[book-dp1] Chapter 7 HTML/PDF comparison issues

[Chihiro2000GitHub]

Hi @mmcky @jstac

Below are the discrepancies I found in Chapter 7 when comparing the PDF and the HTML versions.

I'd be grateful if you could take a look when you have a moment 😊

1. Figure 7.2 is not rendered correctly in the HTML

PDF: Figure 7.2 displays two panels, showing the convex and concave cases in Du’s theorem. The axes, 45-degree lines, dashed guide lines, blue curves, points, and labels are all visible.

HTML: The corresponding figure appears to be broken. Most of the graphical elements, including the axes, lines, curves, dashed guide lines, and points, are missing. Only some labels are visible, as shown in the screenshot below:

Issue: The HTML rendering of Figure 7.2 does not match the PDF. The figure should be restored so that all graphical elements are visible as in the PDF.

2. Subscripts in aggregator notation are lowercased in the HTML (Section 7.3.1.4)

PDF: In the list of common aggregators, the subscripts in the aggregator notation are uppercase:
A_MIN, A_UZAWA, A_CES, and A_ADD.

HTML: The corresponding subscripts appear in lowercase:
A_min, A_uzawa, A_ces, and A_add.

Issue: This is a minor notation difference, but the HTML notation does not match the PDF. The same lowercase rendering also appears later in the chapter.

3. Inconsistent rendering of EIS in Exercise 7.3.14

PDF: In Exercise 7.3.14, the two occurrences of EIS appear in the same style.

HTML: The first occurrence of EIS in the displayed equation appears in math style, while the later occurrence in “Confirm that EIS = ...” appears in a different, regular text style.

Issue: This is a very minor difference, but the two occurrences of EIS should probably be styled consistently, as in the PDF.

In the HTML:

In the PDF:

[mmcky]

Thanks @Chihiro2000GitHub — all three confirmed and triaged. None of these touch the LaTeX source, so they're all in the no-@jstac lane (item 3 is optional).

#	Finding	Root cause	Filed
1	Figure 7.2 broken (only labels visible)	Committed du.svg is text-only — 13 glyph-outline paths, no graphics layer	QuantEcon/book-dp1#362
2	Aggregator subscripts lowercase (A_min vs A_MIN)	A_{\textsc{…}} + KaTeX has no small-caps (\textsc→\text)	QuantEcon/book-dp1#363
3	EIS styled inconsistently	KaTeX \text{} font vs body font — rendering difference, not a content issue	QuantEcon/book-dp1#364 (optional)

Item 1. Good catch — the figure is genuinely broken. The HTML pulls a composite du.svg that only ever captured the text overlay (the labels $v_1$, $v_2$, $Tv$); the actual chart graphics from du_concave.pdf / du_convex.pdf were never composited in, so every line, curve, axis and point is missing. Fix is to regenerate the SVG with both layers — derived asset only.

Item 2. Confirmed, and it's the same root cause as the \iidsim "iid"→"IID" fix from Chapter 5 (#359): the subscripts are math-mode small caps (A_{\textsc{min}}), and KaTeX has no small-caps font, so they render lowercase. The PDF keeps its small caps. We'll uppercase them on the HTML side without touching the source.

Item 3. This one is a rendering difference rather than a real inconsistency. In the source, EIS is written the standard way — \text{EIS} inside the displayed equation (it has to be text-within-math) and plain EIS in prose. In the PDF both use the same text font, so they match; in HTML, KaTeX renders \text{} in its own font while prose uses the theme body font, so they diverge. It's the same KaTeX-vs-body-font class as the accepted limitations in #24 and #30 (item A). I've filed it as optional — a reviewer can decide between leaving it as a known limitation or a small cosmetic source tweak (wrapping the prose EIS in \text{}).

We'll follow up here once these are deployed.

[mmcky]

Update — items 1 and 2 are fixed, deployed, and live-verified on https://quantecon.github.io/book-dp1/ :

• Item 1 — Figure 7.2. Regenerated the figure with its graphics layer (#362, PR #366). The composite SVG was text-only; it now carries the axes, 45° lines, dashed guides, the blue curves and the green fixed points alongside the labels. Worth a quick visual confirmation on ch-val when you have a moment — both the convex and concave panels should now look like the PDF.

• Item 2 — aggregator subscripts. The HTML now shows A_MIN, A_UZAWA, A_CES, A_ADD (uppercase), matching the PDF (#363, PR #365). This was the same KaTeX small-caps gap as the \iidsim fix from Chapter 5.

• Item 3 — EIS styling. Filed as QuantEcon/book-dp1#364 and left optional / for review. On closer inspection this is a rendering difference rather than an inconsistency in the source: in the displayed equation EIS is \text{EIS} (KaTeX's font) and in prose it's plain text (the theme body font); LaTeX renders both in the same font so the PDF matches, but KaTeX's \text{} font differs from the body font. It's the same KaTeX-vs-body-font class as the accepted limitations in [book-dp1] Text and rendering discrepancies in Common Symbols table #24 and [book-dp1] Chapter 4 HTML/PDF comparison issues #30 (item A). The issue lays out the two options (leave as a known limitation, or a small cosmetic source tweak) for a reviewer to pick.

Thanks again for the careful Chapter 7 pass!

[Chihiro2000GitHub]

Thanks @mmcky 😊

For item 1, Figure 7.2 looks much better now, and the main graphical elements are visible.

I noticed one minor rendering difference: in the PDF, the two panels are labeled (a) and (b) below the convex and concave cases, but these labels do not seem to appear in the HTML. This seems minor, but I thought I would mention it just in case.

For item 2, I confirmed that the aggregator subscripts now appear in uppercase and match the PDF.

For item 3, that explanation makes sense to me. I understand that this is a rendering difference rather than a real inconsistency.

Thanks again for the quick fixes!

[mmcky]

thanks @Chihiro2000GitHub -- you're reports are exceptional. We really appreciate them. Thanks.