Discussion Forum > Is it time to reconsider AF4?
I hadn't tried this yet but I wonder if AF4 would work with the random method. Something like this: draw a line at the end of the list as in AF4. Start from the first task in the backlog, "roll the dice" and work the list randomly according to the random rules. However, instead of continuing into the active task section once you reach the line, you continue counting from the beginning of the list to the next active task. Leave the active list aside for the time being. However, if you land on an uncompleted task and the "slide" takes you past the line, then you start working the active list as you would with simple scanning. Once there is nothing else you feel like doing in the active list, you roll the dice again and count from the last task you worked on in the active list and work randomly from there. Even if it takes you to a task still in the active list (i.e. you roll a low number).
It seems that this might allow one to work in a more concentrated way on a stack of more urgent tasks from the active list while the randomizer continues to favour the older tasks almost simultaneously. I'll try it starting tomorrow and let you know how it goes.
Since there might be fewer pages in the backlog, I also thought of tweaking the random number range - perhaps instead of the number of lines on a page (L), one could create a simple formula - number of lines on the page less 3 multiplied by the total number of pages. (L-3p). So if you were working with 36 lines and your total number of pages was 5, your random number would be between 1 and 21 (36 - (3 X 5) = 21). If you had 10 pages, your random number would be between 1 and 6. I wonder how that would feel. You'd never get out of the backlog! But then not really, most likely if you had 10 pages most of the tasks would be on your active list. I digress...
It seems that this might allow one to work in a more concentrated way on a stack of more urgent tasks from the active list while the randomizer continues to favour the older tasks almost simultaneously. I'll try it starting tomorrow and let you know how it goes.
Since there might be fewer pages in the backlog, I also thought of tweaking the random number range - perhaps instead of the number of lines on a page (L), one could create a simple formula - number of lines on the page less 3 multiplied by the total number of pages. (L-3p). So if you were working with 36 lines and your total number of pages was 5, your random number would be between 1 and 21 (36 - (3 X 5) = 21). If you had 10 pages, your random number would be between 1 and 6. I wonder how that would feel. You'd never get out of the backlog! But then not really, most likely if you had 10 pages most of the tasks would be on your active list. I digress...
February 6, 2022 at 3:17 |
Paul MacNeil
Paul MacNeil
This sentence is confusing (maybe it was all confusing): "However, instead of continuing into the active task section once you reach the line, you continue counting from the beginning of the list to the next active task." By active here I meant "uncompleted." Here's a clarification of what I meant:
"However, instead of continuing into the active task section once you reach the line, you continue counting from the beginning of the list until you reach the next "uncompleted" task by landing on it or sliding to it, as per the random rules."
"However, instead of continuing into the active task section once you reach the line, you continue counting from the beginning of the list until you reach the next "uncompleted" task by landing on it or sliding to it, as per the random rules."
February 6, 2022 at 3:23 |
Paul MacNeil
Paul MacNeil
I was very fond of AF4 at the start but over time I felt it didn’t give enough due to the things that need lots of action. My remedy which I used on and off for many years inverts the rules:
1. Write everything down. This is the first list.
2. Scan the first list once. (Inverting AF4.) Do anything.
3. Scan the second list repeatedly. (Inverting AF4.). Do anything.
4. Anything unfinished rewrite to the second list. (same as AF4.)
5. Write new tasks in the first list. (Inverse of AF4.)
The effect of this is to emphasize unfinished and active work, rather than emphasize old work. Old work will be activated gradually, and when it becomes active then it will gain focus.
1. Write everything down. This is the first list.
2. Scan the first list once. (Inverting AF4.) Do anything.
3. Scan the second list repeatedly. (Inverting AF4.). Do anything.
4. Anything unfinished rewrite to the second list. (same as AF4.)
5. Write new tasks in the first list. (Inverse of AF4.)
The effect of this is to emphasize unfinished and active work, rather than emphasize old work. Old work will be activated gradually, and when it becomes active then it will gain focus.
February 6, 2022 at 18:00 |
Alan Baljeu
Alan Baljeu
My experience was similar to Alan. Serial No-List is a kind of inverse AF4 -- cycling repeatedly through the newest stuff and periodically running through the older stuff.
February 7, 2022 at 5:16 |
Seraphim
Seraphim
I see that Seraphim created a thread on "Revisiting AF4R." This reminded me of AF4. I always thought that this was one of the more elegant long-list systems. There was once a beautiful pdf demo of it that may be defunct now. I never used AF4 for long (I spent much longer with Simple Scanning and AF1), but I have no doubt that AF4 has much to recommend about it.
Why might we reconsider it now? Back when Mark did the "Review of the Systems" in 2011, he listed the following under "Disadvantages" of AF4:
<< When the Closed List is down to a few items, the vast majority of tasks are in the Open List. This means that there is often a considerable time lag before any pressure is applied to complete a task that is in the Open List. One result of this is that the little and often principle doesn’t work well with the more difficult tasks. >>
It seems that later variations (such as the Ultimate Time Management System) picked up on this logic and applied greater pressure to work on the old tasks. One can also compare this to a post-AF4 system like Superfocus with it's compulsory second column of must-work-on tasks.
There may be nothing wrong per se with the idea of ramping up the pressure to do tasks. (For what it's worth, I have good memories of using Superfocus.) But this approach does seem inconsonant with (some of) Mark's later developments in long list, which have entailed giving the intuition an even freer reign. In Simple Scanning, arguably (one of) the best long-list system(s), there is little pressure to do old tasks.
If the best approach to long list really is to ease up on the pressure and rules constraints (a la Simple Scanning), then Mark's 2011 criticism of AF4 might be a non-issue.
I haven't used a long list in some time. Back when I did, I found Simple Scanning > AF4 because of the (near) total freedom of the former. But, hypothetically, AF4 might have a couple of advantages over Simple Scanning:
1. The old list + dismissal give a gentle push to get moving on stuff (or at least to evaluate tasks);
2. The dismissal process might help people whose Simple Scanning lists get too long (cf. adding "Weed list" as a task in Simple Scanning);
3. The two-list format of AF4 and the different scanning procedures in each list (i.e., one pass only in new list) create a modicum of variety, whereas Mark has complained of Simple Scanning occasionally being "boring" (going "round and round the list").
And yet AF4 is relatively unconstrained rules-wise--in a sense, broadly similar to Simple Scanning. Urgent tasks shouldn't be too much of a problem in AF4, for instance, since it should be relatively easy to go from one list to the other.
Anyway, the above is just food for thought, in case it helps anyone. Of course, some of it may be wrong. I haven't used AF4 in years. I just thought this might add another wrinkle to the recent discussion of long-list systems (Simple Scanning, FV, etc.).