French SI Brochure feedback from BIPM (pages 21-40)

[metanorma]

French SI Brochure feedback from BIPM (pages 21-40)

[Intelligent2013]

page	comment	to fix
21	first equation: wrong grouping of decimals after the comma: 6,022 140 76.	adoc
21	second paragraph after 1. equation: "M" needs always to be initalics	adoc
21	2. equation: the 12 in 12C needs to be a superscript, not a subscript; C not in italics	adoc
21	below 3. equation: an opening parenthesis is missing in M(12C)	adoc
22	definition of candela: 2nd last line it should read m-2,not m2; Negative and positive exponents are not at the same height	adoc, jEuclid?
22	2. equation: replace decimal point by comma; insert a space between h and Kcd (2 x)	adoc
23	3. paragraph: replace "Article 5.4.7" with "section 5.4.7" and analogous at several other places	rb
24	first paragraph of 2.3.4: isolated » on new line	adoc
24	Table 4: angle solide: m2 not in italics	adoc
24	Table 4: 3re column vertically not well aligned with 1st and 2nd column	xslt
24	Table 4, footnote b: pi not in italics	adoc
27	Table 6: energie volumique: last column: kg m-1 s-2; there is a misprint in the SI brochure	adoc
27	Table 6: increase width of first two columns so that there are less linebreaks	?
28	4. paragraph: pi not in italics	adoc
28	4. paragraph : isolated » on new line	adoc
28	5. paragraph : isolated » on new line	adoc
30	side note: accents missing on four prefix names: mébi, tébi, zébi,pébi	adoc
30	μ not in italics	adoc
31	last line : μ not in italics	adoc
32	Table 8: pi not in italics	adoc
32	Table 8 footnot b: μ not in italics	adoc
33	1. paragraph: 3rd line from bottom: m in italics	adoc
34	3. paragraph: split in two paragraphs before "Les symboles…"	adoc
35	section 5.4.1, 2. paragraph: "livre rouge" not in italics ; isolated » on new line	adoc
36	Figure: insert ticks for the axis labels	image
36	sectio 5.4.2, last side note: spaces not correct, compare with original	adoc
37	2. line: reference to page number 149 wrong, should be 140	adoc
37	5.4.4: isolated » on new line	adoc
38	5.4.7, 2. paragraph: μ not in italics	adoc
38	5.4.7, 3. paragraph: isolate « at end of line	adoc
38	last side note,last line: space missing between" confusion" and "supplémentaire"	adoc
39	5.4.8: pi not in italics	adoc
39	4. paragraph: rayon r, r in italics; omega in italics	adoc
40	Full name of CIPM and CGPM instead of acronym	adoc
40	1. paragraph: "procès verbaux" in italics	adoc

[opoudjis]

For rb, see #40

[manuelfuenmayor]

21 | first equation: wrong grouping of decimals after the comma: 6,022 140 76. | adoc

I understand that this is suposed to be handled automatically (and not by markup).

21 | second paragraph after 1. equation: "M" needs always to be initalics | adoc

21 | 2. equation: the 12 in 12C needs to be a superscript, not a subscript; C not in italics | adoc

21 | below 3. equation: an opening parenthesis is missing in M(12C) | adoc

All three fixed.

22 | definition of candela: 2nd last line it should read m-2,not m2; Negative and positive exponents are not at the same height | adoc, jEuclid?

This is a rendering issue. Not fixable by markup.

22 | 2. equation: replace decimal point by comma; insert a space between h and Kcd (2 x) | adoc

Fixed

24 | first paragraph of 2.3.4: isolated » on new line | adoc

Ok, I believe that « » are quotation marks, right? So, in order to avoid the issue, I've removed the white space between quotation marks and their adjacent words.

24 | Table 4: angle solide: m2 not in italics | adoc

Fixed.

24 | Table 4, footnote b: pi not in italics | adoc

This is not fixable by markup. @opoudjis?

27 | Table 6: energie volumique: last column: kg m-1 s-2; there is a misprint in the SI brochure | adoc

This a rendering isssue. Not fixable by markup.

27 | Table 6: increase width of first two columns so that there are less linebreaks | ?

I can change attributes of the table in order to indicate that certain columns must be wider than others (e.g. [cols="2,2,1]). But this seems not to have any influence in the output.

28 | 4. paragraph: pi not in italics | adoc

Not fixable by markup

28 | 4. paragraph : isolated » on new line | adoc

28 | 5. paragraph : isolated » on new line | adoc

White space removed. (See comment above.)

30 | side note: accents missing on four prefix names: mébi, tébi, zébi,pébi | adoc

Fixed.

30 | μ not in italics | adoc

31 | last line : μ not in italics | adoc

32 | Table 8: pi not in italics | adoc

32 | Table 8 footnot b: μ not in italics | adoc

Not fixable by markup (unless I pasted the literal symbol). @ronaldtse ?

33 | 1. paragraph: 3rd line from bottom: m in italics | adoc

34 | 3. paragraph: split in two paragraphs before "Les symboles…" | adoc

35 | section 5.4.1, 2. paragraph: "livre rouge" not in italics ; isolated » on new line | adoc

All three fixed.

36 | Figure: insert ticks for the axis labels | image

I changed the image file.

36 | sectio 5.4.2, last side note: spaces not correct, compare with original | adoc

Fixed. Line breaks were inserted manually.

37 | 2. line: reference to page number 149 wrong, should be 140 | adoc

Not sure what to do here. I changed "149" by "140", but should I put a cross-reference here? If so, this cross reference must link to an external document (the English version of Brochure). @opoudjis?

37 | 5.4.4: isolated » on new line | adoc

White space removed.

38 | 5.4.7, 2. paragraph: μ not in italics | adoc

Not fixable by markup (unless I pasted the literal symbol).

38 | 5.4.7, 3. paragraph: isolate « at end of line | adoc

White space removed.

38 | last side note,last line: space missing between" confusion" and "supplémentaire" | adoc

Fixed.

39 | 5.4.8: pi not in italics | adoc

Not fixable by markup (unless I pasted the literal symbol).

39 | 4. paragraph: rayon r, r in italics; omega in italics | adoc

Fixed "rayon r" part. But "omega in italics" is not fixable by markup.

40 | Full name of CIPM and CGPM instead of acronym | adoc

40 | 1. paragraph: "procès verbaux" in italics | adoc

All two fixed.

[manuelfuenmayor]

The fixes are in PR #45

[Intelligent2013]

Regarding isolated » on new line issues.
I think we should not remove space between double angle quotation and text.

Source renders with spaces:

I think that would be bettet to replace space to non-break space character (hex A0).

[manuelfuenmayor]

Thanks @Intelligent2013! I take your suggestion. Changes applied.

[opoudjis]

@manuel489

37 | 2. line: reference to page number 149 wrong, should be 140 | adoc
Not sure what to do here. I changed "149" by "140", but should I put a cross-reference here? If so, this cross reference must link to an external document (the English version of Brochure). @opoudjis?

That's a nightmare scenario, but we do provide for it under collections. The problem is that this is going to force a massive unwelcome change to how we process crossreferences between documents.

To reference another document in a collection, use:

* [[[english-doc,repo:(current-metanorma-collection/si-brochure-en)]]]

in the bibliography, and

<<english-doc,anchor=id>>

I think this will break. Do it anyway. We will have to deal with it when it does.

[opoudjis]

24 | Table 4, footnote b: pi not in italics | adoc
This is not fixable by markup. @opoudjis?

@Intelligent2013 bringing you in here.

The XML here is:

<stem type="MathML"><math xmlns="http://www.w3.org/1998/Math/MathML"><mn>2</mn><mi>π</mi><mtext> rad</mtext></math></stem>

Now, the MathML spec explicitly says that Greek constants like pi must be marked up as mi, not mn: https://www.w3.org/TR/MathML3/chapter3.html#presm.mi . The problem is, MathJax, and clearly Euclid, are blanket rendering anything in mi as italics, and Greek lowercase letters are not italicised in maths. (Practice varies for capital Greek letters, and MathJax treats it as an amsmath config option: https://tex.stackexchange.com/questions/87238/greek-letters-in-italic-in-math-equation.)

As https://tex.stackexchange.com/a/440589 indicates, this is a local preference:

With a modern toolchain that supports unicode-math, you can select math-style=upright to get all upright letters, and math-style=ISO to get all slanted math letters. Upright numbers and Greek capital letters with all other letters in math mode italicized is the default, but you can select it explicitly with the option math-style=TeX. There is also a fourth math-style=French option. Whatever you select, symbols such as summation and product series remain upright.

And sure enough:

https://tex.stackexchange.com/questions/159973/french-math-style-with-default-font

The package mathastext called with the frenchmath option does the job.

I've realized that French typography uses a different style for mathematics, namely upright uppercase roman letters and lowercase greek letters (roman lowercase remaining in italic), while default LaTeX is italic. What I want is to use this style with the default math font.

This is a rendering option in TeX. We need to find out if this is a rendering option in jEuclid and in MathJax.

This will NOT be addressed by manually injecting italics or CSS in MathML. We might get away with that, just, in Word; we will not get away with that in PDF.

The BIPM HTML, risibly, used GIFs for Greek letters. No, I am not making this up: https://www.bipm.org/fr/CGPM/db/26/1/ . They therefore have no ability whatsoever to provide any guidance for the proper HTML rendering of MathML, and little authority to demand French rendering.

[opoudjis]

Flabbergastingly, the MathJax TeX font, which is the default font MathJax uses, has no support for upright lowercase Greek characters at all; mathjax/MathJax#2123: a non-italic pi is impossible within MathJax even with the "mathvariant" override attribute (which I really do not want to use; cf. mathjax/MathJax#592). We would have to shift to STIX-Web font to display any upright Greek fonts at all.

[opoudjis]

mi/@mathvariant is working in HTML if I change the MathJax font to STIX. @Intelligent2013 is investigating whether it is respected in jEuclid. If it is, I will need to postprocess the MathML to make it behave as expected in French mathematical typesetting. If it is not, our answer has to be that we do not support French mathematical typesetting.

[Intelligent2013]

I've checked mathvariant attribute and it is working in jEuclid, here is resulted PDF:

jEuclid set mathvariant style for mi tag (from jeuclid-core\src\main\java\net\sourceforge\jeuclid\elements\presentation\token\Mi.java) depends on text length:

        if (StringUtil.countDisplayableCharacters(this.getText()) == 1) {
            this.setDefaultMathAttribute(
                    AbstractJEuclidElement.ATTR_MATHVARIANT, "italic");
        } else {
            this.setDefaultMathAttribute(
                    AbstractJEuclidElement.ATTR_MATHVARIANT, "normal");
        }

Regarding https://www.w3.org/TR/MathML3/chapter3.html#presm.mi, the default value of mathvariant depends of length of content:

If we append zero-width space after pi inside mi:

<stem type="MathML"><math xmlns="http://www.w3.org/1998/Math/MathML"><mn>2</mn><mi>π​</mi><mtext> rad</mtext></math></stem>

then we get non-italic pi:

[opoudjis]

Ok. This becomes new ticket: metanorma/metanorma-standoc#367

[Intelligent2013]

Regarding issue:

page	comment
22	... Negative and positive exponents are not at the same height

The reason of misaligment between -2 and -1 is the max height of previous chars - k has a more height than other chars g s r, therefore -1 shows a bit higher.
I've checked how it displays in MathJax (https://www.mathjax.org/#demo), and it shows similar (but with less differences):

I see only one solution - isolating cd sr k from g into separated mtext like this:
<mtext> cd sr k</mtext>

I,.e. from:

<msup><mrow><mtext> cd sr kg</mtext></mrow><mrow><mrow><mo>−</mo><mn>1</mn></mrow></mrow></msup>

to

<mtext> cd sr k</mtext><msup><mrow><mtext>g</mtext></mrow><mrow><mrow><mo>−</mo><mn>1</mn></mrow></mrow></msup>

Can we do it? If so, then we'll get this result:

[opoudjis]

Um...

... That solution does such violence to the MathML, I would much rather we refuse to implement it at all. Whatever @ronaldtse has planned for UnitsML, it will decidedly not permit this kind of rearrangement of text: units will be treated as semantically meaningful elements.

The kg can and likely will be separated from the sr. But the kg being broken up into bits for the sake of rendering is likely not to happen.

At any rate:

https://en.wikipedia.org/wiki/International_System_of_Units

The kelvin is defined by setting the fixed numerical value of the Boltzmann constant k to 1.380649×10⁻²³ J⋅K⁻¹, (J = kg⋅m²⋅s⁻²), given the definition of the kilogram, the metre, and the second.

The superscript on the K is going to be misaligned no matter what you do, because the K is taller than the m. So this solution won't scale even if it were semantically permissible.

[opoudjis]

This could be addressed by adding an msup/@superscriptshift attribute, forcing the same minimum height on all superscripts. But good luck in guessing what that might be.

So this does work:

<math><mrow><mtext>cd sr </mtext><msup superscriptshift="4pt"><mrow><mi>kg</mi></mrow><mrow><mo>−</mo><mn>1</mn></mrow></msup><msup superscriptshift="4pt"><mi mathvariant="normal">m</mi><mrow><mo>−</mo><mn>2</mn></mrow></msup></mrow></math>

But... how do I know what the right point measure for superscripting should be? We really should not be being forced to do this amount of fine tuning of rendering. And w3c/mathml#27 indicates that this functionality will be removed from MathML Core in the future.

... The real issue here, actually, is simple. BIPM until now have been getting Word superscripting as how units are rendered, which means a fixed height for all superscripts. They will simply not be getting that from us: if there is to be any machine readable units, what they will be getting is what MathML gives them. And what MathML gives them is uneven exponents. My recommendation: that they resign themselves to it. msup/@superscriptshift is not something we should be using.

[manuelfuenmayor]

That's a nightmare scenario, but we do provide for it under collections. The problem is that this is going to force a massive unwelcome change to how we process crossreferences between documents.

To reference another document in a collection, use:

* [[[english-doc,repo:(current-metanorma-collection/si-brochure-en)]]]

in the bibliography, and

<<english-doc,anchor=id>>

I think this will break. Do it anyway. We will have to deal with it when it does.

@opoudjis , problem is, there is no bibliography section in original document. I would need to create one. Should I do so?

[Intelligent2013]

Issue:

page	comment
22	... Negative and positive exponents are not at the same height

fixed.

Before:

After:

[opoudjis]

@opoudjis , problem is, there is no bibliography section in original document. I would need to create one. Should I do so?

Do so for now. I will have to implement code to remove it from ultimate rendering, and this is an issue for ISO 10303 as well, but the reference needs to be in there before we can process the reference at all.

[manuelfuenmayor]

@opoudjis , problem is, there is no bibliography section in original document. I would need to create one. Should I do so?

Do so for now. I will have to implement code to remove it from ultimate rendering, and this is an issue for ISO 10303 as well, but the reference needs to be in there before we can process the reference at all.

Changes pushed to master branch directly.

[opoudjis]

metanorma/metanorma#146 will address removing (hiding) those internal bibliographic entries.

[manuelfuenmayor]

Based on the latest PDF provided by @opoudjis: si-brochure-fr(12-01).pdf

These are the issues that don't fulfill with BIPM requirements yet.

page	comment	issue
21	second paragraph after 1. equation: "M" needs always to be in italics	not fixed
22	definition of candela: 2nd last line it should read m-2,not m2; Negative and positive exponents are not at the same height	not fixed
27	Table 6: increase width of first two columns so that there are less linebreaks	not fixed
36	sectio 5.4.2, last side note: spaces not correct, compare with original	fixed in #59
37	2. line: reference to page number 149 wrong, should be 140	not fixed
39	4. paragraph: rayon r, r in italics; omega in italics	omega part not fixed

[Intelligent2013]

22 | definition of candela: 2nd last line it should read m-2,not m2; Negative and positive exponents are not at the same height

Fixed. I've increased a bit the height of exponents.

Before:

After:

27 | Table 6: increase width of first two columns so that there are less linebreaks

I can't change column width for concrete table. Columns width set automatically depends on longest word in cell. For this table:

• 1st column - width 13 (concentration)

• 2nd column - width 10 (kilogramme)

• 3rd column - width 11 (
  )

• 4nd column - width 19 (
  )

As workaround solution you can try to split long stems in last column into two stems.

37 | 2. line: reference to page number 149 wrong, should be 140

This page number shows only in two-languages PDF. In French only it doesn't show.

[opoudjis]

39 | 4. paragraph: rayon r, r in italics; omega in italics | omega part not fixed

See #40 (comment)

21 | second paragraph after 1. equation: "M" needs always to be in italics | not fixed

Metanorma is implementing the French mathematical typesetting rule, that uppercase variables are upright. Where that is ignored, you need to introduce explicit font styling: stem:[ii(M)_{rm(u)}].

Note that this won't work until metanorma/metanorma-standoc#386 is merged in; it isn't yet, because of a bug somewhere else in the stack.

[manuelfuenmayor]

One remaining issue in this thread:

page	comment	issue
21	second paragraph after 1. equation: "M" needs always to be in italics	fixed
22	definition of candela: 2nd last line it should read m-2,not m2; Negative and positive exponents are not at the same height	fixed, #41 (comment)
27	Table 6: increase width of first two columns so that there are less linebreaks	not fixed
36	sectio 5.4.2, last side note: spaces not correct, compare with original	fixed
37	2. line: reference to page number 149 wrong, should be 140	will not fix, #41 (comment)
39	4. paragraph: rayon r, r in italics; omega in italics	fixed

[manuelfuenmayor]

27 | Table 6: increase width of first two columns so that there are less linebreaks | not fixed

@opoudjis it seems that is not possible to increase the column width in tables by using AsciiDoc attributes like:
[cols="2,2,1,1"] (a four column table where the first two columns are wider than the last ones.)

Is there another way to address this issue?

[opoudjis]

There is a ticket requesting that I introduce column width: metanorma/metanorma-standoc#251

There is no other way to address it, but that would. I don't like doing it, being a purist, but I don't think I can avoid this in the long term. WIll realise that ticket.

[anermina]

Since the new ticket is opened to address this issue, I'm closing this one. I'll create a new ticket for these remaining issues which require more time to be solved, so that we can track what's left for the BIPM.