SFv3 examining the mechanics

Author: Bob Pankratz | Created: 2011-03-4 4:14 | Last Activity: 2011-03-9 17:32

Original Post

2011-03-4 4:14
SFv3 has some interesting mechanics; that seem to make it work so well.

I am using paper to do the process, but I did give pause to think over the mechanical process to see if there was some ways to translate this into software. I'm more than happy to do it on paper but I know there are people that are looking for ways to do this in software so these observations may have merit for that too.

The elements of SFv3

1) A Page of finite length
2) A low urgency column
3) A high urgency / in progress column
4) A closed ring of pages
5) A method to add pages and remove pages

The interdependencies

1) The length of the page is directly related to the maximum number of passes across a page before it close. Result: the capacity in number of items in column two determine how long a page will be open relative to your average number of column two tasks on each pass across a page. Short pages will result in slower motion through the list or a higher dismissal rate; along with pressure to start and finish column 1 tasks in one go; or to record shorter tasks. Longer pages will result in a faster speed through the pages; and tolerance for bigger tasks not completed in one go. Question: does a longer page result in more multitasking.

2) The harder you work to close column 1 on the current page reduces the maximum number of passes on the next page by prematurely filling column two, if you fail to complete the items you start on the current page. If you multitask too much the system will put pressure on you to back off finish what you started or to you will have to thin out the list of things you think you need to do.

There are other interdependencies feel free to add to the observations; I need to think on them further.

My first observation here is that there is a complex mechanical process going on here that Mark has made simple on paper; but will be complex in a computer; thinking like a computer does however shed some light on what's going on here.

The heuristics I see so far are:

A) Tasks are assigned a page assignment.

B) Each time you pass through a page; that page is getting a mark against it, representing how any times it's been passed through. This is the pass count of the page.

C) When the page pass count reaches it's max the tasks still assigned to that page must be completed or dismissed.

D) The max for pass count each page is unique and relative to the number of urgent tasks processed by that page; and the number of uncompleted items that pass through that page.

E) The number of task completed on a given page does't effect the next page. But the number of tasks started but not finished will shorten the max pass count of the next page.


That's almost enough info to write a program to do this from scratch that could manage the tasks; it's not quite enough to figure out how to put this into an electronic system; with a set of rules as simple as those for paper; but it's a start; it needs more thought. Feel free to add to the list.

Reply by Alan Baljeu

2011-03-4 13:03
In fact I'm doing this already in OneNote. In place of crossing out tasks, each section has a count of how many C2s were completed so far. If I ever reach 20 I'll dismiss the section, but it hasn't happened yet.

Reply by Seraphim

2011-03-4 16:26
Bob Pankrtaz wrote:
<<< Short pages will result in slower motion through the list or a higher dismissal rate; along with pressure to start and finish column 1 tasks in one go; or to record shorter tasks. Longer pages will result in a faster speed through the pages; and tolerance for bigger tasks not completed in one go. >>>

Interesting observations. I understand how short pages can lead to greater pressure on C2, and thus a higher dismissal rate, greater pressure to start and finish C1 tasks in one go, etc.

But I don't see how short pages leads to slower motion through the list. Can you elaborate on that? Is it because of the pressure to take action (and finish) more C1 items lest they are dismissed because C2 is getting filled up?

Since I am looking for speeding up progress through the pages, I'm especially interested in this.


<<< Question: does a longer page result in more multitasking. >>>

Could you elaborate a bit more on this also?

Reply by wairererose

2011-03-5 0:21
This summary is fascinating. I'd love to see a software for it (I'm unlikely to use it because paper is so very, very portable, although if there were an iPod version I might consider it), but never gave any thought to the different things that would have to apply for each part - programming is not one of my obvious talents (or hidden ones either). I did think of one thing though: Will tasks require a column assignment as well as a page assignment?

Reply by Will

2011-03-5 6:35
Bob, my list was slightly different, based on a programming style forty years out of date:

Each task has
- Description
- Page #
- C1/C2 flag
- Open/closed/dismissed flag
- current task flag

There is a global number of lines /page (which need not be the same for C1 and C2.
There is also a flag for whether any C1 tasks have been worked since you turned the page. And a final page number.

Events are:
- turn page (reset C1 worked flag, change current page number to next page with open tasks)
- review task
- select current task
- finish working on current task (reenter on final page if incomplete, checking whether the page is full and incrementing if necessary)
- add new task (on final page, checking ect)
- add new urgent task (checking whether there is space on current page: if not check next page, etc)

Reply by sabre23t

2011-03-5 9:01
My SFv3 implementation in Toodledo.com (TDD) is very similar to how you defined it, Will.

Each task ...
- Description = Task description
- Page # = Folder (name "01", "02" etc) due to free TDD limits I also use Folder for C2 task (name "01a", "02a" etc)
- C1/C2 flag = Priority (C1 is "0 Low", C2 is "2 High")
- Open/Closed/Dismissed = Completed check box for Open/Closed, use Folder name "dismiss" for dismissed.
- current task flag = Star flag
- one additional field for task number in a page = Tag (name "00" to "25" my chosen page length)

I have my SFv3 list in TDD sorted by Folder than Tag with hiding of completed tasks.

I do the events manually in TDD. ;-)

Reply by Will

2011-03-5 10:58
Oops: forgot to add the dismissal rule to the new page event.

I think the global flag for whether C1 tasks have been worked actually should be a three way flag:
1.Dismiss: C2 full
2.Dismiss: no C1s worked
3. Don't dismiss

Check flag on "new page". If 1 or 2, dismiss all open tasks on old page.
Set page# to next page with open tasks
Check #C2s vs pagelength and set to 1. on new page: if not 1. then set to 2.
Reset to 3. on completing first piece of C1 work

Reply by Will

2011-03-5 11:18
Oops: forgot to add the dismissal rule to the new page event.

I think the global flag for whether C1 tasks have been worked actually should be a three way flag:
1.Dismiss: C2 full
2.Dismiss: no C1s worked
3. Don't dismiss

Check flag on "new page". If 1 or 2, dismiss all open tasks on old page.
Set page# to next page with open tasks
Check #C2s vs pagelength and set to 1. on new page: if not 1. then set to 2.
Reset to 3. on completing first piece of C1 work

Reply by Will

2011-03-5 11:18
Oops: forgot to add the dismissal rule to the new page event.

I think the global flag for whether C1 tasks have been worked actually should be a three way flag:
1.Dismiss: C2 full
2.Dismiss: no C1s worked
3. Don't dismiss

Check flag on "new page". If 1 or 2, dismiss all open tasks on old page.
Set page# to next page with open tasks
Check #C2s vs pagelength and set to 1. on new page: if not 1. then set to 2.
Reset to 3. on completing first piece of C1 work

Reply by Will

2011-03-5 15:38
Sorry about that. Something bizarre seems to have happened when I tried to change a post after the allotted 15 minutes.

Reply by Bob Pankratz

2011-03-8 7:53
Seraphim wrote:
<>

I played with this before I posted that; it was easy since I'm using stenopads. What I did was use the school's paper cutter; and I cut off the bottom half of the steno pad. Which gave me 12 lines to work with instead of 24 and no way to cheat.

What I found was one of two things happened:

1) I slowed down stayed on a given page longer and worked for longer stretches on tasks and completed more of them.

2) I started more tasks and had more simultaneous items in play at once.

Basically .... C2 filled faster if I didn't change my habits. Because C2 filled faster I found I had fewer passes across each page. So since many of my actions for customers can't be dismissed; I had to start them (more multitasking), finish them in less passes (slow down and work longer on ea), or renegotiate them with the customer (thin the list) before the page was dismissed. If I started two many I didn't really run out of space at first because my C1 was also taking more pages; but as I looped through and started bunch of tasks I eventually started to chase the last page; and one time did fill it. So I was left with the conclusion (limited testing disclaimer here) that if the page was short; that I would personally have to slow down and multitask less.

In the end I've settled on using about 21-22 lines on each page; I currently have about 4-5 pages active at a time. I cover about 3 pages a day; and the oldest pages usually are down to 2-3 items. Page seem to close in 4 passes; and because of that most pages are at most 6-8 days old.

This also left me with the impression that starting with a big existing list has a real drag on the process through the list. It seems to take a long time to get momentum if have 4-5 pages of col1 full of stuff; a lot of urgent stuff in col2 and nothing crossed off on any of the pages.

I would suggest the following for people starting with an existing list.

1) Fill Col 1 75% of the way by scanning your old list and adding items that standout to the new col 1

2) Set the rest of the list aside.

3) Add a col 2 on page 2 of the new SF book that says. Add 4-5 tasks from old book to COL1

4) Recur the "add from old book task" in COL1 until the old book is completely processed.

I feel in my case that would have led to better balance in the system sooner. I also think it may be a way to rebalance the system if someone lets col2 run amok .

Reply by Bob Pankratz

2011-03-8 8:17
Will wrote:

<< Bob, my list was slightly different, based on a programming style forty years out of date:>>

Thanks Will I probably should have consider it from a functional based perspective to help balance the ideas in my head. Hadn't gotten that far; thanks for adding in the info.

I was thinking about it strictly from an iPad programming approach since that size device is appealing; the need for sudden stop without the ability to save state had me thinking over the heuristics, and compare them to possible user requirements for options; always good to do before designing objects.

My thought thread was, a program could capture data as it went if the objects were built right. Then if the heuristics of how the method works are well understood; then the usage can be compared to the heuristic.That would allow the program to have a feature set that included a coaching algorithm. So the user could say; help me (1) slow down and focus (2) help me multitask more; (3) help me work faster in smaller burst.... even more clever would be if the software could automatically react to the rate the user enters tasks into column 1 or 2. It really may be as simple as adjusting the virtual page length based on the current average rate of input over a finite period of time.

It's all just a speculative idea right now. But paper works so well; so it seems to me that if you want to use software for something like then the software needs to be smarter and more helpful than paper. That's sort of how the MLO algorithm was born back in the day; it came from my pure laziness about prioritizing my lists. Granted it took 3 years to get it right but I think we did; as I get about 1 email a month asking for changes and I use to get 20 a week.

Reply by Will

2011-03-8 8:33
Bob,

"Objects"?? Since my time... ;o)

I'm just trying to get my head around webOS so I can selfishly stick something on my palm Pre. I was thinking of separate C1 and C2 lists sharing page numbers. It ought to be straightforward to display C1 at black text on white background and C2 as white on black.

I think I'd be sticking with manual mode for most of the time, but I like the idea of a coaching mode to give you the feel of different ways to get thorugh the list.

Reply by Bob Pankratz

2011-03-8 9:12
<"Objects"?? Since my time... ;o)>

LOL...

I still have my COBOL and Fortran and Pascal Books

Objects is kid speak for "Structures" for the lazy...

Reply by Alan Baljeu

2011-03-8 15:11
I've observed that column 2 itself is a paper thing. You'll notice that only the current page and the next have active tasks, so in a digital version of SF you wouldn't actually need a column 2. There's different possible interfaces, but essentially you get:

- a series of pages of tasks
- a list of active items, NOT on any page.
The active items can be finished, or they can be worked on a bit (checked off). Finished items disappear.

When no unchecked items remain, "Turn the page" is enabled. This command shows the next page, and the list of remaining active items are unchecked.

This matches SF rules except for the "Dismiss page if C2 is full". That rule was obviously influenced by the physical limits of paper, but it or something similar is needed. I've resorted to tracking a "C2 count" on each page. It's weird, but it allows this rule to remain in effect.

Reply by Bob Pankratz

2011-03-8 19:23
Good observation Alan,

There is a corner condition, working on the last physical page col1 near full but not full, stop working a col 2 item and it gets reentered on the first physical page that has col2 space. Then staying on that last page you fill col1 and add a new page and uncomplete items wind up in col2 of this now new last page. So now there are two pages with col2 enteries. This has happened to me 4 tines so far, It really a physical artifact tht a software system might better do without as you observed. Still have to track how many col 2 get check off on a virtual pge some the lifespan remaining for the page can be track.

There is also the case were pages are near complete, I find my col2 item then spread forward a page or two. This happens around the time I am about to close several pages at once. I actually find that motivates me to finish those straggerly item. Still need to work the system more to see if that's an artifact or a desirable side effect.

Reply by wowi

2011-03-9 12:29
In fact, I'm using such a mechanics on paper. Reason: I have a small notebook which I don't want to give up just for getting more items on a page. Thus, I use two opposite pages as my column 1. I could put the column two on the next two pages, but this means a lot of page turning. Thus, I use a separate sheet of paper for my column 2.

It contains a list of items which represents the current column 2. I put all urgent things on the next page, thus, the current column 2 is a closed list which can be closed just after I have entered the page (i.e. after I have turned the page and have checked, whether there is still space on the new page). I put a line under the list, starting with the column 2 of the next page.

In addition, I have a counter on each page showing how many items in column2 have been done on the previous passes. If the counter is above the threshold then the page is full. Thus, items in column 1 need to be acted upon or dismissed. Column 2 is not touched in this case - but it may be further extended with unfinished and urgent items. When col.1 of the full page has been completely dealt with, the current column 2 is passed to the next page (if that page is not full as well).

Currently, my "algorithm" includes some slight simplifications:
- if a page is not full when I enter it, all the items of its col2 filled during working on the previous page will be added to this page (i.e. there is no overflow of single items but only of all items produced during the visit of the previous page).
- I add the column to a page only when I turn the page. Thus, the effect at the end of the list mentioned by Bob does not appear in this setting, and the complete column 2is moved to the new last page.

I found this is a simple way to live with a small notebook and have a reasonable number of items on each page (which are visible at one glance). By using a notebook with moveable pages it is easy to handle the extra sheet of paper for capturing column 2. And it's easier to use this mechanism then to describe it in words ...

Wolfgang

Reply by Seraphim

2011-03-9 17:32
Great idea Wolfgang! I went from a pocket book to a larger one, to get more lines per page, but I really miss the portability of the small book.