Before talking about how files speak to each other, it helps to understand what a file format actually is. A file format is a language. Word speaks .docx. Excel speaks .xlsx. PowerPoint speaks .pptx. Each format encodes certain kinds of information , .xlsx encodes cells, formulas, charts, and named ranges. .pptx encodes slides, layouts, text boxes, and embedded objects. .docx encodes paragraphs, headings, styles, and referenced objects.

When information moves from one application to another, one of three things happens. It is exported. The source application writes a new file in a different format. The data travels. The original file is unchanged. The new file is a snapshot , a record of how things looked at the moment of export. CSV exports, PDF exports, and Excel exports from Access are all in this category. The exported file has no connection back to the source. It cannot be updated automatically.

The paste options in Word, Excel, and PowerPoint are this same three-way choice, expressed as a menu. The default Ctrl+V is almost always the embedded static copy. The other options are visible when you click the small paste icon that appears after pasting, or when you choose Paste Special from the Edit menu. When you copy an Excel table and paste it into Word, the menu offers:

Most people approach recurring reports by rebuilding them each time. They pull the new data, start a fresh Excel workbook, build the pivot table, create the chart, copy it to PowerPoint, write the summary in Word, and send it. Thirty minutes, maybe forty-five. Every month. The template pattern collapses this. You build the report once, correctly, and then every subsequent month you fill in the blanks.

There is a category of integration work that Claude does well and a category where it adds no value. Claude does well when the task is: express this structured data as natural language, or express this natural language as structured instructions. Translating a pivot table summary into a paragraph for the VP of sales. Turning a plain-English description of a report into SQL that queries the database. Writing the subject line and first sentence of an email that introduces the attached deck.

Everything in this chapter is a variation on the same underlying question: where should this piece of information live, and how should it get everywhere it needs to be? A clinic's address should live in the database, once, and be referenced everywhere. The monthly sales chart should live in Excel and be linked , or embedded, if the deck is going outside the organization. The memo's summary paragraph should be composed with help from Claude but verified against the actual numbers before it leaves your hands.

When information moves between applications, one of three things happens: it is exported (a static snapshot in a new format, no connection to the source), it is linked (a live reference that updates when the source changes but breaks when the source moves), or it is embedded (a self-contained copy inside the destination file that does not update). File formats are languages. CSV is the universal translator , flat, plain text, readable by everything, faithful to structure. JSON preserves hierarchy that CSV cannot.

15.1 (The three-way choice.) You paste a chart from Excel into a PowerPoint deck using the default Ctrl+V. Two weeks later, the Excel data is updated with new figures. Does the PowerPoint slide update? Explain why or why not, using the language of export, link, and embed. Then describe the one-step change that would make the slide update automatically when the data changes , and name the trade-off that change introduces.

The ideas in this chapter didn't appear from nowhere. Charles Goldfarb was working on GML and SGML at IBM , the markup-language architecture that lets one document become a database, a web page, a printed book, and an Excel input file without ever being retyped decades before most people had heard of integrating applications and file formats. Here's a prompt to find out more , and then make it better.

Create a single-file D3 v7 HTML chart using a quantitative y-axis with a zero baseline, 4-6 labeled categories on the x-axis, direct value labels above marks, one primary red series and neutral supporting marks. The figure should represent: A small Excel table (three columns: Product, Quantity, Revenue , with blue header row, currency formatting on Revenue) shown transforming to its CSV equivalent (three lines of plain text, comma-separated, no color, dollar signs present as characters but not as formatting) , student should see what survives the translation and

When information moves between applications, one of three things happens: it is exported (a static snapshot in a new format, no connection to the source), it is linked (a live reference that updates when the source changes but breaks when the source moves), or it is embedded (a self-contained copy inside the destination file that does not update). File formats are languages. CSV is the universal translator , flat, plain text, readable by everything, faithful to structure. JSON preserves hierarchy that CSV cannot.