind the
times by the time they were able to ship.

Sometimes you have to design your stuff to account for development
time being long enough that some reasonable decisions for today may be
bypassed by Moore's law enough that they're unreasonable at ship date.

I think .NET is in that category. Its demands will be reasonable by
the time it gains critical mass. (In fact, it cannot happen any other
way.) With a company as big as MS behind it, it *will* gain critical
mass. They can buy success for a crappy product, as history has
proven.

I don't use .NET, I don't intend to use it, and I'm only vaguely
interested because it appears it's going to affect C++. I do like
some of the features they're pushing toward C++, and considering some
of the big guns (Stroustrup, Plaugher, EDG, Sutter, and other
experts), the good results are not surprising.

Not that I'm in favor of pushing MS technology, but I do like many of
the C++ extensions and hope many become part of the official C++
standard, not just "another" standard.


Yes. Though if you plan to continue working with windows, during the
waiting period it makes sense to dabble.


I'd guess it picks up momentum for the next 18 months, but won't be
truely "mainstream" for at least another 2-3 years.

--
Chris (TeamB);