New Application Development

[JamesBowen]

Hi all.

My boss has drummed up some new business in the Asia/Pacific (Japan and Korea) market and tolled our new clients that our database is Unicode.

However there was a miss communication saying that our database was Unicode, and that out database can be converted to Unicode.

My real problem is that the current application in CHUI and I need to change the interface. Due to the previous developer not understanding the principles of "MVC" I am going to have to rewrite the entire application.

So if I am going to rewite the application from scratch, how should I do it? Should it be Web-based application or Windows GUI ABL/ADM?

My gut feeling is that it should be Windows GUI running on OpenEdge WebClient is this the best approach?

My other problem is my ADM/SmartObject Skills are not that great so development time will be longer.

Has anybody got any ideas and suggestions.

Also our current Printing can not handle Extended characters (i.e. Unicode) which is done on Linux.

Current
OE 10.1B running on Linux. RH4SE

Thanks.

 

[timk519]

My real problem is that the current application in CHUI and I need to change the interface. Due to the previous developer not understanding the principles of "MVC" I am going to have to rewrite the entire application.

You don't need MVC to separate the UI from the BL, and then bolt on the UI's as appropriate.

This may be a serious restructuring of the program, but it's not a re-write - unless the base code really that bad.

 

[tamhas]

Of course, given a lot of the ABL code I have seen in existing applications, it could be that bad ...

But, more positively, there are a bunch of questions you should be asking yourself and your boss (along with that bit about promising things that don't exist) Budget, time frame, how bad do you really want the business, what else might you want to do with the software in the future, are there other development demands in the queue, etc. And, too, you need to analyze where you are ... how much of the code needs to change, is there any UI/BL separation, how many screens and programs, how much is just fine the way it is, etc.

Having done that, you need to figure out where you want to go. There are at least 7 or 8 basically different technologies you can use for a GUI or WUI front end and they each have their good and bad points. Some, for example, are ideal if you only need to throw up a few new screens; some are good is you have an existing two tier application; some are good if you have a reason to want to reach out to suppliers or customers; etc.

And, you should be prowling around PSDN and OE Hive for material on OERA so that you have a better idea of a good target architecture, although there aren't yet any really good model code sets suitable for production use.

You might want to check out http://www.oehive.org/UMLFromABL as it proceeds because this will be a tool which helps take an existing, poorly documented system, and to analyze it so that it can be restructured.

From all of that one can build a plan (and hope the budget is there to support it). If you don't have a good architect in the company, you might want to buy some consulting in the planning of the transformation process
and the development of the new architecture.

 

[JamesBowen]

Ok. I have taken some advice on board from my Progress Account Manager. They have recommended that stick with using Webspeed for developing my Client Side application.

Reasons :
1. Unicode compliant.
2. We already have the OpenEdge licensing setup.
3. A existing web Framework is already setup (although it does need some changes).
4. Zero deployment time.

Consideration:
1. How to get Dynamic HTML populated lookups i.e. a modal lookup selection list.?
2. Keyboard short cut keys? i.e. CTRL+S to save CTRL+A to Add etc.
3. Javascript, to handle client-side validation. i.e. Date validation I know very little Javascript.

Thanks

 

[tamhas]

Want to buy a bridge?

 

[JamesBowen]

Want to buy a bridge?

Nope but I do want to buy a Shed to hide in.

 

[tamhas]

Well, if you believe the rep, you'll need it.

 

[TomBascom]

Other than shooting down ivory tower dreams of a rewrite with a dose of reality just what is wrong with the reps response? I sure sounds a lot more realistic...

Sure, the current state of affairs might not be perfect but it is a bird in the hand. I somehow doubt that a guy whose boss has done what he has done is going to buy into the originally proposed scheme. You might be able to get some traction towards a transformation but I wouldn't be holding my breath if I were you.

 

[tamhas]

Nothing wrong with considering Webspeed as an option ... it is just the notion that it is necessarily simpler or faster or easier than anything else which I am questioning.

I especially question the notion of grab it and run without consideration of other issues since another path may well be better if one knew more about what else might be in the pipeline.