[Harbour] GTWVW => GTWVG Port - A Clarification by Pritpal Bedi

CCH: Quoted verbatim from a thread on the Harbour Developers List


Hi Budiyanto, All

> If you mean revising (or rewriting) GTWVW into Harbour, then I'll be very
> much interested, as well as
> many others. But please keep it backward compatible as much as possible. I
> remember you said here
> in the past that GTWVW is "impossible" to port to the new GT API because
> there is a conflict with
> GTWVW's use of first parameter as window designator (for most functions).
> Do you finally find a way
> to solve that problem? Or do you plan change the convention in GTWVW? (I
> hope it is the former.)
>
> Am I interpreting it right with what you mean by "porting GTWVW in GTWVG"?
>

When GTWVW came into existance I had already ported my applns to GTWVG,
the then GTWVT+WVTGUI, so I never looked into GTWVW deeply. Everything
I needed was available in GTWVG ( current name ). Rather at that time
I thought pseudo emulation of controls was good enough to appeal my clients
and I was saved of the time to learn a new protocol. With only little
efforts I was able to roll my applns with a new look and feel.

Over the time concepts changed. Multi-Window and Multi-Threading forced me
to look into real windows world. I was convinced that a lot can be achieved
on this front as MT and GT modal of Harbour is much superior to any other in the
market today.

I looked around for effective GUI modal to base my next developments
targetting not only my own needs but the Clipper community in general. The available
modals in-use for Harbour around were as follow:

1.  FWH
2.  VXH
3.  HWGUI
4.  MINIGUI
5.  XAILER
6.  GTWVW

Every modal had its pros and cons.  Some had documentation missing,
some were pure IDE based, some dynamic but heavily based on #define(s)
( due a large number of parameters for a single activity ), and some lacked
OOP.
Every modal had its own parameters passing and calling convensions.
And, for sure, all modals lacked adaptation to MT, the biggest mainstay
in the coming future.

What I wanted to base next efforts on development of a GUI :

1) It must be adaptable to MT modal.
2) It should be able to take advantage of MW GT whereever possible.
3) It must adhere to standardization of parameters and calling convensions.
4) It must be having a well written documentation.
5) If possible, it must have a user-base.
6) And above all, it must be able to base Multi-Platform GUI.

After a lot of thinking and spadework I could not accept any of the above.
I just picked Xbase++. This compiler's subsystems were almost fitting
my ideology, is having a large user-base, ships with descent documentation,
complete with MT capabilities and nicely fits into Multi-GT environments.

Moreover I had worked with this compiler for some time and know how
best its parameters passing and calling conventions were. Also I wanted to
have a well define class framework to dream for a multi-platform GUI framework.

I knew to follow an already established framework is always a difficult
task.

To create a new protocol you are the master of your own thoughts. But to
copy feature-by-feature is very difficult. But I choose the difficult way.

Also I wanted to have an integrated library which could absorb different
scattered GUI modals into one as all were having one aspect in common -
Harbour.

My first results are quiet encouraging and now GTWVG can host:

* pure console,
* pseudo GUI console,
* Xbase++ consols and dialogs with xbase parts,
* pure consoles with Xbase parts,
* and a mixture of any of the above.

In this evolution process GTWVW becomes the next natural GEM to
be inducted in this whole picture, and probably FWH too, who knows.

If you invest a little time in GTWVG's Xbase++ implementation, you
will wonder how a Xbase++ .prg ( with only those controls which have
have taken some shape ) can be compiled in Harbour. And this is the
same intention for GTWVW too. No code change whatsoever.

As before, today also i say that it is difficult to port GTWVW as is
in new GT system. As GTWVG and GTWVW share the same code base at the
core levels it is easier to embed GTWVW GUI controls into GTWVG.
End user will not experience any change to his/her code. This way
it is not as difficult as it seems to be.

Also to rewrite it again will be sheer waste of time and resources.
In fact I could never take advantage of GTWVW because it become mutually
exclusive with GTWVG, and which I do not want at this time. Possibly
in the coming few days I will lay the base structure how GTWVW
functionality can be included in GTWVG. There onwards there should be
cummulative effort to achieve this goal. I alone can achieve but it may
take a bit longer.

To conclude, GTWVW is not going to be rewritten but will be included
in GTWVG and for sure end-user will need a minimum to change his code.


Hope I am clear enough.

Regards
Pritpal Bedi