Michael Kay
Michael Kay
>It seems to me that when an expression contains transient() the result may differ, depending on the chosen order of evaluation. Absolutely. That's why this isn't the default. >To summarize,...
Note my 2020 Balisage paper on asynchronous XSLT: https://www.balisage.net/Proceedings/vol25/html/Kay01/BalisageVol25-Kay01.html Four years later, we're getting close to releasing a product that includes some of these ideas. I do think that asynchrony...
The issues of determinism (or functional purity), referential transparency, and asynchrony relate to each other, but I think they are are separable concepts. I think there are three levels of...
There are very few cases where we made a conscious decision in Saxon to not conform with the spec, and this was one of them. We decided (especially for `fn:collection()`)...
I don't feel 100% comfortable with this: particularly for maps. It feels wrong that `map:contains(M, X)` is true if the map contains the key X, while map:contents(M) returns the "values"....
Following the discussion yesterday, I'd like to suggest for arrays: ``` array:to-sequence() ``` I think this expresses what the function does fairly clearly and intuitively, and signals that with a...
Closed by #1197
I propose to close this issue without action; I think the current spec is acceptable.
A number of the examples for csv-to-xml use CRLF delimiters, and the tests for these examples have been failing. For the time being I am fixing the tests by adding...
Fix #296