diff --git a/pdf/css/remark.css b/pdf/css/remark.css
index b0cbe369c67f8649d4d39e2fadc6ed71c2ef0c77..a9c332cbe35f6007bc1b50bb9f20b848c503c4d0 100644
--- a/pdf/css/remark.css
+++ b/pdf/css/remark.css
@@ -44,4 +44,18 @@ table thead th {
.full-width-pre > pre {
margin-left: -4em;
margin-right: -4em;
+}
+
+blockquote {
+ background: #dee2e6;
+ margin-left: 0;
+ padding: 0.5em;
+}
+
+blockquote > *:first-child {
+ margin-top: 0;
+}
+
+blockquote > *:last-child {
+ margin-bottom: 0
}
\ No newline at end of file
diff --git a/pdf/pdf.html b/pdf/pdf.html
index a0a26287c74f3554b76ae0233cdf36ae4c58e280..cb331dc61321a4dd429c151a9fd3606d6c26b6f1 100644
--- a/pdf/pdf.html
+++ b/pdf/pdf.html
@@ -368,24 +368,35 @@ What about suficient?
| Name | Purpose |
| ---- | :--- |
| cmap | character maps |
-| maxp | memory requirements (#characters, #points, ...)
-| head | global information (max bounding box, font style, ...)
-| OS/2 | windows font metrics
-| name | multilingual strings about font (copyright notice, font name, ...)
-| cvt | list of values for instructions
-| fpgm | program run upon first usage of font
-| gasp | "prefered rasterization on gray-capable devices"
-| prep | program run upon character drawing
-| GDEF | ligatures
-| GPOS | "glyph placement in sophisticated text layout"
-| GSUB | ligatures
-| hhea | sizing per glyph (ascender, decender, ...)
-| loca | offsets of glyph data blocks
-| hmtx | width of glyphs
-| glyf | list of glyph data blocks
+| maxp | memory requirements (#characters, #points, ...) |
+| head | global information (max bounding box, font style, ...) |
+| OS/2 | windows font metrics |
+| name | multilingual strings about font (copyright notice, font name, ...) |
+| cvt | list of values for instructions |
+| fpgm | program run upon first usage of font |
+| gasp | "prefered rasterization on gray-capable devices" |
+| prep | program run upon character drawing |
+| GDEF | ligatures |
+| GPOS | "glyph placement in sophisticated text layout" |
+| GSUB | ligatures |
+| hhea | sizing per glyph (ascender, decender, ...) |
+| loca | offsets of glyph data blocks |
+| hmtx | width of glyphs |
+| glyf | list of glyph data blocks |
In short: Read out tables, patch them together differently, hope offsets work.
+---
+
+# Subsetting
+
+Fonts usually have 100's of glyphs.
+Do we need more than 60?
+
+
+
+
+
---
# Bonus: How to extend ttf?
@@ -405,6 +416,124 @@ like colored glyphs (emojis!):
class: center, middle
+# PDF Writers
+
+---
+
+# State of the art
+
+Overview:
+- PHP provides [TCPDF](https://github.com/tecnickcom/TCPDF) and [FPDF](https://github.com/Setasign/FPDF) (and "new" [tc-lib-pdf](https://github.com/tecnickcom/tc-lib-pdf)).
+- Python has [PyFDPF](https://github.com/reingart/pyfpdf), [PyPDF4](https://github.com/claird/PyPDF4).
+- Some action in the [go](https://github.com/jung-kurt/gofpdf) and [JavaScript](https://github.com/MrRio/jsPDF/blob/master/src/jspdf.js) universe.
+
+Libraries are hard to use.
+
+```php
+private function printH3($text, $startX, $startY)
+{
+ $this->pdfDocument->SetXY($startX, $startY);
+ $this->pdfDocument->SetFontSize(22);
+ $this->pdfDocument->SetFont('opensans', 'b');
+ $this->pdfDocument->MultiCell(0, 0, $text, 0, 'L', false, 1);
+ $this->pdfDocument->Ln(1.6);
+}
+```
+
+Usually html & CSS to PDF "converters" are used -> spin up chrome headless, and use it to print the page.
+
+---
+
+# Architecture (TCPDF)
+
+Two big tasks:
+- Read (and possibly write) `.ttf`
+- Write `.pdf`
+
+Issues:
+- "One File approach": Leads to 24k (`pdf`) and 3K (`ttf`) lines of spaget.
+- Code quality generally **very** low. [1](https://github.com/tecnickcom/TCPDF/pull/361/files) [2](https://github.com/tecnickcom/TCPDF/commit/978eb8c8247cc1069a2935784125596fc3507326) [3](https://github.com/tecnickcom/TCPDF/commit/14fd6779f320c2873f9deab00a9b77f2a657bc98)
+- API unintuitive, badly documented (`AddPage` vs `startPage` vs `addTOCPage`?)
+- Resulting code is untestable, hard to follow, large [print table row](https://github.com/mangelio/web/blob/e6ae315bada04313505e87e513d17778f11df548/src/Service/Report/Pdf/Report.php#L465)
+
+> You cannot meaningfully improve upon this existing work.
+
+---
+
+# Target
+
+Implementation requirements:
+- Add content (text, images, drawings)
+- Within Layout (columns, tables, grid, ...)
+- Fully testable (what text is printed? which color does it have?)
+- No output technology specific details
+
+```php
+// create printer
+$printer = new Printer(new PdfPrinter());
+$printer->setColor("dark grey");
+
+// create layout
+$layout = new ColumnLayout($printer);
+$layout->setColumns(2);
+```
+
+Architecture:
+- `document-generator` provides printer & layouts (nothing PDF-specific!)
+- `pdf-generator` implements interfaces required by `document-generator`
+- `html-generator`, `email-generator`, `word-generator`, ...
+
+---
+
+# `document-generator`
+
+Concept:
+- abstracts PDF specific details (can also attach HTMl / CSS export)
+- `Layout` (recursively!) defines columns, tables, grid, ...
+- `Printer` allows to print text, images within said layout
+
+```php
+/**
+ * @param string[] $rowContent
+ */
+private function printTableRow(TableRowLayoutInterface $row, PrintFactoryInterface $printFactory, array $rowContent)
+{
+ $printer = $printFactory->getPrinter($row);
+
+ $columnLength = \count($rowContent);
+ for ($i = 0; $i < $columnLength; ++$i) {
+ $row->setColumn($i);
+ $printer->printParagraph($rowContent[$i]);
+ }
+}
+```
+
+---
+
+# `pdf-generator`
+
+Concept:
+- abstract lower-level details continously to tame complexity
+- `Printer` prints text, images & drawings
+
+Architecture:
+- `Frontend` (to attach to `document-generator`)
+- `IR` (pdf-agnostic `Printer` with colors, sizing, fonts, ...)
+- `Backend` (actually writes the PDF; 5 more layers of abstraction)
+
+```php
+$printer = new Printer(new Document());
+$printer->setCursor(new Cursor($xPosition, $yPosition, 1));
+$printer->printText('Hello World');
+$result = $printer->save();
+```
+
+... same architecture for `ttf`!
+
+---
+
+class: center, middle
+
# Reflection
---
@@ -422,7 +551,7 @@ What about long term storage (PDF)?
Better Long-Term storage?
- Word / Excel -> nearly same problems as PDF
-- HTML / CSS as structure stays -> but the "web" changes rapidly
+- HTML / CSS as structure stays -> but writing a browser even harder
</textarea>