Re: Designing a Linked Data developer experience

On 30/12/2018 08:50, Ivan Herman wrote:

First of all, I do not fully agree with:

The reason we have a DOM API is to have a standard way to expose the browser’s DOM to scripts.

This is obviously an important use case, but it goes beyond that. The DOM API is to handle any kind of XML data and, to the point, from any type of programming language. I did a bunch of programming in Python at some point (e.g., an RDFa extractor, but also things that were not HTML related) and the fact that there is a standard DOM API specification implemented in Python libraries meant that I could easily handle the underlying data was a huge win. There are other types of libraries providing access to XML data in a more 'python specific' way, but the DOM API meant an easy move from one environment to the other. These standards also mean the portability of programmers and not only the portability of programs, which is also important:-)

Yes, that's the point I was getting at.

I did look, at some point, at libraries for RDF in different languages (JS, Python, etc) and the fact that each of them take a different approach on how to 'model', in terms of a programming language, the access and handling of RDF data becomes a drag. I "grew" into handling RDF data via Python's RDFLib, and it was a royal pain when I had to do something in Javascript where there are a number of libraries that may not be similar to the way RDFLib handles things. The 'portability of programmers' simply breaks when it comes to programmatically handle RDF data.

This is the sort of document that I had in mind. Having a W3C Recommendation provides a level of stability which transcends the ever-changing view of what is an effective web programming environment.  It's a pity the group ran out of steam.


Richard Light

Received on Sunday, 30 December 2018 13:23:00 UTC