On Sun, Oct 11, 2009 at 7:17 PM, Lyle <[email protected]> wrote: > Hi All, > According to 'Software Engineering 8' (Sommerville, 2007, p13) the 3 > key challenges facing software engineering are:- > 1. The heterogeneity challenge > 2. The delivery challenge > 3. The trust challenge > > Doesn't Perl 6 and parrot completely smash the 'heterogeneity > challenge'? Even legacy languages can be put on parrot and easily pulled > into Perl 6 and other lanuages, in theory... > > I thought this might spark some discussion... > > > Lyle > > Heterogeneity description > http://en.wikipedia.org/wiki/Heterogeneity > > Sommerville, Software Engineering 8, 2007 >
It was hard to reply without a bit more context, but based on a little searching I think we're talking about heterogenous networked systems. This description popped to the top of my googling: "Computers are pervasive in modern life. When users perform transactions, they usually do not limit their access to their own computers. Instead, the users tend to access systems distributed globally across networks with heterogeneous technology and support systems. Further, legacy systems integrate with modern systems creating new challenges in building systems that are flexible. Heterogeneity challenge is the challenge involved in developing flexible systems that work well in different technological background." Based on that.. I don't think parrot is hugely relevant. I guess one way of approaching a legacy system integration would be to try and get the legacy platform running on parrot, but it seems like a bit of a stretch. More relevant would be simple, widely accepted standards for increasingly high level protocols. Perl is of course pretty well suited to wrapping up a legacy system and providing a modern interface to it, via your choice of web service protocol. Nobody mention SOAP! It's interesting that the above question should mention transactions - I see there are some web service standards for handling distributed transactions ( http://www.ibm.com/developerworks/library/specification/ws-tx/ ), hadn't come across those before. They look like they may be tied into SOAPland, anyone tried using them? I guess we've seen proprietary interop protocols emerge and fall out of favour (COM, CORBA), being replaced by open standards for web services based on web technologies (SOAP and co), which have in turn been superceded by simpler, more human-friendly versions. The focus of effort has shifted somewhat from the basics of communication towards concerns like authentication, authorisation and data definition. We also have increasing availability of and commercial interest in networked platforms for parallel processing. It seems obvious that as good distributed solutions appear at each level, adoption will shift to the next level up - transactions, security models, user preferences, application configuration, content storage and retrieval, content indexing and searching, who knows. Perhaps before too long it will be normal to design applications to be composed of major independent components running on different platforms. Alex _______________________________________________ BristolBathPM mailing list [email protected] http://mailman.bristolbath.org/mailman/listinfo/bristolbathpm
