A few days back, I asked about eating near Heathrow. I asked for two reasons.

First, I'm going to be there in a couple of weeks. Much depends on dinner. Thanks for all the great advice. (Consensus: don't try. Get on the Heathrow Express, and eat near Paddington or Notting Hill Gate.)

Second, notes about travel plans are a fascinating note-taking problem. Travel notes have four important purposes:

• Planning. I know we're arriving late and departing early. We sought to eat. London is a great metropolis; why waste a night if we can spend it to good effect?
• Doing. It's always good to know what your plans were, because they're bound to change. Having a rigid schedule is good, but being able to drill down to find the plans you'd considered and rejected can be vital when your schedule comes unglued, or when an unexpected opportunity comes along.
• Telling. Sharing stuff you know is now easy, and it creates a casual web of knowledge. Esther Dyson has long been one of the best tech journalists around; she famously knows more about hotel swimming pools around the world than just about anyone, because she travels a lot and cares about swimming pools. It used to be hard to share information like this; the casual web makes it easy.
• Returning. You might pass this way again; you need to remember where it was that you had those wonderful muffins. After trying a tasty dish, I'll often say, "I'm going to make this!" and then I often forget to try.

You've got to write it down, because you're going to need it.

Let's start a series of notes about travel notes. I'll begin by making a new Tinderbox document and adding these notes about why travel notes matter. For good measure, I'll also drag into it some of the email I received about dining at Heathrow. I'll drop a distinctive color scheme on the document as well, so I'll know I'm working on my travel notes.

I'll also make a new agent in this web log to collect notes about TravelNotes, and let it export a page. And I'll make another agent that takes the top items off that page and makes them into a dedicated RSS feed. With Tinderbox, this took me about 5 minutes from start to end.

(click to enlarge)

When starting a new project like this Travel Notes repository, some people dive in and add lots of data.

Others like to build the perfect structure first, with plenty of apparatus and automation. (Merlin Mann had some choice words for the application of this tactic to personal productivity at last month's Tinderbox Weekend!)

We'll begin, instead, by sketching out some very preliminary ideas, using Adornments to set aside space for some notes we want to take.

• The blue adornments on the left side are meta spaces -- notes about these notes. I've got a meta space for features I want to add, and another one for topics I want to discuss, and yet another area for observations about the project. Another meta space will hold some prototypes.
• The lavender adornments represent places. Some of these are trips I've planned; those have a thin violet border and white titles. Others represent trips I'm considering -- travel I expect I'll need to undertake, trips I'd like to plan if I can find the time; those have a ticker, lighter border and blue titles.

Adornments are a lightweight organizers. They organize space and describe intentions, but they represent a very slight commitment. Not sure where something fits? You can place it at the edge of an adornment, or just outside an adornment, or astride two adornments.

This is an example of spatial hypertext -- using proximity to represent a weak connection. It's also an example of the limitations of outliners: if you don't know how things need to be organized, you can't get started with the organization.

It's possible we'll revise this scheme, and revamp the map according to date. Or we might organize by region instead of place. Some other organizing principle might emerge. I don't know what I'm doing yet: the principle of least commitment advises me to sketch a simple and easily-changed organizing system and get on with the work.

We could be even simpler: why organize at all? I might try to rely completely on search. That's the idea behind One Big Text File, and also the idea behind similarity-seeking tools. I want some organization because it helps remind me what I want to think about: for example, it's good to start thinking about travel you'd like to undertake as well as travel your business and family require. If you know what you'd like, you can work for it: you don't want to wake up and realize you never saw the pyramids because you didn't have an urgent business reason for going.

Before we get much further with the Travel Planning Notes structure, it's time to go ahead and make notes. It's sometimes tempting to make elaborately automated documents when you ought to be simply making some notes. Don't get tied up in infrastructure: go ahead and start making notes. Write it down.

What should we note down? Everything we don't want to forget! The (proposed) dates for the trip. The people I'm hoping to see. Places I'd like to visit. Things I want to do. Things I need to get done.

I've started a simple color coding scheme here. Red things are people. Green things are dates. I don't know this will be useful, at this point, but it helps focus my thinking and suggests new ways to organize and analyze.

It's usually preferable to make lots of small notes instead of a few long ones. Small notes are easier to locate. Because they are tightly focused, small notes make more sense as link destinations. By sticking to one point, short notes are easier to update when things change; if someone is going to be in Hong Kong when I'm in Chicago, I can move them to the next Chicago trip.

The key to success in a project like these travel planning notes is continuous incremental advantage. Each bit of work spent on the project needs to make your work better -- faster, smoother, more accurate, more creative, less stressful.

We've only just started building travel notes, but already they're better than the best we'd expect from other technologies.

• Trying to remember it all is the reason we got started on this in the first place. You can't do it. (If you can do it, and it doesn't cause you stress, you're probably not reading this topic)
• Lots of little bits of paper is a non-starter. The yellow stickies and backs of envelopes and legal pads cause clutter, and they don't really help you find what you need.
• One Big Text File is probably the most competitive technology. We've got all the advantages -- ease of adding information, lightning-fast search for finding things, easy printing when you need a note on paper. But we've also got spatial hypertext for brainstorming, color coding, and links -- and we're just getting started.
• Databases have a big impedance mismatch; we need to design the data model (and get the right data model) before we can even begin populating the database. With a database, we'd still be getting started.
• Outliners share some of the problems with databases and have most of the limitations of one big text file. We need to know how to organize the file right away. We can't know this yet. So we end up dithering, or we end up with One Big Text File in an outliner.

We can deal with some subtle knowledge representation issues that might otherwise give us fits. My mom lives in Chicago, so that's easy. What if your mom lives in Fargo in the summer, but always spends the winter in Dallas with Aunt Jane? Easy enough in Tinderbox: make an alias, drag it to Dallas, and put it right next to Aunt Jane.

Thanks to aliases -- shown here by italics -- the same note (not a copy) can be in two places.

In our Travel Planning Notes, we've been making some prototypes for People, Lodging, Food, Museums, and trip Dates. All these prototypes are gathered together in one corner of the document, which makes them easy to find and which serves as a handy legend to help us (or our assistants) remember that those notes with the gold borders are hotel reservations.

We can customize Tinderbox in a few seconds to make this legend even smarter.

We simply add an action to the adornment underneath the prototypes: IsPrototype=true; Now, any note we create here will automatically be set up to be a prototype. Tinderbox "knows" you're using this part of the map to collect prototypes, and recognizes that if you add a new note here, you're probably intending for it to be a prototype.

We didn't expect, when we designed Tinderbox, that you'd use adornments to do this. We didn't expect, for that matter, that this is something you'd want to do. You don't have to send us a feature request, and you don't have to hire anyone to do custom programming. Just do it, and you're done. Change your mind? You can modify the action, or remove it, just as quickly.

We've been talking a lot about long-range planning. Let's look at improving some short-range matters: making sure we bring what we need on each trip, without wasted time or unnecessary anxiety. What we need, of course, is a packing list.

To begin, I make a BIG list of everything I might want to take along on the kinds of trips I usually take. This is the prototype packing list, and it's just a typical prototype.

Now, when I want to plan what should be packed for a specific trip, I just make a new packing list, based on the prototype. It starts out with the same text as the prototype list -- everything I'm likely to need.

I won't need everything on every trip, of course. I can quickly delete the items I don't want to pack, and add the extraordinary things I'll want to bring only on this trip.

As I revise the list, I may discover things I want to add to the prototype packing list, so they'll be on the next packing list I make. That only takes a moment. It's another example of incremental formalization in Tinderbox; I don't have to worry about getting the prototype exactly right, because I can always change it later.

What could be easier? I just double-click on the trip, tell Tinderbox I want a packing list, and I'm all set.

Because I'm feeling very lazy today, I added a rule to the packing list prototype: if(Name=untitled){Name="pack list";Width=2;}. Now, I don't even need to name the packing lists -- Tinderbox will name it automatically. If I do give the packing list a name -- if, for example, I make separate packing lists for Linda and for me -- Tinderbox respects it.

It's great to have a good packing list, but it's important to check the list. You've got to write things down, but sometimes you also need to read them again.

Of course, we could instantly print the packing list. But that list is formatted for the screen; if we start printing it, we'll be tempted to make it look better on the page.

If we want it to look good on the page, a nice approach turns out to be HTML Export. That's right -- we export the packing list to HTML, even though we have no intention of putting it one the Web. HTML is a good text format, and CSS stylesheets give us good control over how out lists look on the page. Set up some simple templates, press preview, and here's what we get in Safari.

I confess I got a little carried away here, and wasted about 45 minutes tinkering with styles to get it just right. A little extra space between categories, a nice header, flexible scaling in case I need a wallet-size version of the checklist. Your mileage will vary (and you can probably make it look good in much less time.

Nice side effects:

• The list is a standard HTML microformat, so you can import it into all sorts of programs
• You should have no trouble, for example, getting your list on your PDA, or emailing it
• Your Tinderbox note isn't mixed up with print formatting

Tinderbox agents can be useful for complex and sophisticated tasks, but some of the most useful agents can be set up in seconds.

You'll recall that we created a few prototypes to represent different kinds of notes -- notes about people, notes about dates, notes about packing. For example, for trips that are already scheduled, I always have at least one note giving the dates of the trip. Unscheduled treks, such as notes about places I want to go but for which I have no immediate plans, don't have any Date notes.

I can easily set up an agent to gather all these Date notes in one list.

While I'm at it, I can have the agent sort the notes by date, too. Instantly, I've got a high-level itinerary.