The Future of PFC

The Present

PFC is enjoying it's highest popularity since the initial release. There are many reasons for PFC’s success. Neutral standards, third party support, QA, documentation, training, technical support, free peer support in newsgroups and mailing lists, and overall technical quality have converted a great majority of PowerBuilder developers into PFC believers. Compared to the prior year, the 98 LA conference had twice the number of PFC sessions. Every PFC session was packed. Few sessions such as Millard Brown's "LUW service" were standing room only! But… just as the PFC sessions were testing the capacity limits of the LA Convention Center rooms something strange was in the air, and it wasn’t just smog.

Rumours

While there were exciting announcements about PowerBuilder’s future directions, no public announcements regarding the future of PFC were made. And the rumours started to fly...

I heard some rumours about Sybase cutting down on resources for development of the PFC. I even read Sybase was planning on discontinuing the PFC.

..... I have heard that Sybase is not going to give any more new releases on PFC after PB6.5. Is it true? We all feel like that the PFC & PB are the best part of their application software suite. It is sad that they are not going to do any more enhancement for their best product.

The facts are that Sybase did not announce any of the above. They have never mentioned that there will not be any new PFC releases. Sybase did not say that PFC would no longer be supported. Finally Sybase most certainly did not say that PFC would be discontinued. Still the rumours persist. At a times like this silence could be very unsettling. Imagination runs wild trampling anything that stands in the way. Why the silence? In contradiction to the pessimistic outlook there could be other explanations. Non-disclosure is a common practice for projects in early development stages. Could it be that Sybase does have future plans for PFC, but is not ready to announce them publicly? When projects are early in the design stages base architectural directions can turn 180 degrees from one day to the next.

If mere silence has caused the rumours, can you imagine what a wrong announcement can do?

But unrelenting need to know is not going away. If Sybase is not directly going to tell us their plans for PFC how are we going to find out? One way is by doing some detective work.

What if…

What if I was placed in charge of the PFC enhancement project? I can dream can’t I? As every conference announcement suggested PowerBuilder is going full steam into a partitioned, distributed, multi-tier, component based environment. The down side of the new direction is that it introduces more complexity. There is a tremendous need for a class library that would tie all the pieces together and give a roadmap to the new archiecture.

What would I do to PFC? The first thing I would do is to extend it to provide middle tier functionality in the distributed environment. Unfortunately the first task is the most challenging. The existing PFC library contains a lot of useful features for the UI layer. However taking the PFC as is and moving it to the middle tier is not practical. PFC was never intended to be used in a distributed environment. Still, about 30% of PFC 6.5 services could be used in the middle tier. The decision of carrying the luggage of 70 % of visual services to use the PFC on the middle tier may be justified by PFC end users on a case by case basis. However, that decision cannot be applied as a rule. The official Sybase guideline must be "PFC is a client UI library".

That leaves a second approach. Given that existing PFC would be used on the client the task at hand would be to create a separate distributed PFC counterpart for the server side and provide an interface to the original PFC on the client. I would call it something slick like dPFC. Perhaps I would de-couple the 30% of the services from PFC and use them in dPFC. Any additions would have to be very loosely coupled; in fact I would use a modular architecture. I would take the existing PFC and put it on the client.

Business object is a new buzz-word. The business object would be at the base of my distributed framework. This is where the bulk of application logic would reside. The business objects should be able to function inside a transaction server or on their own.

continued on ...