There was an error while loading. Please reload this page.
📝 docs: Track M's core is built — status, the M0 findings, and what review taught The page said "planned, not built" and stopped being true when #48 merged. Status flips, each slice carries its outcome (M0 changed the plan: the third finding — typography living in classes — grew M2), and a new section states the rules five review rounds taught the contract: fence the inexpressible, honour the differently-expressible, name the vacuous, and a broken component fails the SEND. EN and pt-BR together.
📝 docs: Track M — email rendering Author an email in C# with the same components a page is written with, and get HTML an email client will actually render. Email is a THIRD REALIZER beside web and Photon — a target whose engine is merely very restrictive, which is architecturally the same shape as a target with no DOM. Nothing in the authoring layer changes. Two pieces already exist and the page says so, because a plan that ignores them would overstate the work: WebRealizer.Lower with no style sink keeps styles INLINE rather than atomising them into classes, and HtmlRenderer.RenderNode turns the lowered tree into a string. What is missing is the medium — nested tables for Outlook's Word engine, literal colors, absolute image URLs, the 600px shell, and a fence over everything email cannot do. M0 MEASURES before M1 builds: render an existing component through those two calls and open it in real clients. Half a day, no new code, and it decides the size of the rest. Written that way on purpose after a day in which every confident conclusion drawn from a harness rather than a real run turned out wrong. Page, both sidebars and both roadmaps, EN and pt-BR together.