Web Services Architecture

[Sean Wheller]

Hello,

Seasons greetings to all.

Attached is a proposal, in graphic image format, for a TurboCASH based on XML
Web Services.

I make this proposal because I feel that this architecture provides a number
of benefits to the project:

Client Side
1. platforms on which TC can run increases
2. programming languages to develop client interfaces increases
3. deployment possabilities increase
4. separates concerns (rapid application development)
5. implements indirection (smooth upgrades)
6. Promotes application integration

Server Side
1. platform for development (SDK XML-based component framework)
2. client-server architecture (point-to-multipoint)
3. interoperability (machine-to-machine)
4. portability (oneday, Linux, Solaris, Mac OS, Windows)

The images attached are kept to the higher-level in order to make them as
simple to understand as possible. I hope they will encourage questions,
discuss, debate and most of all new ideas based around the same architecture.

Hope this helps,

--
Sean Wheller
Technical Author
se...@inwords.co.za
+27-84-854-9408
http://www.inwords.co.za

[Yogi Yang]

I personally think this would be an over kill

[Sean Wheller]

On Monday 02 January 2006 17:21, Yogi Yang wrote:
> I personally think this would be an over kill

Could you substantiate or give an alternative/better solution?

[Philip Copeman]

Yogi Yang wrote:

>I personally think this would be an over kill
>
>
>
>
>
>
>

I see TurboCASH and TurboCASH Server as two separate products. YEs our
current verison of TurboCASH will carry us clearly in the Windows market
for soem time yet. There is no need to amke major changes to thsi
program, mere get it talking to as many other products as possible.

THe TurboCASH Server will give a lead into the Linus Market, will
intereface woth the curretn TurboCASH and will help us build a web based
accoutnign bsuiness. Let me work on s[pecifiying it further.

[Sean Wheller]

On Tuesday 03 January 2006 17:34, Philip Copeman wrote:
> I see TurboCASH and TurboCASH Server as two separate products. YEs our
> current verison of TurboCASH will carry us clearly in the Windows market
> for soem time yet. There is no need to amke major changes to thsi
> program, mere get it talking to as many other products as possible.

For purpose of simplicity in developing the code it can be one and the same
product. Computers these days are strong enough to we workstation and server
at the same time. The minute a computer makes a share to the network it
becomes a server.

Turbocash installed on a workstation for use by one person, should scale
easily to meet the requirements of a company with 5 computers and 5
employees.

As the company grows the server end of the installation should be capable of
moving to a server computer and each of the client installations just pinted
to the new web service location.

Eventually the business grows more and time comes for e-commerce. Again,
turbocash should scale to integrate with solutions out there.

It can be the same code. Just options in the setup process.

Click:
1. Server Only
2. Client Only
3. Both the Above

>
> THe TurboCASH Server will give a lead into the Linus Market, will
> intereface woth the curretn TurboCASH and will help us build a web based
> accoutnign bsuiness. Let me work on s[pecifiying it further.

The Server can be developed on Windows for the near furture, but give the
users a choice of desktop. Later, if the project gains demand for a
Linux-based server, then you may find that it will bother some of the Linux
developers so much that they will decide to bite the bullet, scratch the itch
and develop a version of the turbocash server for Linux.

But in the interim, open and expand the possabilities for client side
developers. I think a few of the Java people lurking here may get exited. I
know for sure the PHP developers will, those guys jump on everything :-)

[Henry Eko H]

Hi,

great idea,
i pretty much aggree with your proposal.

i'm currently develop my version of these services. still in early stage.
there are benefits from these services / multitier approach.
business logic ( business services) and database handler (dataservices) can
be separated and independent.
so, development using different database is possible.

it will not be an overkill, because i have develops couple systems using
multitier solutions.
the performance is very much scalable.
(as long as no .net solution involved, .net will require alot of computer
resource and will limit low end computer from running it.)

i was once discuss this solution with pieter.
but it will require a major rewrite for turbocash.
the current delphi implementation is almost imposible to adopt these
services.

it will cost 3 menmonths and aproximately 10.000 euros. according to pieter.

my version of these service will come in two stages.
the current development is in early stage one.
providing basic functionality to lookup / references and entry screens. with
minimal to no accounting logic.
these will be tckernel like extensions.

stage two,
providing full functionality. moving all accounting logic and core logic to
services. providing new improved GUI.
off course, these is a major rewrite.

Salam
Henry Eko H

Kokutama Komputer
Roxy Square Lt. 1 Blok A9/ no. 15
Jl. Kiai Tapa no. 1 Tomang, Jakarta 11440
Telp. 021-56954332 Fax. 021-56954332