>For instance, in serious disciplines such as Mathematics, there is a great focus on scholarship, and building up on predecessors. In computing, we find it hard

>For instance, in serious disciplines such as Mathematics, there is a great focus on scholarship, and building up on predecessors. In computing, we find it hard to come up with a definition of "progress". There is constant "change", sometimes what looks like "improvements" and "corrections", but it is hard to tell whether stuff is really getting better. At best in software, we find "extensions" of concepts/systems. Contrast with mathematics, where there is a clear mounting to higher concepts over time. This is absolutely true. One real example of this phenomenon in my opinion: the rise of JSON. In the 90's, the XML was invented to be the silver bullet of the myriad of ad-hoc formats for structuring documents that existed in the wild. So in the XML stack you could find the XSL to ensure the validity of an XML document, the XPath to be a querying language, the XSLT to be a language for writing transformations, etc. I guess that in the 90's many people regarded XML as the dawn of an era of strict correcteness between data exchanges and the elevation of the contract-based design. At some point though people deemed XML too verbose and started using JSON. Now with JSON there are multiples conflicting implementations of basic things like a schema mechanism and files and API's that rely many times on natural language to be defined, so an obvious retrocess only because of "verbosity". You can see this random walk in many other places, with frameworks reinventing the same thing over and over and getting traction besides implementing something worse. I guess that since technology is a field that receives a big influx of people, the dynamic of the field changes to support people who don't understand completely what they are doing, this doesn't happen with mathematics/chemistry/physics because it's you who have to adapt to them.