FrameWorks Before and after PFC
Almost twenty years ago, before PFC, user group meetings were mostly totally dedicated to frameworks. How deep or shallow… Inheritance, encapsulation and polymorphism. Who had the smartest string functions or ability to do file system operations better than anyone. I remember getting a buzz out of playing a wavefile using soundA. It was all very technical. On projects frameworks were an afterthought for the techos while waiting for the final final user specs or while we were RADing. Who needs specs when you can write a quick and dirty actual system in half the time.
Every new project seemed to have a Lead who managed to sell or donate their framework enhanced since adopted from a previous project and which would get it all going and programmers had to learn it and adhere by its’ inherent standards, whilst forgetting Cobol and worse. PB was better than VB and SQL Windows came and went. Datawindows above all conquered the up and coming client server market running on windows 3.1.1 for work groups.
My point here is. 15-20 years ago PB was it. New project = PB. In 1998 we had 16+ people attending FastTrack to PowerBuilder 5 and 6.5 every month in Sydney, Australia alone.
PFC set the foundation of all new projects in the mid to late nineties. Frame, sheets, menus, tabs and 3D sunken five deep. Game, set and match. By late ’96 I was a lead on a project and found myself wanting to write my own, cos that’s way more fun, but hearing myself praising PFC. Got my kicks from writing new window services. Sticky Notes was one.
By 2000 PB had gone from best C/S tool with added PFC to 3tier Jaguar with… well you really need a different sort of framework to pool your poor stateless components. Not many could justify a Jaguar and true to form, Sybase struggled selling it for HOW MUCH?! We had gone from high development license cost and free for all deployment to a different model all together.
After PFC… well PFC is now embedded in a few “Legacy” systems. If you wanted something else you couldn’t justify it. Metaphorically speaking the PFC foundation is as heavy as solid concrete and all the utilities heavily embedded. At least your house won’t move in a hurry. Might also keep a few of us enjoyed till retirement in not so many moons.
So is there an after PFC. Is there something to replace it. Are there Tech Leads and Architects on projects who can take the user requirements and who can justify PB Apps at all? Apps!! have we gone full circle? What are Apps on Windows going to look like next year? Could a modern PB framework take control away from terrible awful browser applications where you have to enter a simple date from 3 dropdown. A drag when you are born more than half a century ago. Who will design it? What are the real requirements of a modern generic app?
To me, the ideal framework would enable you to produce a maintainable core business application app in a very short time frame.
I don’t know what else can snatch PB from the jaws of the slow attrition to oblivion.
Hi Lars,
Thanks for sharing your experience.
As a framework guy myself, I'm interested but have to confess that I don't understand some sentences. I still have to improve my English.
Can you explain "Got my kicks from writing new window services. Sticky Notes was one." ?
In your conclusion are you criticizing PFC, calling for enhancements or for a fresh new framework ?
We have our own framework, a legacy one, and sometimes I think that it would have be better to have PFC, but I also like to control it all.
And sometimes I just think that too much framework kills the framework, when optimizing performance especially.
Have you ever experienced other existing frameworks, like Chris Pollach's STD, or Kodigo ? I'm always thinking that I have to look at them, but did not yet. It could be useful to have a review of the different options in this site.
Regards
Guillaume
Hi Guillaume,
I was referring to enhancements or extensions to the PFC I had fun doing on a project 16 years ago. It was a new window service in the context of PFC where functionality is delegated to a service object being a descendant of n_cst_winsrv. (Not a windows NT service or anything). From memory is was the ability to attach a sticky note to any window by invoking a hand-written n_cst_winsrv_sticky service. Actually, it was probably a datawindow service.
I have worked with other frameworks most of them not memorable, but not STD or Kodigo. Should probably have a sticky beak.
Your English is plenty pretty good I am sure. You are not the first to not comprehend my odd sentences. Perhaps a reflection of my warped sense of humour. I tend to forget or perhaps ignore that I am writing to a wider varied audience.
Not sure what I am suggesting. It is important to learn from what has been built and to be prepared to throw it away and start a fresh. Have you ever had the thought. "If I did this again, I would do it differently".
Lars
Thanks for the clarification.
I like to throw and start fresh, like you said, I kept that from my Lego period.
I searched for "Sticky Notes" on google and found a desktop application, I was wondering how that could have been done with PB ?! 🙂