I'd like to suggest a few things:
- Turn this into a group EDTF reference project.
- Use C++ (using SWIG can easily provide services to a number of languages)
- Instead of GPL a BSD license.
- Build as a class library for DateTime. Include also comparison etc.
This is, I think, particularly important in a number of applications.
Using ==, for example, to mean congruence.
1986 == 1986-01-02 etc.
- Problem with natural language is that the normalization are often non-
trivial. 1960-talet in Swedish might be 1960s in English but "20 Jh" in
German is the same as "20th Century" in English but the "19-Talet" in Swedish
and 19 in EDTF.
Pre-requisite I think will be a consensus on expressions. A number of
constructs that we have created I think will be quite difficult to express in
plain language. There are a number of cultural issues that will arise.
- Once we have this it would be natural to also want to have a parser to turn
these expressions into normalized EDTF expressions.
EDTF ---> internal representation --> Plain Language
\--> normalized EDTF expressions
Plain Language Expressions --> EDTF
On Tue, 26 Apr 2016 15:53:47 +0200, [UTF-8?]SaaÅ¡ha Metsärantala wrote
> > an EDTF - Human readable library.
> I plan to implement a GNU-licensed such library as soon as the new standard
is ready, but it will be a multilingual one, that is: You will not only input
EDTF, but also which human language you want the output to be in. For example
> Input: 196x and sv
> Output: 1960-talet
> for Swedish,
> Input: 196x and fi
> Output: 1960-luku
> for Finnish and
> Input: 1910~ and en
> Output: c1910
> for english etc. The choice of available languages will depend on the
language knowledge of interested persons willing to help.
Edward C. Zimmermann, NONMONOTONIC LAB