Do languages evolve to minimise portability?
Governments promote standards because following them helps their citizens save money. The UK and US have contrasting positions, with the UK focusing on savings achieved through the repeated use of standardized items and the US focusing on the repeated use of skills people acquired through using a standardized item.
What benefit, if any, do product vendors receive from adhering to a standard? A few vendors may be in a position to use economies of scale to gain market share (by driving out of business less well funded competitors), other vendors follow a specified standard as a means of winning a contract with large customers, and the rest don’t care.
In the program language world, for 20 odd years, there were just the Cobol and Fortran standards. These started out idealistically driven because hardware vendors had to convince potential customers that it was commercially viable for them to write software for the products they were selling (i.e., the early computers). Once business started to take off the computer industry’s compete/cooperate dynamic, mixed with large egos wanting their own way and the sales driven mandate to keep existing customers happy (i.e., don’t break existing code) took over. There was lots of infighting and progress was slow, but at least there were lots of people who thought it important enough to be involved.
In the mid-1980s there was a seismic shift in approach to language standards with Pascal and C becoming ISO standards, the first created by a group of mainly academics and the second by a group containing many small company consultants/vendors. The age of ‘activist’ language ISO standard creation had begun; which is not to say that anything changed in the Cobol and Fortran standards’ world. For a look back at this decade see: Computer Standards & Interfaces, volume 16, numbers 5 & 6, Sep 1994.
If you were to put somebody who knew nothing about computer languages in a room with the standard’s committees for Modula-2 and C++, I don’t think they would be able to tell them apart. Inside the bubble its not possible to distinguish the language that died before its standard was published and the language that continues to grow like Topsy.
Activists want to get things done and will select the route of minimal bureaucratic resistance. These days this means that new language standards are often created locally and fed into the ISO process higher once they are done.
Open source compilers have solved many of the source portability problems caused by language dialects, that standards were intended to solve, by dramatically shrinking the market of the compilers that supported those dialects.
Where has the user been in all this? Well, where-ever they are, they are rarely seen at language standard meetings.
These days portability problems are caused by the multitude of languages, not by the multitude of dialects of a language.