On 19 May 2011, at 14:03, Alan Ruttenberg wrote: >> "Current tools", i.e. the ecosystem in which we expect broad adoption of SW >> technology, is not just a temporary, minor obstacle; >> it is likely THE most important aspect our technical proposals must fit to. > > This view is incredibly short sighted, and shows little awareness of > the actual historical evolution of technological innovation. Are we > still using Mosaic to browse the web? Do you not think the development > of the many cms for web, or the build your own site from template > tools offered by providers had nothing to do with broad scale > adoption of HTML.
I think you're missing Martin's point, Alan. HTML was initially designed so that it could be easily produced with "current tools", which in the beginning meant, "text editors". Now, HTML's designer could have said, "let's not be so short-sighted, eventually people will use CMSes and visual editors, they won't see the markup any more, so let's make the language more compact and efficient by not making the markup itself human-readable." This is a reasonable argument from an engineering point of view, especially if one takes a long-term perspective. However, it is quite likely that HTML would have never taken off if that design had been chosen. That's because every fundamentally new technology first has to win the hearts of the early adopters (which almost always means, techies). They don't have tool support yet. They don't have the assurance that this random new technology is going to stay around for the long-term. All they have is yet another new unproven technology vying for attention. The harder it is to get going with it, the less early adopters will actually start using it, and the slower uptake will be. You have to win over the early adopters, because they are the ones who are going to build the nice tools that will finally get the majority on board. This is why new technologies have to slot right into existing ecosystem *and* have to come in a form that works for early adopters. (The story is a bit different if you have technology giants pushing new technology to a locked-in market.) Examples of this are endless; much of the history of the Internet can be seen as a struggle between OSI's efficient, well-designed ASN.1 based stack of binary protocols (e.g., X.400) and the worse-is-better text-based IETF stack of protocols (e.g, SMTP, HTTP). Closer to the present, I'm sure you've come across more than one example of ASN.1 based or other binary formats being supplanted by more recent XML alternatives. People are often bewildered how really well-designed solutions fail while clearly inferior solutions take the market by storm. There can be many reasons for this, but a failure to appreciate the importance of simplicity in technology adoption is certainly one of them. Best, Richard
