Software still isn't very smart today despite decades of innovation effort and the reason is schemas. Schemaless software is a new alternative that provides an easy way to produce software that’s far smarter than the software we routinely use today. This isn’t science fiction superintelligent software, it’s much like the software we use today but without the smartness limits of the schema-structured information modeling done today. Lifting the limitations imposed by schemas can increase the smartness of software by a factor of ten routinely with the possibility of an occasional hundredfold improvement.
Schemaless software may seem to be an oxymoron like “dry water” because nearly all software is designed with schemas today but in the early days of digital computing all software was schemaless. When schemas came along in the 1960s most programmers found them unnatural because the freestyle programming methods then used were far more flexible than the rigid formats imposed by schemas. The freestyle structures of early schemaless programming were chosen to fit the natural nomological structure of the information being modeled and this intuitive approach was ideal in the days before modern software engineering when programming was done by subject matter experts.
Why is schemaless software smarter? It’s smarter because it supports a far greater range of practical meaning structures and these structures are the ultimate source of all software smarts. These are smarts that comes from far greater editorial range rather than exotic algorithms and it means that practical software can finally be just as smart as the information modeling we do with practical documents.
Does anyone really want smarter software? We’ve all gotten used to today’s simplistic software and we’ve developed survival skills for dealing with its many limitations but smarter software has the potential to provide far greater practical value that the software we routinely use today. This new value includes far greater ease of use as well as substantially greater practical sophistication. Every IT manager knows that there’s always lots of demand for enterprise software solutions that can’t be built due to unmanageable complexity and many of these are good candidates for schemaless software solutions. In just the same way there’s long been lots of latent demand for sophisticated consumer software products that can now be provided as schemaless solutions.
Schemas got their start in the age of data processing as a way to build on the legacy of punch card data modeling but schemaless is ideal for the Internet age where the grand challenge is building smart syndicated information systems at Internet scale. Today many still believe that schemaless is a digital dinosaur from the early days of digital computing but technological tradeoffs change over time and now it’s clear that schemaless set for a big comeback.
No comments:
Post a Comment