| From 1995 to 2002 I led biz dev for DataChannel's XML parser [ the first written, most widely licensed to other platforms eg MSFT, and likely most widely adopted ]. I coordinated our participation in XML & all the related web services standards bodies. We got to participate in a lot of amazing projects, whether relatively simple open standards for DTDs that made it easy to interchange crisply [ eg MathML, many others ], or for industries with considerable interchange challenges [ eg SWIFT ]. These projects were typically fairly lightweight, and fortunately many simply 'worked' and succeeded. In this capacity XML seemed to "just work" and seemed to make the world a better place -- "victory!" Of course you can't really discuss XML without discussing "Web Services," which is what many consider the "xml ecosystem."
Around 1996 Norbert Mikula, Mike Dierken, John Tigue, myself and probably a few other characters began riffing on various ideas of how XML + HTTP could be used. What if you could simply lookup a signed DTD for how your data was suppose to be delivered to you? Lookup a signed DTD for how to invoke its API / RPC calls? Hell, why not have a DNS-like directory of what services are available, whether on your intranet, or from vendors? That's where the original 'whitepages / yellowpages / greenpages' naming convention in the very earliest days of web services came from btw. Naive? Over-optimistic? Absolutely... But we didn't know any better, so why not. We started discussing it with lots of smart people in SGML and XML space, the concepts started turning into prototypes and it started to build momentum. People who were interested fairly wisely said "a standards body should lead this stuff [ instead of some vendor ]," so the early work landed inside OASIS [ the original SGML standards gangsters ]. Their wisdom gleened from long years building incredibly complex SGML doc systems spared the early XML services from many potential debacles btw. Momentum began to build, and super interesting things started being built. As Fortune500 IT, startups, middleware vendors and app servers became increasingly interested in tapping in this magic a very interesting dynamic changed: the platform players perceived all this interop, open standards as a serious competitive threat. Microsoft and IBM in particular got very involved very quickly, and the once simple & elegant concepts quickly devolved into multiple competing standards that were a horrible, illogical, impractical mess. This happened in less than two years... Open-standards based web services essentially became an irrational choice vs just "staying with your existing vendor's stack..." That glimmer of amazing potential magic was quickly and effectively killed. Am I saying XML & web services were perfect? of course not. I am saying a thread of amazing potential was pulled, then around 1999 it was cut. And it was sad to see first hand. I hope GraphQL, JSON, the incredible variety of cloud services and other interesting bits tap into that same magic. Blockchain... imho interesting, but I fail to see many use cases where business side buyers [ aka budgets ] must have a blockchain-based solution. Looks like tech still searching for need and product/mrkt fit to me... think long-term it offers many areas an interesting evolutionary step, but I don't see it as hugely revolutionary gamechanging stuff. Thanks for the walk down memory lane... |