Posted 15 April 2005 - 03:01 PM
And to think that I missed out on all of this. Neil Mellis must be seething.
VIP is lovely, Rob's IDE is lovely. For that matter, @MODX is lovely if only because it still works.
Java is better (and so is .Net, before the flame-wars start) because it really is, well, better, now.
Get out, while the getting's good. PRO-IV's core is old, it's time to get you and your business away from it, just like it was time to run away from COBOL and/or assembler (for most businesses) 5-10 years ago. This is not a change from the arguments around SuperLayer, when it came out (it was the salvation of PRO-IV, lest we forget) or Studio (it was the salvation of PRO-IV, lest we forget), or VIP (it was the salvation of PRO-IV, lest we forget), or Rob's IDE (it will be the salvation of PRO-IV, lest we forget). The problem, in the end, is not about a development tool, it's about a language that never quite could manage to grow up. Now we all have lots of investment in it, some of us have spent 20+ years of our working lives in PRO-IV, but it really is time to start looking at the alternatives.
Large(ish) companies like the one I work for have enormous capital tied up in PRO-IV, but the cost of a shift is (in my opinion) less onerous than it once was - and the benefits can be huge! No more legacy database (we converted from PRO-ISAM), no more confusion of presentation logic with business logic, no more reliance on piss-poor implementation of SQL (I want type 2 SQL, why won't the kernel just bloody handle it???) Want to use distributed applications architecture (SOA, ESB, WS-* if you want acronyms) - go ahead! But not with PRO-IV (not really - you can - but it's a band-aid approach at best).
I've spent some of my time in the last 2 years building J2EE applications and it's becoming remarkable what's actually possible using some of the Open Source frameworks out there - like PRO-IV's abstraction of the data access (file definitions do it for you)? Try Hibernate. Can't be bothered doing transaction handling code (PRO-IV does it implicitly) - it's all there in Spring. Want to enforce relational integrity in your lovely RDBMS? Hibernate again. Want a real cross-platform UI design tool? Guess what, it aint in PRO-IV. But it is available inside Eclipse. Which is, oh yes, free.
PRO-IV's value proposition, when I started, was that (in the final analysis) you could write a simple data access/update screen in about 10 minutes, and it just worked. That was unique to the 4GL concept - no buggering about with data access, service layers etc. But the simple fact is that PRO-IV's been largely overtaken by what's available in Java and/or .Net, and Java and .Net (and probably many other things) do the re-usable component thing (viz global screens, updates, etc) simpler, better, faster these days. I can write a web page to display a list of something or other in a nicely formatted table, and link to a maintenance page, and have all of it managed transaction-wise, in not much more than the time that it takes to do the screen layout - that is those 10 minutes that PRO-IV used to do it in - in Java, using *free* software components. And I can then connect it all up to consume an external web service, and other things that PRO-IV can't really do natively, and have transactions managed properly, and security, and it's all version controlled through CVS or Subversion, and its compatible with release management tools, and, and, and...
So PRO-IV's lot will, probably, be to continue down the path of recurrent iterations of whatever the new tool is for development, and inevitably, whatever is next after VIP will be the saviour of PRO-IV developers ererywhere, and we'll all argue about whether it's any good, any why it ignores various principles of good UI design, and why it insists on being windows-only, and so on and so on. Rob's IDE will go the same way. And that's not to say that the IDE is bad, or VIP is bad, but that, in the end, the underlying platform is *fundamentally* flawed, because it can't escape its basic limitations.
I suppose that in the end my point is that PRO-IV's edge in 'RAD' is now seriously eroded - I'd suggest that it's now *more* expensive to maintain a well-built application in PRO-IV than it would be a well-built application in Java, and that means that you all have suddenly got a business case for moving apps to another platform (if you like .Net better than Java I'm not going to argue). That business case may be a multi-million spend, but it's time to take a serious look, because otherwise someone else just might.