State of the Art in Web Development

Five years ago, I wrote for Internet World magazine a feature story on the State of the Art in Web Development. I checked back on that article with some trepidation, because the one thing constant in IT commentary is how embarrassingly wrong one can be – especially over time.

What I Got Write ..uhh Right

What I was trying to emphasize in this article was the emergence of the Web desktop not only as the GUI of choice but also as the new development paradigm of choice. The comparison was with classic internal IT development projects that were still using pre-Y2K mode of operation – monolithic 1-3 years, waterfall methods with often commendable goals such as large scale application integration and legacy application reuse. In contrast Web projects, tended to emphasize Agile methods with their test-first, iterative, and feature-driven cycles and with working program deliverables as their model.

But there were other underlying trends that I tended to try to forsee. Outsourcing plus Service provision through a battery of gradually more complete offsite or packaged programs. But I did not have a or the battery of new Service Web-based Service providers including a belated, rushing-to-market Microsoft Revived – Alive .. uhh Live. But that is another story of how Redmond will try yet again to turn what it considers its greatest enemy, the Web Interface, inward and onto Vista and other Windows proprietary services.

However, I certainly managed to get the basic trend to Web based GUI delivery right as the new crop of portal-based applications and SaaS -Software as a Service truly live on the Web GUI. But I missed the full flowering of RIA_Rich Internet Application technologies that would attempt to deliver the 6As of Presentation Layer programming. Yes, I allude to the needs and requirements; but no I did not see the emergence of Flash particularly in its Open Source format (check out Laszlo and others), the resurrections of both Java and JavaScript (especially in its AJAX incarnation), and the huge problems of meeting the cross platform (especially down to PDAs, mobiles and embeddeds), online and offline requirements and what major stumbling blocks they would present.

Also I caught a strong whiff of the middleware and ERP as model for all application building. Let the pros build the CRM, SCM, EAI, and in general most enterprise applications – then buy the one with the best fit plus integration, personalization and customization features. Also I missed out on the importance of SOAP/WSDL/XML Web Services in the guise of SOA-Service Oriented Architectures and ESB-Enterprise Service Bus as being the leading proposed means of application integration. Given the security problems and complexity of some the latest Web Service APIs(see Oasis and WSI for examples), I am still not certain that SOA+ESB+Web Services will become the dominant development framework. There are some nasty long-duration transaction and persistence layer problems and related middleware and messaging layer difficulties that keeps the problems up in the air.

But all in all, I did lead people in the right directions particularly on the nature of development processes – no small achievement. Yes – its hard reaching all the way back to “pat, pat”.

(c)JBSurveyer 2006