\
**Category:** General\
**Created:** [October 19, 2019, 9:51pm UTC](https://swi-prolog.discourse.group/t/html-1-from-html-write-quoting-of-html-tag-attributes/1407 "2019-10-19T21:51:01Z")\
**Posts on this page:** 3\
**Page:** 1
**Author:**  [@mgondan](https://swi-prolog.discourse.group/u/mgondan)\
**Post date:** [October 19, 2019, 9:51pm UTC](https://swi-prolog.discourse.group/t/html-1-from-html-write-quoting-of-html-tag-attributes/1407/1 "2019-10-19T21:51:02Z")
Dear developers,
I noted that html//1 always wants to quote the tag attributes:
Example (with & replaced by $):
> html(math(mfenced(separators(‘$nbsp;’), [mtext(a), mtext(b)])), M, ), print\_html(M).
Result is (with spaces inserted into the tags to make them visible, and & replaced by $)
> \< math\>
> \< mfenced separators=“$amp;nbsp;”\>
> \< mtext\>a\< /mtext\>
> \< mtext\>b\< /mtext\>
> \< /mfenced\>
> \< /math\>
This is because the html attributes always undergo quoting in html//1. I would like to
suppress it using the backslash (as in the content of html tags), but this does not work for the attributes,
> html(math(mfenced(separators(\‘$nbsp;’), [mtext(a), mtext(a)])))
raises an error that an atom is expected. I could fix the problem by inserting the following line into library/http/html\_write.pl (around line 830)
> html\_quoted\_attribute(\Text) → !, [Text]. % inserted
>
> html\_quoted\_attribute(Text) →
> { xml\_quote\_attribute(Text, Quoted, utf8) },
> [Quoted].
I don’t think this breaks existing code, or will it? You might wish to integrate this into the code. Thank you for your consideration.
Best wishes,
Matthias
---
**Author:**  [@jan](https://swi-prolog.discourse.group/u/jan)\
**Post date:** [October 20, 2019, 2:02am UTC](https://swi-prolog.discourse.group/t/html-1-from-html-write-quoting-of-html-tag-attributes/1407/2 "2019-10-20T02:02:02Z")
The intend is to use `&(nbsp)` and have html//1 emit the right term. Why doesn’t that work?
---
**Author:**  [@mgondan](https://swi-prolog.discourse.group/u/mgondan)\
**Post date:** [October 20, 2019, 8:13am UTC](https://swi-prolog.discourse.group/t/html-1-from-html-write-quoting-of-html-tag-attributes/1407/3 "2019-10-20T08:13:40Z")
Dear Jan,
I was referring to the tag attributes, not to its “content”. Indeed, as you write, quoting can be suppressed for the content, so this works:
> html(math(mfenced(separators(‘’), [mtext(a), \[‘ ’], mtext(b)])), M, ), print\_html(M).
and I think your suggestion refers to this:
> html(math(mfenced(separators(‘’), [mtext(a), ‘&(nbsp)’, mtext(b)])), M, ), print\_html(M).
I was actually referring to the “separator” (i.e., the tag attribute). But this one does not work, because the ampersand is quoted:
> html(math(mfenced(separators(‘ ’), [mtext(a), mtext(b)])), M, ), print\_html(M).
> html(math(mfenced(separators(‘&(nbsp)’), [mtext(a), mtext(b)])), M, ), print\_html(M).
Both result in html output like
> \\\a\\b\\\
And in the attribute it seems not possible to suppress the quote, so this does not work:
> html(math(mfenced(separators(\[‘ ’]), [mtext(a), mtext(b)])), M, ), print\_html(M). error message html(math(mfenced(separators(\\'\ '), [mtext(a), mtext(b)])), M, []), print\_html(M). error message
I see that my use case is somewhat special. But I was wondering if quoting of tag attributes is a desirable feature at all.
Thank you for your consideration.
Best wishes,
Matthias