<p>Every writer I know has the same private ritual: they open a blank page, type one sentence, read it back, and delete it. Then they do it again. Somewhere around the fourth attempt they decide the problem is the coffee, or the chair, or the fact that they have never really had anything to say.</p>
<p>The problem is none of those things. The problem is that they are trying to write the finished sentence first.</p>
<h2>Two different jobs, one keyboard</h2>
<p>Drafting and editing feel like the same activity because they use the same tools. They are not. Drafting is about finding out what you think. Editing is about making what you think legible to someone else. Doing them at the same time is like trying to sketch and frame a picture in one motion.</p>
<p>When you insist on a good first sentence, you are asking your drafting brain to do quality control on material it has not produced yet. It cannot. So it stalls, and the stall feels like proof that you have nothing to say.</p>
<blockquote>The first draft is you telling yourself the story. Every draft after that is you telling it to someone else.</blockquote>
<h2>What bad writing actually buys you</h2>
<p>A deliberately bad draft is fast, and speed is the whole point. Fast writing outruns the internal editor. It lets you find the paragraph on page three that turns out to be the real opening, the one you would never have reached if you had spent forty minutes polishing page one.</p>
<ul>
<li>Write the ugly version of the sentence and move on. Mark it and come back.</li>
<li>Let yourself repeat words. Repetition is a signal, not a failure, and it shows you what the piece keeps circling.</li>
<li>Never delete in the drafting pass. Strike through instead, so you can see the shape of what you rejected.</li>
</ul>
<h2>The permission is the technique</h2>
<p>There is no trick underneath this. The technique is the permission. Once you accept that the first pass is allowed to be genuinely poor, the page stops being a performance and becomes a workspace, and workspaces are where things get made.</p>
<p>Write it badly. You can always fix bad writing. You cannot fix a blank page.</p>

The case for writing badly first

<p>Open any book printed before about 1960 and measure the margins. On a well-set page the outer margin is often wider than the gutter, the bottom is deeper than the top, and the block of text sits noticeably high and to the inside. None of this is decoration. It is the page doing structural work.</p>
<h2>Margins are a reading instruction</h2>
<p>White space around a text block is what tells the eye where the column begins and ends without needing a rule or a border. Narrow the margin and the eye loses the edge, overshoots the line ending, and has to hunt for the start of the next line. That hunting is invisible to the reader. They do not think the margin is too small. They think the piece is hard to read.</p>
<blockquote>Nobody has ever complained that a page had too much air. They complain that it felt cramped, or busy, or tiring, which is the same complaint wearing a coat.</blockquote>
<h2>The screen forgot</h2>
<p>Web layout inherited its margins from a system that had none. Early pages ran edge to edge because the browser window was the page, and we have been slowly rediscovering the printed answer ever since. Measure your line length. Somewhere between sixty and seventy-five characters is where most readers stop noticing the mechanics of reading, and the fastest way to get there is not a smaller font. It is a bigger margin.</p>
<h2>A practical starting point</h2>
<ul>
<li>Set the measure first, then let the margin be whatever is left over.</li>
<li>Give the bottom of the page more room than the top. A block sitting dead centre reads as if it is sliding off.</li>
<li>Keep the margin proportional as the screen grows. A fixed thirty-two pixels looks generous on a phone and mean on a monitor.</li>
</ul>
<p>Margins are the cheapest quality signal in design. They cost nothing, they require no new asset, and they are the first thing to disappear when a layout is rushed.</p>

What a good margin does for a page

<p>I have a shelf of beautiful, mostly empty notebooks. Each one was bought with intent. Each one has between four and nine used pages, then a clean run of paper to the back cover. Four of them are full, and those four are the only ones I still reach for.</p>
<h2>What the finished ones had in common</h2>
<p>They were cheap. Not disposable, but cheap enough that ruining a page cost nothing emotional. The expensive ones stayed empty because every page felt like it needed to earn the paper.</p>
<p>They were also single-purpose. The full notebooks were each about one thing: a project, a house move, a year of reading. The empty ones were all going to be about everything, which in practice means nothing gets written because nothing quite belongs.</p>
<blockquote>A notebook with a subject fills itself. A notebook without one becomes a museum of good intentions.</blockquote>
<h2>The thirty-six taught me something too</h2>
<p>Abandoned notebooks are an honest record of what I thought I would care about. Read together they are almost a diary of enthusiasms: three weeks of bookbinding, a fortnight of learning chess openings, one memorable afternoon of bee-keeping research. I do not regret any of them.</p>
<ul>
<li>Date the first page. It turns an abandoned notebook into a dated artefact rather than a failure.</li>
<li>Never start on page one. Starting in the middle removes the ceremony, and ceremony is what kills notebooks.</li>
<li>Carry one, not three. The choice of which book to write in is where the thought escapes.</li>
</ul>
<p>The shelf is not a graveyard. It is just a very slow filing system.</p>

Notebooks I have actually finished

<p>Ask a new designer to kern a word and they will reach for a number. Minus ten here, plus five there, applied evenly, as if the alphabet were made of identical rectangles. Then they will look at the result and quietly feel that something is wrong without being able to say what.</p>
<h2>The eye counts area, not distance</h2>
<p>What we perceive as even spacing is not equal distance between letterforms. It is equal <em>area</em> of white between them. That is why a capital A next to a capital V can almost overlap and still look correct, while two capital H letters need real daylight between them to look like they are merely adjacent.</p>
<blockquote>Type is set for the eye, and the eye has never once measured anything.</blockquote>
<h2>Where it actually matters</h2>
<p>At body size, leave it alone. The type designer has already solved it, and your adjustments will make it worse across ten thousand words. Kerning is a display problem: logotypes, headlines, anything set above about thirty pixels where a single awkward pair becomes the only thing anyone can see.</p>
<ul>
<li>Squint at the word. Bad pairs show up as dark clumps or bright gaps the moment detail drops away.</li>
<li>Read it upside down. It stops your brain from reading and forces it to look at shapes.</li>
<li>Fix the worst pair, then reassess. Kerning is relative, so one change moves everything.</li>
</ul>
<p>There is no correct value. There is only the point at which you stop noticing, and that point is the answer.</p>

Kerning is a feeling, not a number

<p>Hemingway supposedly stopped writing each day mid-sentence, so that the next morning he knew exactly where to begin. Whether or not he really did, the instinct is sound, and it generalises far beyond writing.</p>
<h2>Finishing is expensive</h2>
<p>The last ten per cent of anything costs about as much as the first ninety. Not because the work is harder, but because finishing requires closing every open question at once, and open questions are the fuel you use to start again tomorrow. Close them all and you wake up to a cold engine.</p>
<blockquote>Stop while you still know what comes next. Momentum is easier to keep than to manufacture.</blockquote>
<h2>Unfinished is not the same as abandoned</h2>
<p>The distinction matters. Abandoned work has no next step, only guilt. Deliberately unfinished work has exactly one obvious next step, written down, waiting. The trick is to leave the note.</p>
<ul>
<li>End the session on the easiest remaining task, not the hardest.</li>
<li>Write one line about what you were about to do. Not a plan, a sentence.</li>
<li>Leave the file open. Friction at the start of a session is the real enemy.</li>
</ul>
<h2>A useful discomfort</h2>
<p>It feels wrong at first. There is a small satisfaction in tying off a day cleanly, and this asks you to give it up. What you get in return is that the blank page never really appears, because you have already written the first line of tomorrow.</p>

On leaving things unfinished on purpose

<p>The twelve-column grid won because it divides neatly. Halves, thirds, quarters, sixths, all available, all at once. It is a genuinely excellent tool and it has quietly made a great many layouts worse.</p>
<h2>Options are not the same as answers</h2>
<p>Give a layout twelve columns and every element becomes a negotiation. Should this card be four wide or five? Should the sidebar take three or four? Each question is small, each answer is defensible, and collectively they produce a page with no point of view, because nothing was ever forced to be more important than anything else.</p>
<blockquote>Constraint does not limit a layout. It tells the layout what it is about.</blockquote>
<h2>What two columns force you to do</h2>
<p>With two columns you get one decision per element: wide or narrow, primary or secondary. That is it. The result is a page with an obvious spine, an obvious hierarchy, and an obvious reading order, and it takes a fraction of the time to build.</p>
<ul>
<li>Put the argument in the wide column and the evidence in the narrow one.</li>
<li>Let the narrow column go empty. Emptiness is a legitimate use of a column.</li>
<li>Break the grid once per page, deliberately, for the single thing that matters most.</li>
</ul>
<h2>When to reach for more</h2>
<p>Dense data wants more columns. Dashboards, catalogues and comparison tables all genuinely need the flexibility. Editorial pages almost never do. If the page is mostly words and pictures arranged in service of an idea, two columns and a wide margin will beat twelve nearly every time.</p>

Notes on making things slowly.

The case for writing badly first

More writing

What a good margin does for a page

Notebooks I have actually finished

Kerning is a feeling, not a number

On leaving things unfinished on purpose

The quiet power of a two-column grid

About