Arrive in Pittsburgh 09:02 pm 09/17/04 Fri
Depart 09:15 am 09/20/04 Mon
No concrete plans yet, except for hanging out with kenoubi, so if you want to get together, let me know :)
Second, I wanted to pose a question related to the task-management techniques I've been working on. The question part is directed at anukul and Joe, but the summary of the task-management techniques is for the benefit of anyone who's interested.
I've been listening to the Getting Things Done audiobook and the system works like this. You basically put unresolved things into an incoming queue, and then set up buckets to hold actions that take no thought to complete, just energy. The only decision making occurs when you're processing the queue, and once the queue is empty you wind up with lists full of tasks to complete when you have some energy and time, and you don't have to interrupt your flow when you're doing them with "what now?" but just keep doing stuff and crossing it off your lists.
In practical terms, you have an inbox for unforeseen stuff that comes up (mail you receive, stuff you think of, stuff you notice) and a "projects" lists of goals you want to accomplish over a short or long term (such as: finish thesis, take vacation in Madrid, organize MP3s). Both of these are sources of tasks; you process your inbox daily and review the project list once a week. You generate tasks by looking at an outcome you want and determing the next specific action which will advance that task, and which must be simple enough to require no strategic thought to complete. If that task will take less than 2 minutes you do it immediately, and otherwise you put it on your task list. When you sit down to work, you whittle down your task lists, which are full of simple, easy things to knock down. In actual implementation, you have several different task lists organized by "context": a list of things you need to buy/drop off next time you're out on errands, a list of people you need to call when you're by a phone, a list of things to do around the house, on a computer, etc.
I have started to use my Palm desktop apps Calendar and Memo as the nexus for this stuff, meaning my palm-phone contains all my action items and reminders wherever I go, and this is great. I also don't have to struggle with Graffiti to edit things too much, since I can use my computer to do it and then sync.
So the process is so far all very algorithmically efficient and pain-free, but after listening to the CDs twice through and working with the system for a week or so, I am still wondering something:
Let's say you have a project "Sell lemonade." In order for this project to succeed, you eventually have to build a lemonade stand, buy lemonade ingredients, hire cashiers, find a good spot by the side of the road, and set up. Let's say this is a hobby type of task, so it's on your projects list and not tracked by any other system at your workplace, and let's say you have the plans for the lemonade stand already. You review your project list on Sunday and go "Ah! Sell lemonade. I have the plans, I need to buy lumber" and you put "Buy lumber" on your Errands task list. On Thursday afternoon you're out shopping and you get the lumber. Now Saturday would be a good time to build the stand, but you don't re-review your projects list until Sunday. So if you follow the system to the letter (and, if you're busy, you probably aren't going to be extra zealous) you won't notice until Sunday that you missed a good opportunity to do the next task for that project and you have to let it go for another week.
My question is: To solve an issue like this, should you (a) review your project list more often to try to reduce the error rate; (b) trace each just-completed task back to its source project and see if you should generate the next task for that project on the spot (or would this ruin the flow of task-completion - I suspect it would); or (c) try to anticipate this kind of thing at the time of generating the original task, and "chain" a couple of them in the task list - perhaps insert contingent tasks. For example, in the "maintenance" task list (along with clean the garage and strip/refinish the old chair) you could put "build lemonade stand (contingent on buying lumber)" and if you didn't get around to buying the lumber, you just pass over this task until you do. The disadvantage to (c) is that you could end up with a lot of clutter - the clean simplicity of "pick a task, any task" is lost.
The question applies also to processing your inbox. Let's say you pick up an item that's an old pair of glasses. This item means: Oh yeah, I need to get new glasses. (You put "new glasses" on your Projects list.) Now you need to find the next action, so you think, "well, I need to update my prescription first. Oh, and I need to find a new eye doctor in this area since I just moved here." So the first task, finding a new eye doctor, will take only a minute, so you consult a website and choose a doctor. Great, now that that's done, and you find a place to store your old glasses in the closet. That's the end of the official process for that one item. But you could have generated the next task, Make appointment for eye exam, at this point as well. Let's say you don't want to wait a week until you review your projects list again, and the appointment will only take two minutes to make too, so you do that. But now you're caught up in following a chain of actions from one item, which means you're not processing your inbox - because you're supposed to do-or-record only one task from each item in the inbox. How/when is it appropriate to chain?
Thanks for any input.
Edit: (7/27/05) According to this 43Folders post, there can be more than one next action generated during a projects list scan. Also, I assume you can scan your projects list as often as you want to, and/or keep working on a project without explicitly generating a next action if time and circumstances allow. So, I suppose that answers the question.