Computed-Score Priority and Start Dates
256 views
Skip to first unread message
chrisleeuk
Jul 10, 2009, 4:31:41 AM7/10/09
to MyLifeOrganized
I generally like the Computed-Score Priority but I've got one issue
with it that I wonder if anyone has solved, or if a change would be
required in the software.
The problem I have relates to how dates are handled.
Specifically Start Dates.
Say I have two tasks both ending tommorow.
Task A has a start date set a month ago.
Task B has a start date set a week ago.
I understand why Task A will appear higher on the priority list,
because I've had a whole month to do Task A and it's still not
complete.
However I just don't work that way in practical terms.
I only want to use a start date to hide a task that can't start until
say next week.
I don't want the start date to change the weight of the two tasks I'll
use importance and urgency for that.
I've tried dropping the Start Date in the Weight factors but this has
made little difference.
A second issue is tasks with no start date, these generally end up way
below tasks that do have a start date. Again I don't want the start
date to have any say at all in the ordering, because the only reason
one of the tasks had a start date to begin with is because it could
not start until some time in the future.
Any ideas how to solve this one?
with it that I wonder if anyone has solved, or if a change would be
required in the software.
The problem I have relates to how dates are handled.
Specifically Start Dates.
Say I have two tasks both ending tommorow.
Task A has a start date set a month ago.
Task B has a start date set a week ago.
I understand why Task A will appear higher on the priority list,
because I've had a whole month to do Task A and it's still not
complete.
However I just don't work that way in practical terms.
I only want to use a start date to hide a task that can't start until
say next week.
I don't want the start date to change the weight of the two tasks I'll
use importance and urgency for that.
I've tried dropping the Start Date in the Weight factors but this has
made little difference.
A second issue is tasks with no start date, these generally end up way
below tasks that do have a start date. Again I don't want the start
date to have any say at all in the ordering, because the only reason
one of the tasks had a start date to begin with is because it could
not start until some time in the future.
Any ideas how to solve this one?
Timothee
Jul 11, 2009, 12:01:29 PM7/11/09
to MyLifeOrganized
It seems to me too, that Start Date shouldn't affect priority. Can't
anyone think of a reason it should? A scenario where it makes sense
to?
anyone think of a reason it should? A scenario where it makes sense
to?
chrisleeuk
Jul 13, 2009, 4:40:37 AM7/13/09
to MyLifeOrganized
The only scenario I could come up with was the one I touched on above.
MLO seems to assume that because you have had a long time to do a task
it is more urgent, but that would only be true if the amount of time
between the start and end date was a measure of how long the task
takes to do.
Another way to think of this would be to say, hey this is urgent you
have had three months to do this and it's due tommorow so its more
urgent than that other thing.
As I said above, that's just not how I work.
I think we do need the option to take the start date out of the
computer scoring.
This might be making the weight slider drop down to a 'no effect'
setting or a check box saying 'Don't use start date in scoring, or
whatever.
Anyone else agree on this one?
MLO seems to assume that because you have had a long time to do a task
it is more urgent, but that would only be true if the amount of time
between the start and end date was a measure of how long the task
takes to do.
Another way to think of this would be to say, hey this is urgent you
have had three months to do this and it's due tommorow so its more
urgent than that other thing.
As I said above, that's just not how I work.
I think we do need the option to take the start date out of the
computer scoring.
This might be making the weight slider drop down to a 'no effect'
setting or a check box saying 'Don't use start date in scoring, or
whatever.
Anyone else agree on this one?
chuckdevee
Jul 13, 2009, 6:41:33 AM7/13/09
to MyLifeOrganized
Great idea. Definitely agree. This is a big problem for me. The
weightings for start and due date are misleading in that they suggest
you can actually nullify the effect of a start date. In actual fact,
if you enter a start date you put the calculatation into what amounts
to 'egg-timer' mode for that task. Your score depends on the amount of
time that has passed since the start date, as a proportion of total
time for the task (due date less start date). By contrast, a task
without a start date is scored on a time-to-deadline basis.
If I micro-managed my life and had masses of spare time to manage my
tasks, this would be OK I guess, but I'm very busy most of the time
and I have many, many tasks in MLO. In most cases my start date
simply signifies when a task becomes viable, or simply when I expect
or want to start it. I really don't want to be scored using this start
date at all. The only time that is really important to me is how long
until the due date.
A tickbox next to the Start Date slider in the weightings section
which renders it useless and enforces scoring on a time-to-deadline
basis for all tasks (ie assumes that a start date has not been set)
would work really well I think. I might even be able to start properly
understanding and using the computed score priority.
weightings for start and due date are misleading in that they suggest
you can actually nullify the effect of a start date. In actual fact,
if you enter a start date you put the calculatation into what amounts
to 'egg-timer' mode for that task. Your score depends on the amount of
time that has passed since the start date, as a proportion of total
time for the task (due date less start date). By contrast, a task
without a start date is scored on a time-to-deadline basis.
If I micro-managed my life and had masses of spare time to manage my
tasks, this would be OK I guess, but I'm very busy most of the time
and I have many, many tasks in MLO. In most cases my start date
simply signifies when a task becomes viable, or simply when I expect
or want to start it. I really don't want to be scored using this start
date at all. The only time that is really important to me is how long
until the due date.
A tickbox next to the Start Date slider in the weightings section
which renders it useless and enforces scoring on a time-to-deadline
basis for all tasks (ie assumes that a start date has not been set)
would work really well I think. I might even be able to start properly
understanding and using the computed score priority.
Eberhard
Jul 13, 2009, 11:16:36 AM7/13/09
to MyLifeOrganized
+1
Kudos
Jul 13, 2009, 2:12:59 PM7/13/09
to MyLifeOrganized
+1
Andrey Tkachuk (MLO)
Jul 14, 2009, 11:31:20 AM7/14/09
to MyLifeOrganized
Bob Pankratz (the author of the algorithm) and I are looking at the
code at the moment...
A.
code at the moment...
A.
ratz
Jul 14, 2009, 12:31:33 PM7/14/09
to MyLifeOrganized
The start date should effect the priority because MLO was conceived to
be a mini project management system and start dates do matter in many
instances.
We are looking at the weighting factors to see if setting them to the
minimum can gracefully remove them from the computation all together.
In the past that didn't work; rev 1 of the algorithm would get divide
by zero errors; so the weights let you tweak but not removed the
effect of due dates). Rev 2 is done differently and may handle that
case; just not something that was looked at; at the time. Rev 2 was
conceived to make the depth of the outline less of a factor in the
priority and to make sliders at the center as a default possible; aka
our focus was on that; not dates.
This is a recursive algorithm so putting IF gates into it to handle
special case causes exponential increases in the computation overhead
and slow the algorithm down. Not a big deal on the PC side, but the
PPC version this is a BIG deal so we have to be careful about change.
We can't just say "IF john set this preference" then skip this step.
We can do parallel code branches but both have the problem that then
every option someone dreams up has a huge increase in the size of the
code or the or the speed of the code. Or in other words it's really
not as simple as it would seem....
If there is a graceful way to modify the algorithm we'll do it; I
think setting the slider to the minimums may work if we modify
existing branch checks in the code; we calculate the weighting factor
separate from everything else and then feed it into the engine. I
think we'll do something like. If the startdate weight is at the
lowest setting then only use the due date weight formula in all cases;
if both weighting factors are zero then don't weight factor at all;
one key feature will get lost if we do that; if you zero out the due
date weight then you'll loose the special feature that a task that is
due in 1 day and less than 3 days overdue; it gets a pretty major
priority boast in the hopes that you will either get it done and
rescue it; or reschedule it; this saves you from those time bombs that
are hiding in your list a page down. After 3 days the algorithm
assumes you are an optimist that lies to him/herself and it drops the
task back into a normal aging progression.
be a mini project management system and start dates do matter in many
instances.
We are looking at the weighting factors to see if setting them to the
minimum can gracefully remove them from the computation all together.
In the past that didn't work; rev 1 of the algorithm would get divide
by zero errors; so the weights let you tweak but not removed the
effect of due dates). Rev 2 is done differently and may handle that
case; just not something that was looked at; at the time. Rev 2 was
conceived to make the depth of the outline less of a factor in the
priority and to make sliders at the center as a default possible; aka
our focus was on that; not dates.
This is a recursive algorithm so putting IF gates into it to handle
special case causes exponential increases in the computation overhead
and slow the algorithm down. Not a big deal on the PC side, but the
PPC version this is a BIG deal so we have to be careful about change.
We can't just say "IF john set this preference" then skip this step.
We can do parallel code branches but both have the problem that then
every option someone dreams up has a huge increase in the size of the
code or the or the speed of the code. Or in other words it's really
not as simple as it would seem....
If there is a graceful way to modify the algorithm we'll do it; I
think setting the slider to the minimums may work if we modify
existing branch checks in the code; we calculate the weighting factor
separate from everything else and then feed it into the engine. I
think we'll do something like. If the startdate weight is at the
lowest setting then only use the due date weight formula in all cases;
if both weighting factors are zero then don't weight factor at all;
one key feature will get lost if we do that; if you zero out the due
date weight then you'll loose the special feature that a task that is
due in 1 day and less than 3 days overdue; it gets a pretty major
priority boast in the hopes that you will either get it done and
rescue it; or reschedule it; this saves you from those time bombs that
are hiding in your list a page down. After 3 days the algorithm
assumes you are an optimist that lies to him/herself and it drops the
task back into a normal aging progression.
chuckdevee
Jul 14, 2009, 2:56:50 PM7/14/09
to MyLifeOrganized
Correct me if I'm wrong but, after having looked at the formulas, it
doesn't matter of you adjust the start date, or the due date, it does
EXACTLY the same thing in the formula. So it's a bit misleading to
suggest people can independently alter weightings for start and due
date.
Why not keep it as simple and clear as possible and tell users exactly
what's going on.
MLO should just have a single Time Factor slider and two clearly but
succinctly explained options to set it to either:-
- Proportion of task period elapsed.
- Time to due date.
Users can then decide on the relative importance of this Time Factor
in the calculations.
For speed, that would be two seperate algorithms in MLO I guess.
And personally, I wouldn't use that adjustment for tasks between 1 day
to go and 3 days overdue.
There are plenty of other ways now in MLO to highlight these tasks if
you really want to - using the special formats.
I really think you should keep the algorithm as simple, clear and
straightforward as you possibly can.
doesn't matter of you adjust the start date, or the due date, it does
EXACTLY the same thing in the formula. So it's a bit misleading to
suggest people can independently alter weightings for start and due
date.
Why not keep it as simple and clear as possible and tell users exactly
what's going on.
MLO should just have a single Time Factor slider and two clearly but
succinctly explained options to set it to either:-
- Proportion of task period elapsed.
- Time to due date.
Users can then decide on the relative importance of this Time Factor
in the calculations.
For speed, that would be two seperate algorithms in MLO I guess.
And personally, I wouldn't use that adjustment for tasks between 1 day
to go and 3 days overdue.
There are plenty of other ways now in MLO to highlight these tasks if
you really want to - using the special formats.
I really think you should keep the algorithm as simple, clear and
straightforward as you possibly can.
ratz
Jul 14, 2009, 3:17:37 PM7/14/09
to MyLifeOrganized
On Jul 14, 1:56 pm, chuckdevee <don_sm...@me.com> wrote:
> Correct me if I'm wrong but, after having looked at the formulas, it
> doesn't matter of you adjust the start date, or the due date, it does
> EXACTLY the same thing in the formula. So it's a bit misleading to
> suggest people can independently alter weightings for start and due
> date.
>
implement
this without a speed hit speed hit I can tell you that you are wrong.
I'm
not sure what you miss read of if the docs are off (the code has been
tweaked over time); but I think I'm better of spending my time on code
There are spread sheets that graph and model the forumlas on my HD
somewhere and I can assure you the dates matter.
Richard Collings
Jul 14, 2009, 4:59:28 PM7/14/09
to myLifeO...@googlegroups.com
Hi Bob
Really interested in what you say about Rev2 trying to make depth in the
outline less of a factor in the scoring process. This is my major hate in
MLO - the way in which lower level items are automatically made more
important than higher level items to the extent that it is often impossible
to promote the higher level items in front of the lower level items -
particularly if one has applied the Week priority boost. It causes me so
many problems that I just don't use the MLO scoring.
Is there anyway that this recursive boosting can be made optional?
Thanks
Richard
Really interested in what you say about Rev2 trying to make depth in the
outline less of a factor in the scoring process. This is my major hate in
MLO - the way in which lower level items are automatically made more
important than higher level items to the extent that it is often impossible
to promote the higher level items in front of the lower level items -
particularly if one has applied the Week priority boost. It causes me so
many problems that I just don't use the MLO scoring.
Is there anyway that this recursive boosting can be made optional?
Thanks
Richard
ratz
Jul 14, 2009, 10:32:22 PM7/14/09
to MyLifeOrganized
You are already using REV2. Which has been around for 3-5 years.
First some comments then some news about REV3.
The depth of the outline really shouldn't come into play UNLESS
you use the WEEKLY GOAL setting. The weekly goal applies
a HUGE boost and that does cascade down because that's
the original design intent. It was the HOLLY CRAP feature
to make something POP to the top.
If you don't use that weekly goal (never like that myself);
things inherit correctly; whenever someone has an outline
that is out of whack I usually find they have a item near the top
with aggressive importance or urgency which propagates
down a really deep tree. Which is what it was designed to do;
and it is simply a case that that user just didn't expect the impact
to be so strong. Lower the priority cascade down too but people
always see to over look LOWERing the importance to balance the
outline... any how I digress.
The boosting factor I'm speaking of is that we BOOST the priority
of tasks that are overdue; not like we do with weekly goals; but we
do exponentially increase as the moves past the due in 1 day mark.
****
REV 3 of the algorithm was just sent to Andrey; this is the first
redux in about 3-5 years.
****
I took a good long look at it and decided it was time to re-factor the
code.
I found a couple things I no longer liked, Remember when the
algorithm REV2 was last written you had to have both a start and a
due date they weren't optional.
When Andrey gave you the ability to have no start or no due date;
the algorithm could handle it BUT it was not optimized for those cases
by any means; In the end it did work just not like everyone thought it
should
Such is the problem with changing things because someone likes
purple instead of green... opinions are like.....and sometimes the
results aren't perfect.
Here's the enhancements
0) Made it faster never hurts to protect our PPC user's batteries.
1) Added the ability to turn off the over-due boosting via a
preference.
tasks will not use the more aggressive calculation when over due
if this is disabled. (consider it the "don't nag me if
procrastinate"
option)
2) Made it possible to have due and start date weighting factors of
ZERO
or to out right disable them.
3) Change the computations for the duedate=startdate condition.
When start=due now the start acts only as a snooze function
and only the due date calculation is used. Old logical / code,
was using the start <> due calculation but was convoluted so
bad that you couldn't tell when you read the code; and it only
threw things out of whack on very deep outlines.
4) Fixed the miss handling of the startdate with nodue date condition.
Startdate now acts a a snooze only; previously the algorithm
treated this by using the StartDate as the DUE date; this happened
when it became possible to have NO Due Date. Most people
assume no due date means finish this task some day but who
cares when; the algorithm is now designed to think that way about
this case as well.
5) Made some structural changes that make it much much
easier to adapt as MLO evolves in the future.
It's a pretty big write so it will take him a few days to look over
decide
if he likes it and then integrate it. Or it will take him 1 day;
sometimes
he's crazy fast.
I usually try not to give him code that requires him to change the
gui or the object model. This time I did both *gasp* so it might
take a bit of integrate and I have no insight into which version
if any will use this code. I did hack together a version that
should plug into the existing MLO version so he could do quick
beta testing with select users. If you get tapped to test it realize
that the testing would be desktop only and probably break
your PPC sync if you use it; so beware what you ask for.
ok that's all; I'm going back to pondering on how to create
AutoFocus and AutoFocus2 templates for MLO without
begging for 30 new features.... ok maybe 5 or 6.
..... See
http://www.markforster.net/autofocus-system/
http://www.markforster.net/blog/2009/6/27/autofocus-2-time-management-system-af2.html
and
http://www.markforster.net/forum/post/835234
First some comments then some news about REV3.
The depth of the outline really shouldn't come into play UNLESS
you use the WEEKLY GOAL setting. The weekly goal applies
a HUGE boost and that does cascade down because that's
the original design intent. It was the HOLLY CRAP feature
to make something POP to the top.
If you don't use that weekly goal (never like that myself);
things inherit correctly; whenever someone has an outline
that is out of whack I usually find they have a item near the top
with aggressive importance or urgency which propagates
down a really deep tree. Which is what it was designed to do;
and it is simply a case that that user just didn't expect the impact
to be so strong. Lower the priority cascade down too but people
always see to over look LOWERing the importance to balance the
outline... any how I digress.
The boosting factor I'm speaking of is that we BOOST the priority
of tasks that are overdue; not like we do with weekly goals; but we
do exponentially increase as the moves past the due in 1 day mark.
****
REV 3 of the algorithm was just sent to Andrey; this is the first
redux in about 3-5 years.
****
I took a good long look at it and decided it was time to re-factor the
code.
I found a couple things I no longer liked, Remember when the
algorithm REV2 was last written you had to have both a start and a
due date they weren't optional.
When Andrey gave you the ability to have no start or no due date;
the algorithm could handle it BUT it was not optimized for those cases
by any means; In the end it did work just not like everyone thought it
should
Such is the problem with changing things because someone likes
purple instead of green... opinions are like.....and sometimes the
results aren't perfect.
Here's the enhancements
0) Made it faster never hurts to protect our PPC user's batteries.
1) Added the ability to turn off the over-due boosting via a
preference.
tasks will not use the more aggressive calculation when over due
if this is disabled. (consider it the "don't nag me if
procrastinate"
option)
2) Made it possible to have due and start date weighting factors of
ZERO
or to out right disable them.
3) Change the computations for the duedate=startdate condition.
When start=due now the start acts only as a snooze function
and only the due date calculation is used. Old logical / code,
was using the start <> due calculation but was convoluted so
bad that you couldn't tell when you read the code; and it only
threw things out of whack on very deep outlines.
4) Fixed the miss handling of the startdate with nodue date condition.
Startdate now acts a a snooze only; previously the algorithm
treated this by using the StartDate as the DUE date; this happened
when it became possible to have NO Due Date. Most people
assume no due date means finish this task some day but who
cares when; the algorithm is now designed to think that way about
this case as well.
5) Made some structural changes that make it much much
easier to adapt as MLO evolves in the future.
It's a pretty big write so it will take him a few days to look over
decide
if he likes it and then integrate it. Or it will take him 1 day;
sometimes
he's crazy fast.
I usually try not to give him code that requires him to change the
gui or the object model. This time I did both *gasp* so it might
take a bit of integrate and I have no insight into which version
if any will use this code. I did hack together a version that
should plug into the existing MLO version so he could do quick
beta testing with select users. If you get tapped to test it realize
that the testing would be desktop only and probably break
your PPC sync if you use it; so beware what you ask for.
ok that's all; I'm going back to pondering on how to create
AutoFocus and AutoFocus2 templates for MLO without
begging for 30 new features.... ok maybe 5 or 6.
..... See
http://www.markforster.net/autofocus-system/
http://www.markforster.net/blog/2009/6/27/autofocus-2-time-management-system-af2.html
and
http://www.markforster.net/forum/post/835234
ratz
Jul 14, 2009, 10:34:44 PM7/14/09
to MyLifeOrganized
oh good golly rest assured my code is better than my grammar and
formatting in that post...
I need some sleep.
formatting in that post...
I need some sleep.
chuckdevee
Jul 15, 2009, 2:58:44 AM7/15/09
to MyLifeOrganized
Hi Ratz, this is why I thought the srat date and due date were
identical in the calculations..here is the main formula from the MLO
help guide for those tasks with a start date:
date score contribution = ((StartDate WeightFactor + DueDate
WeightFactor )/ (Task Duration / Elapsed)) /500
If it's wrong, then it needs to be changed in the explanation, but
according to this formula, the strart date and due date have exactly
the same effect on the calculation.
Ideally, I'd like there to be an option for the date factor to be
scored based on an assumption that there is no start date, using the
formula you have for these tasks:
date score contribution = (DueDate WeightFactor / (1 – (1 /
Remaining)))/2500
Can you tell me please, do your changes accommodate this? From what I
have read, it's not clear that they do, but maybe I'm misreading
this.
I need to be able to disable the Start Date but still keep the Due
Date weight factor... and I think this is the main thrust of suggested
changes in this thread..
identical in the calculations..here is the main formula from the MLO
help guide for those tasks with a start date:
date score contribution = ((StartDate WeightFactor + DueDate
WeightFactor )/ (Task Duration / Elapsed)) /500
If it's wrong, then it needs to be changed in the explanation, but
according to this formula, the strart date and due date have exactly
the same effect on the calculation.
Ideally, I'd like there to be an option for the date factor to be
scored based on an assumption that there is no start date, using the
formula you have for these tasks:
date score contribution = (DueDate WeightFactor / (1 – (1 /
Remaining)))/2500
Can you tell me please, do your changes accommodate this? From what I
have read, it's not clear that they do, but maybe I'm misreading
this.
I need to be able to disable the Start Date but still keep the Due
Date weight factor... and I think this is the main thrust of suggested
changes in this thread..
chuckdevee
Jul 15, 2009, 3:50:09 AM7/15/09
to MyLifeOrganized
.. actually Bob to be more clear, my key question is:
If I enter a task with no start date, and one with a start date, under
this new algorithm, is there a way that MLO can score the date factor
for these tasks identically so that the start date is irrelevant. And
will I still be able to change the effect of dates on the overall
score by changing the due date weight factor?
thanks
If I enter a task with no start date, and one with a start date, under
this new algorithm, is there a way that MLO can score the date factor
for these tasks identically so that the start date is irrelevant. And
will I still be able to change the effect of dates on the overall
score by changing the due date weight factor?
thanks
Richard Collings
Jul 15, 2009, 4:17:37 AM7/15/09
to myLifeO...@googlegroups.com
Hi Bob - you wrote:
>
> The depth of the outline really shouldn't come into play
> UNLESS you use the WEEKLY GOAL setting. The weekly goal
> applies a HUGE boost and that does cascade down because
> that's the original design intent. It was the HOLLY CRAP
> feature to make something POP to the top.
>
> If you don't use that weekly goal (never like that myself);
> things inherit correctly; whenever someone has an outline
> that is out of whack I usually find they have a item near the
> top with aggressive importance or urgency which propagates
> down a really deep tree. Which is what it was designed to do;
> and it is simply a case that that user just didn't expect the
> impact to be so strong. Lower the priority cascade down too
> but people always see to over look LOWERing the importance to
> balance the outline... any how I digress.
>
> The depth of the outline really shouldn't come into play
> UNLESS you use the WEEKLY GOAL setting. The weekly goal
> applies a HUGE boost and that does cascade down because
> that's the original design intent. It was the HOLLY CRAP
> feature to make something POP to the top.
>
> If you don't use that weekly goal (never like that myself);
> things inherit correctly; whenever someone has an outline
> that is out of whack I usually find they have a item near the
> top with aggressive importance or urgency which propagates
> down a really deep tree. Which is what it was designed to do;
> and it is simply a case that that user just didn't expect the
> impact to be so strong. Lower the priority cascade down too
> but people always see to over look LOWERing the importance to
> balance the outline... any how I digress.
Thanks for this explanation.
As I remember it, it was the use of the weekly goal that was causing the
problem and specifically that the Weekly Goal boost that is applied
recursively.
I asked at the time what the business logic was behind this - and nobody
could supply an answer.
From memory, if you have a situation like this:
Project A
Task 1
Task 2
Task 3
Task 4
And you apply the Week Goal to Project A, then Task 1 gets boosted by the
Weekly Goal factor twice and Tasks 3 and 4 get boosted three times which
makes it impossible to bring Task 1 in front of Tasks 3 and Task 4. I
was (and still am) totally bemused by this - why should Tasks 3 and 4 be
more important than Task 1 - the fact that Task 2 logically breaks down into
two smaller tasks does not automatically make those tasks more important, in
my view. And when I asked on the board previously, nobody could explain
why this was. And nobody said, please don't change this, I find it really
useful because .....
So it seems that we have a feature which causes lots of problems to new
users, which some people hate (me!) and which nobody uses/defends. Am I
missing something? Incidentally, what does HOLLY CRAP mean!!
Is there any chance of getting this changed? Or at least making it an
option. For me, the Weekly Goal boost is useful but it should just boost
all the tasks to which it applies equally.
Regards
Richard
ratz
Jul 15, 2009, 2:02:34 PM7/15/09
to MyLifeOrganized
Yes if Andrey implements the preference to turn off the boost
you can have exactly what you want. I used this thread to find
the problem in the first place.
And yes IMO the case you refer to was incorrectly selecting
the wrong formula because of the changes that resulted from
the addition of null dates. It worked but not optimally IMO.
The docs you refer to really document REV 1.5 of the algorithm
They will be updated when REV 3.0 gets put in; such it the
world of docs; no one notice that null dates meant updating that.
The old code was highly optimized compound if branchs very fast but
hard to read. The new stuff needs to handle more cases
and is now a case statement with other optimization to
get the speed back. It's now REALLY obvious which
branch algorithm is getting used. This should prevent
future changes from have really hard to locate logic bombs.
It's not my day job and I'm super busy but I think the
6 hours of work I did on it yesterday should leave use in a
good place for a while.
In the end I'm a volunteer and the CSA is something
I'm responsible for and Andrey graciously puts
in MLO for me. The Hierarchal method is his.
I've always said blame me if you don't like it
I'll do what I can when I can to improve it but the
delays will be long. The last 3 years have been
crazy and I suspect in the future I can be more responsive.
you can have exactly what you want. I used this thread to find
the problem in the first place.
And yes IMO the case you refer to was incorrectly selecting
the wrong formula because of the changes that resulted from
the addition of null dates. It worked but not optimally IMO.
The docs you refer to really document REV 1.5 of the algorithm
They will be updated when REV 3.0 gets put in; such it the
world of docs; no one notice that null dates meant updating that.
The old code was highly optimized compound if branchs very fast but
hard to read. The new stuff needs to handle more cases
and is now a case statement with other optimization to
get the speed back. It's now REALLY obvious which
branch algorithm is getting used. This should prevent
future changes from have really hard to locate logic bombs.
It's not my day job and I'm super busy but I think the
6 hours of work I did on it yesterday should leave use in a
good place for a while.
In the end I'm a volunteer and the CSA is something
I'm responsible for and Andrey graciously puts
in MLO for me. The Hierarchal method is his.
I've always said blame me if you don't like it
I'll do what I can when I can to improve it but the
delays will be long. The last 3 years have been
crazy and I suspect in the future I can be more responsive.
ratz
Jul 15, 2009, 2:03:02 PM7/15/09
to MyLifeOrganized
Yes, and Yes.
ratz
Jul 15, 2009, 2:24:54 PM7/15/09
to MyLifeOrganized
The weekly goal was an option carried over from the Hierarchal Method;
that just got grafted into the CSA.
It simply affects Urgency; and it's from before the urgency slider was
added. It was a way to make something urgent. It's going to be very
sensitive to outline depth. It was designed to drive things deep in
the outline to the top and it's a very old feature.
The best thing would probably to remove it's effect from computed
score all together the way monthly and weekly are ignored. If you
really need to boost a project you should move the urgency slider
these days as that is the proper way. The weekly goals is redundant
and flawed from the perspective of the CSA and that's why the urgency
slider was created.
I have to think about that; and review with Andrey; it's a one line
change to the algorithm; I don't very much that we'd enhance it to do
anything else as it's after all redundant.
..... but I could see having the Weekly goal checked magically move
the urgency slider up 1 full notch but that would probably freak out a
percentage of the user base. We could probably do that transparently
under the hood without anyone noticing and get the desire effect; with
the cavet that if you maxed out the urgency slider AND check the
weekly goal; the weekly goal would have no effect as you'd already be
at the max setting and you can't go past the max without returning to
the original problem you described. Any how that's all code in the
gui; so I'll share my thoughts with him on how that could work.
that just got grafted into the CSA.
It simply affects Urgency; and it's from before the urgency slider was
added. It was a way to make something urgent. It's going to be very
sensitive to outline depth. It was designed to drive things deep in
the outline to the top and it's a very old feature.
The best thing would probably to remove it's effect from computed
score all together the way monthly and weekly are ignored. If you
really need to boost a project you should move the urgency slider
these days as that is the proper way. The weekly goals is redundant
and flawed from the perspective of the CSA and that's why the urgency
slider was created.
I have to think about that; and review with Andrey; it's a one line
change to the algorithm; I don't very much that we'd enhance it to do
anything else as it's after all redundant.
..... but I could see having the Weekly goal checked magically move
the urgency slider up 1 full notch but that would probably freak out a
percentage of the user base. We could probably do that transparently
under the hood without anyone noticing and get the desire effect; with
the cavet that if you maxed out the urgency slider AND check the
weekly goal; the weekly goal would have no effect as you'd already be
at the max setting and you can't go past the max without returning to
the original problem you described. Any how that's all code in the
gui; so I'll share my thoughts with him on how that could work.
Richard Collings
Jul 15, 2009, 4:22:17 PM7/15/09
to myLifeO...@googlegroups.com
Hi Bob
Thanks for the detailed reply. I have a question and then an observation
re:
>
> The weekly goal was an option carried over from the
> Hierarchal Method; that just got grafted into the CSA.
>
> It simply affects Urgency; and it's from before the urgency
> slider was added. It was a way to make something urgent.
> It's going to be very sensitive to outline depth. It was
> designed to drive things deep in the outline to the top and
> it's a very old feature.
>
Does this mean that boosting the Urgency of a top level task will also
generate a depth related boost down the tree below that task - ie that the
urgency boost of the top level task is applied recursively down the tree
(once to the top level task, twice to its children, three times to their
children and so on).
If so, then this just doesn't work for me. Taking my example again:
> > Project A
> > Task 1
> > Task 2
> > Task 3
> > Task 4
If I boost the urgency of A, I would like Tasks 1, 3 and 4 to all receive
the same boost and not to suddenly find that Tasks 3 and 4 appear above Task
1. I just cannot see the logic of this - all I have said is that A is now
more urgent. Why should Tasks 3 and 4 then suddenly become more important
than Task 1?
If this recursive boosting is the case, then I would make a strong plea for
this behaviour to be made optional - ie: to have 'Switch off recursive
boosting' (or similar) which when ticked will result in the boost just being
applied once to the Task in question and to all the children and their
children, etc.
Many thanks
Richard
ratz
Jul 15, 2009, 5:15:43 PM7/15/09
to MyLifeOrganized
Honestly.....I'd have to go look again; I really haven't thought about
urgency in a long time. I believe after thinking about it that
importance is recursive and urgency is not but I will check an make a
authoritative statement. later. ( I was working in a different part of
the algorithm that runs in parrallel so I didn't have to concern
myself with thinking about the urgency topic)
I will say that it's highly unlikely we'll change the way urgency
works because it does what it was suppose to do and and people expect
it to do what it does now. So don't spend a ton of time formulating an
argument; we've been through that 4 years ago.
Fixing the weekly goal is the only real topic open for discussion.
I'll review urgency only so much as finding the right way to fix the
weekly goal issue my above thoughts were open thinking on the fly that
doesn't mean they are the correct solution; just me thinking out loud;
only so much as the weekly goal issue is concerned and sometimes I
draw bad ideas when doing that; we sort through that when I try and
implement them.
Completely separate from thoughts of the weekly goal
If urgency as implemented isn't to your liking you have several
options:
1) Don't use the urgency slider
2) Set the preference to by importance only
3) Use the hierarchal priority method
That should suffice for anyone's needs; the program has got so many
different ways to tweak the priority that it is silly. And this this
program has too many options already and we can't bend the algorithms
to everyone's whims or the program would be unfathomable to new users.
The additive approach your suggesting really isn't' in the cards for
the design.; that's what the weekly goal was suppose to do and it
doesn't work because it's really really hard to track it down the tree
as you recurse. lots of stack space and speed issues and plenty of
places to make mistakes; and it confuses people... really trust me it
does; the last time we went over this everyone had trouble keeping the
additive and multiplicative properties straight during the discussion
and much arguing and crying occurred.
I go off to think about it some more. maybe something simple and
elegant will occur to me ... no promises.
urgency in a long time. I believe after thinking about it that
importance is recursive and urgency is not but I will check an make a
authoritative statement. later. ( I was working in a different part of
the algorithm that runs in parrallel so I didn't have to concern
myself with thinking about the urgency topic)
I will say that it's highly unlikely we'll change the way urgency
works because it does what it was suppose to do and and people expect
it to do what it does now. So don't spend a ton of time formulating an
argument; we've been through that 4 years ago.
Fixing the weekly goal is the only real topic open for discussion.
I'll review urgency only so much as finding the right way to fix the
weekly goal issue my above thoughts were open thinking on the fly that
doesn't mean they are the correct solution; just me thinking out loud;
only so much as the weekly goal issue is concerned and sometimes I
draw bad ideas when doing that; we sort through that when I try and
implement them.
Completely separate from thoughts of the weekly goal
If urgency as implemented isn't to your liking you have several
options:
1) Don't use the urgency slider
2) Set the preference to by importance only
3) Use the hierarchal priority method
That should suffice for anyone's needs; the program has got so many
different ways to tweak the priority that it is silly. And this this
program has too many options already and we can't bend the algorithms
to everyone's whims or the program would be unfathomable to new users.
The additive approach your suggesting really isn't' in the cards for
the design.; that's what the weekly goal was suppose to do and it
doesn't work because it's really really hard to track it down the tree
as you recurse. lots of stack space and speed issues and plenty of
places to make mistakes; and it confuses people... really trust me it
does; the last time we went over this everyone had trouble keeping the
additive and multiplicative properties straight during the discussion
and much arguing and crying occurred.
I go off to think about it some more. maybe something simple and
elegant will occur to me ... no promises.
RichardCollings
Jul 15, 2009, 6:16:11 PM7/15/09
to MyLifeOrganized
Hi Bob
Thanks very much for taking the time to reply in such detail. I can
undertand your reluctance to go over ground that was clearly covered
in some detail some time ago (before I got involved with MLO).
I would be very interested to know if one of the sliders does not
apply a recursive boost because this is what I want (desparately).
What is frustrating for me is that when I have posted previously on
this topic nobody has been able to explain the reasoning behing the
recursive boost - why from a project planning/business point of view,
lower level leaf tasks in a hierarchical structure should be made more
important than other leaf tasks that appear higher up in the
hierarchy?
ie: Going back to my example:
> > > > Project A
> > > > Task 1
> > > > Task 2
> > > > Task 3
> > > > Task 4
why should Tasks 3 and 4 be made more important than Task 1 when I
boost Project A?
I agree that there are lots of different ways of exploiting the
algorithm but there does not appear to be a way of handling my simple
requirement which is
<<When I boost a top level task, I want all the subtasks to receive
the same level of boost irrespective of their depth in the hierarchy
below that higher level task. ie: they retain their relatively levels
of importance/urgency>>
This does not seem an unreasonable request. Incidentally, I am
pretty certain that the hierarchical scoring method does meet this
requirement.
If you can throw any more light on this, I would be very grateful.
Many thanks.
Richard
> > Richard- Hide quoted text -
>
> - Show quoted text -
Thanks very much for taking the time to reply in such detail. I can
undertand your reluctance to go over ground that was clearly covered
in some detail some time ago (before I got involved with MLO).
I would be very interested to know if one of the sliders does not
apply a recursive boost because this is what I want (desparately).
What is frustrating for me is that when I have posted previously on
this topic nobody has been able to explain the reasoning behing the
recursive boost - why from a project planning/business point of view,
lower level leaf tasks in a hierarchical structure should be made more
important than other leaf tasks that appear higher up in the
hierarchy?
ie: Going back to my example:
> > > > Project A
> > > > Task 1
> > > > Task 2
> > > > Task 3
> > > > Task 4
boost Project A?
I agree that there are lots of different ways of exploiting the
algorithm but there does not appear to be a way of handling my simple
requirement which is
<<When I boost a top level task, I want all the subtasks to receive
the same level of boost irrespective of their depth in the hierarchy
below that higher level task. ie: they retain their relatively levels
of importance/urgency>>
This does not seem an unreasonable request. Incidentally, I am
pretty certain that the hierarchical scoring method does meet this
requirement.
If you can throw any more light on this, I would be very grateful.
Many thanks.
Richard
>
> - Show quoted text -
ratz
Jul 15, 2009, 11:01:31 PM7/15/09
to MyLifeOrganized
Yes we seem to have fun explaining this topic, but the software does
do the right thing from a fuzzy logic project management task break
down approach.
Let's have a little philosophy of the methods discussion I think that
will help you see what CSA does what it does.
We have 2 scheme's
(a) Hierarchal is a method that uses and arithmetic progression down
the tree using addition. That method assumes that all tasks are rated
against the universe at large on a fixed scale. This is traditional
prioritization with a few necessities for making mass changes and
boost whole groups with the goals functions and a little hierarchal
smoothing thrown in. This method is great if you have less than 200
tasks and if you are disciplined and consistent I bet you can do 500
without tiring of the effort of prioritizing correctly. This method is
Andrey's baby and it works great for what it was designed to do. I
like it! and so do a lot of people.
(b) CSA is a method, that use an arithmetic progress down the tree
using multiplication of logarithmic reversible number pairs to
calculate a relative priority based on minimal data entered around a
localized position in the tree. (sounds sexy doesn't it? or just Bs?
actually it just some math theory that happens to be pretty it's a
GLOB Sorter if you want to get technical it cluster "like data" into
groups of similar values ). Under this model you set each tasks
Importance and urgency relative to it's immediate parent only. How
important and how urgent is this individual task to completing the
parent task; and only the parent task; not the project at a whole,
that's the KEY the parent task only. That allows for faster data
entry within huge outlines with 500 to 5000s of tasks. Because you
don't have to evaluate the task against your whole life; just it's
importance and urgency to the parent task, and when it's do. That is
localized positioning. IF you set your values that way the CSA will
give you very accurate results for priorities. I know I've been using
it for almost 10 years as lifebalance uses a simpler form of this
approach and I started on that tool in 98. This scheme is designed
specifically for people that have to make decisions about what gets
done AND what does NOT get done. Just because it's due today doesn't
mean it should be done. If figuring out which tasks should even be
reviewed on a give day is a challenge, then CSA is the method you
want. The CSA gets you a nice list of likely suspects to review. This
lines up nicely with GTD that says to own you own intuitive
prioritizations, so we often recommend CSA to GTDers' because it make
a first WHACK at you list for you; and reduces the number of items you
have to consider for you final selection of the correct task to do.
The problems usually creep in when people try to use CSA in a manner
other than intended; it will not make your decisions for you and it
won't process a really short list all that effectively that's why tiny
short lists give weird results; it wasn't designed to do what people
often try to test. It's also not a GANTT chart and it won't schedule
time linear linked tasks; if you need that see MS project and numerous
other tools or fall back to Hierarchal. CSA will always get the top
15-25 things to do in the right cluster at the top of the list out of
1000s of tasks. That's what it's designed to do. Get you a todo list
where the top screen without scrolling down at all has the things that
should be review and action'ed as necessary. The order of that screen
will never be perfect because only your intuition at the time of
choosing will tell you which of the top 15 things is the right one to
do right now right here.
That's what the method does. It really can't be bent to do other
things. But people loose site of that and start to blend the two
different methods characteristics. I you expect the computed todo list
to be ordered 1, 2, 3 ,4 exactly like you are expecting it you will be
disappointed. Don't pound nails with a screw driver; use Hierarchal
instead.
The anomalies I was fixing this week where messing up the output of
the data; and that was true of both large data sets and small data
sets. People do get confused when I jump in to fix something when that
conversation started out as a discussion of a short list. I'm usually
not trying to fix the short list results. Rather I see something that
makes me realize there is a problem with the core approach for it's
intended goal.
do the right thing from a fuzzy logic project management task break
down approach.
Let's have a little philosophy of the methods discussion I think that
will help you see what CSA does what it does.
We have 2 scheme's
(a) Hierarchal is a method that uses and arithmetic progression down
the tree using addition. That method assumes that all tasks are rated
against the universe at large on a fixed scale. This is traditional
prioritization with a few necessities for making mass changes and
boost whole groups with the goals functions and a little hierarchal
smoothing thrown in. This method is great if you have less than 200
tasks and if you are disciplined and consistent I bet you can do 500
without tiring of the effort of prioritizing correctly. This method is
Andrey's baby and it works great for what it was designed to do. I
like it! and so do a lot of people.
(b) CSA is a method, that use an arithmetic progress down the tree
using multiplication of logarithmic reversible number pairs to
calculate a relative priority based on minimal data entered around a
localized position in the tree. (sounds sexy doesn't it? or just Bs?
actually it just some math theory that happens to be pretty it's a
GLOB Sorter if you want to get technical it cluster "like data" into
groups of similar values ). Under this model you set each tasks
Importance and urgency relative to it's immediate parent only. How
important and how urgent is this individual task to completing the
parent task; and only the parent task; not the project at a whole,
that's the KEY the parent task only. That allows for faster data
entry within huge outlines with 500 to 5000s of tasks. Because you
don't have to evaluate the task against your whole life; just it's
importance and urgency to the parent task, and when it's do. That is
localized positioning. IF you set your values that way the CSA will
give you very accurate results for priorities. I know I've been using
it for almost 10 years as lifebalance uses a simpler form of this
approach and I started on that tool in 98. This scheme is designed
specifically for people that have to make decisions about what gets
done AND what does NOT get done. Just because it's due today doesn't
mean it should be done. If figuring out which tasks should even be
reviewed on a give day is a challenge, then CSA is the method you
want. The CSA gets you a nice list of likely suspects to review. This
lines up nicely with GTD that says to own you own intuitive
prioritizations, so we often recommend CSA to GTDers' because it make
a first WHACK at you list for you; and reduces the number of items you
have to consider for you final selection of the correct task to do.
The problems usually creep in when people try to use CSA in a manner
other than intended; it will not make your decisions for you and it
won't process a really short list all that effectively that's why tiny
short lists give weird results; it wasn't designed to do what people
often try to test. It's also not a GANTT chart and it won't schedule
time linear linked tasks; if you need that see MS project and numerous
other tools or fall back to Hierarchal. CSA will always get the top
15-25 things to do in the right cluster at the top of the list out of
1000s of tasks. That's what it's designed to do. Get you a todo list
where the top screen without scrolling down at all has the things that
should be review and action'ed as necessary. The order of that screen
will never be perfect because only your intuition at the time of
choosing will tell you which of the top 15 things is the right one to
do right now right here.
That's what the method does. It really can't be bent to do other
things. But people loose site of that and start to blend the two
different methods characteristics. I you expect the computed todo list
to be ordered 1, 2, 3 ,4 exactly like you are expecting it you will be
disappointed. Don't pound nails with a screw driver; use Hierarchal
instead.
The anomalies I was fixing this week where messing up the output of
the data; and that was true of both large data sets and small data
sets. People do get confused when I jump in to fix something when that
conversation started out as a discussion of a short list. I'm usually
not trying to fix the short list results. Rather I see something that
makes me realize there is a problem with the core approach for it's
intended goal.
Richard Collings
Jul 16, 2009, 3:53:26 AM7/16/09
to myLifeO...@googlegroups.com
Thanks again for the detailed reply. Sadly I have tried both schemes and
can't make either work for me. And there is a steady stream (trickle?) of
other people posting similar comments.
Although the last time, I tried CSA I was also using the Weekly Goal which
from what you have said, distorts the behaviour significantly.
I would definitely put myself in the 'too many tasks' category so perhaps I
should go back and try it again
I am not sure that I understand the math. I tried Googling for GLOB sorter
and couldn't find anything. If the CSA is based on a more widely used set
of theories I would be interested to read a bit more - do you have any
references?
Would I be right in thinking that what you have as your top level tasks is
quite important to the CSA. At the moment, I start off with a Home/Work
split and then split each of these into things like Single Step Actions,
Daily Routines, etc.
What do you have or would you advise to have as the top level tasks and what
principles would you use in terms of how you organise tasks under this. Is
there a template that works well with CSA?
And given what you say about the CSA not being suited to sorting things down
to the level of individual tasks, it maybe points back to the need for a
layering a manual sort, which I desparately need, on top of this (and I
believe Andrey is thinking about) - ie: you use CSA to bring the most
important stuff to the top and then use manual sort to put into an order in
which you want to tackle things today.
Thanks again for taking the time to post.
Richard
> -----Original Message-----
> From: myLifeO...@googlegroups.com
> [mailto:myLifeO...@googlegroups.com] On Behalf Of ratz
> Sent: 16 July 2009 4:02 a
> To: MyLifeOrganized
> Subject: [MLO] Re: Computed-Score Priority and Start Dates
>
>
>
can't make either work for me. And there is a steady stream (trickle?) of
other people posting similar comments.
Although the last time, I tried CSA I was also using the Weekly Goal which
from what you have said, distorts the behaviour significantly.
I would definitely put myself in the 'too many tasks' category so perhaps I
should go back and try it again
I am not sure that I understand the math. I tried Googling for GLOB sorter
and couldn't find anything. If the CSA is based on a more widely used set
of theories I would be interested to read a bit more - do you have any
references?
Would I be right in thinking that what you have as your top level tasks is
quite important to the CSA. At the moment, I start off with a Home/Work
split and then split each of these into things like Single Step Actions,
Daily Routines, etc.
What do you have or would you advise to have as the top level tasks and what
principles would you use in terms of how you organise tasks under this. Is
there a template that works well with CSA?
And given what you say about the CSA not being suited to sorting things down
to the level of individual tasks, it maybe points back to the need for a
layering a manual sort, which I desparately need, on top of this (and I
believe Andrey is thinking about) - ie: you use CSA to bring the most
important stuff to the top and then use manual sort to put into an order in
which you want to tackle things today.
Thanks again for taking the time to post.
Richard
> -----Original Message-----
> From: myLifeO...@googlegroups.com
> [mailto:myLifeO...@googlegroups.com] On Behalf Of ratz
> Sent: 16 July 2009 4:02 a
> To: MyLifeOrganized
> Subject: [MLO] Re: Computed-Score Priority and Start Dates
>
>
>
Toes_NZ
Jul 16, 2009, 7:03:07 AM7/16/09
to MyLifeOrganized
gidday all
i am not an expert on all this but seem to have it working quite well
i am using computed score, by urgency, & importance
i have due and start date at the minimum settings
i mainly use the importance slider, usually on the project heading
which weights all the sub tasks
the say if i have a project with a deadline that is urgent also, i
ramp up the urgency to suit
i came across a website that explained the Eisenhower method, which
makes sense to me with a lot of stuff hitting my desk every day and
having to keep focus on the stuff that is important
basically to work effectivly, the task / project should have normal
urgency and be ranked by importance mainly
if a job is urgent and important, you are fighting fires [which i do
from time to time!]
just google it, there is stacks of info on the web about it
i find gtd is a little to perfect world for me, some parts are great,
and other parts veers a little on the micro managing side
with tasks i believe the 80/20 rule applies, 20% of your tasks are
truly important and create 80% of your productivity
the other 80%, probably should be delegated, or removed
mlo is still my favorite task manager and just seems to work better
the more i learn about it's features
i have quite a few custom views setup
the main ones i use are Due today, due in the past, & my dit short
list [do it today]
i find my productivity levels are very high working in an environment
where i will never get to the bottom of the pile
anyway, i'm off the beaten track
cheers
toes
i am not an expert on all this but seem to have it working quite well
i am using computed score, by urgency, & importance
i have due and start date at the minimum settings
i mainly use the importance slider, usually on the project heading
which weights all the sub tasks
the say if i have a project with a deadline that is urgent also, i
ramp up the urgency to suit
i came across a website that explained the Eisenhower method, which
makes sense to me with a lot of stuff hitting my desk every day and
having to keep focus on the stuff that is important
basically to work effectivly, the task / project should have normal
urgency and be ranked by importance mainly
if a job is urgent and important, you are fighting fires [which i do
from time to time!]
just google it, there is stacks of info on the web about it
i find gtd is a little to perfect world for me, some parts are great,
and other parts veers a little on the micro managing side
with tasks i believe the 80/20 rule applies, 20% of your tasks are
truly important and create 80% of your productivity
the other 80%, probably should be delegated, or removed
mlo is still my favorite task manager and just seems to work better
the more i learn about it's features
i have quite a few custom views setup
the main ones i use are Due today, due in the past, & my dit short
list [do it today]
i find my productivity levels are very high working in an environment
where i will never get to the bottom of the pile
anyway, i'm off the beaten track
cheers
toes
chuckdevee
Jul 16, 2009, 7:10:36 AM7/16/09
to MyLifeOrganized
thanks for your work on CSA Bob.. and for the explanation.
I think this is a really important part of MLO to get right..
I like the concept behind CSA but just think it needs to be simplified
so that users can better understand its behaviour..
mainly, I'd argue for taking out any additional boosts - either from
weekly goals or when a goal approaches its deadline..
Beyond that, I think it might be worth providing some more explanation
(minus the maths) about how the underlying concepts work, in
particular, the importance of localised scoring and the waterfall
effect whereby settings of parent tasks affect sub-tasks..And then,
perhaps some guidance on the sliders.. as these are subjective
measures.. I mean, what exactly does importance actually mean in
relation to a task?
I don't really use the urgency slider but here's how I use the
importance slider for CSA...
A neutral value (MID-POINT) means that task MUST to be done in order
to complete its parent. Tasks that are not essential get scored either
one or two notches below. For the remainder, as all these must be done
in order to complete the parent, in theory they are all equally
important for that parent task. However, those tasks that have a
positive impact on other tasks/goals/aims beyond the parent get scored
a notch or two higher, depending on how significant this impact might
be. Eg if I'm writing a few functions for a programme but one of them
could be really useful elsewhere, then I give it above neutral
importance. I find that if I use this method, it gives me a reaonably
consistent scoring logic for importance across tasks. Does this fit
with your view of how scoring should be used with CSA?
I think this is a really important part of MLO to get right..
I like the concept behind CSA but just think it needs to be simplified
so that users can better understand its behaviour..
mainly, I'd argue for taking out any additional boosts - either from
weekly goals or when a goal approaches its deadline..
Beyond that, I think it might be worth providing some more explanation
(minus the maths) about how the underlying concepts work, in
particular, the importance of localised scoring and the waterfall
effect whereby settings of parent tasks affect sub-tasks..And then,
perhaps some guidance on the sliders.. as these are subjective
measures.. I mean, what exactly does importance actually mean in
relation to a task?
I don't really use the urgency slider but here's how I use the
importance slider for CSA...
A neutral value (MID-POINT) means that task MUST to be done in order
to complete its parent. Tasks that are not essential get scored either
one or two notches below. For the remainder, as all these must be done
in order to complete the parent, in theory they are all equally
important for that parent task. However, those tasks that have a
positive impact on other tasks/goals/aims beyond the parent get scored
a notch or two higher, depending on how significant this impact might
be. Eg if I'm writing a few functions for a programme but one of them
could be really useful elsewhere, then I give it above neutral
importance. I find that if I use this method, it gives me a reaonably
consistent scoring logic for importance across tasks. Does this fit
with your view of how scoring should be used with CSA?
RichardCollings
Jul 16, 2009, 8:50:32 AM7/16/09
to MyLifeOrganized
And I guess the other question that I don't thinkn your response
answers is the question of whether when you boost a parent task
Importance or Urgency, whether that boost gets applied recursively as
you pass down the tree.
If it does, I cannot just get my head around that - all that you have
done is say - 'all these tasks under these parents are now more
imporant relative to other tasks elswhere in the hierarchy'. Why
should altering the priority of the parent, boost lower level tasks
more than higher level tasks. Surely all the bottom level tasks
should remain in the same relationship to each other according to the
urgency/importance settings applied to them and to their immediate
parents?
> > > > If I boost- Hide quoted text -
>
> - Show quoted text -...
>
> read more »
answers is the question of whether when you boost a parent task
Importance or Urgency, whether that boost gets applied recursively as
you pass down the tree.
If it does, I cannot just get my head around that - all that you have
done is say - 'all these tasks under these parents are now more
imporant relative to other tasks elswhere in the hierarchy'. Why
should altering the priority of the parent, boost lower level tasks
more than higher level tasks. Surely all the bottom level tasks
should remain in the same relationship to each other according to the
urgency/importance settings applied to them and to their immediate
parents?
>
> - Show quoted text -...
>
> read more »
chuckdevee
Jul 16, 2009, 9:59:12 AM7/16/09
to MyLifeOrganized
I guess I can understand that one, if indeed it is the case that lower
level tasks get an extra boost..
Imagine 2 projects with the same importance/urgency scores... the CSA
will work to try to get both done at the same time, other things
equal.
So if you have 5 levels of subtasks for Project A, and only 3 for
Project B, you will tend to see more Project A tasks cropping up until
you are at broadly equal levels of depth..
And if you feel that Project B is more important, you can presumably
negate/lessen this effect by giving it a higher importance score..so
that it gets done ahead of Project A. How well this works depends on
the extent of the recursive boost to lower level tasks.. I might try
a test to see how this would actually work..
level tasks get an extra boost..
Imagine 2 projects with the same importance/urgency scores... the CSA
will work to try to get both done at the same time, other things
equal.
So if you have 5 levels of subtasks for Project A, and only 3 for
Project B, you will tend to see more Project A tasks cropping up until
you are at broadly equal levels of depth..
And if you feel that Project B is more important, you can presumably
negate/lessen this effect by giving it a higher importance score..so
that it gets done ahead of Project A. How well this works depends on
the extent of the recursive boost to lower level tasks.. I might try
a test to see how this would actually work..
RichardCollings
Jul 16, 2009, 10:16:47 AM7/16/09
to MyLifeOrganized
It just doesn't work for me - the fact that you have broken one
activity down into more steps and to a deeper level doesn't
automatically make those individual tasks more important if you boost
their common parent. If they are more important/urgent, I can go
in and make adjustments at that level, I don't want MLO )or anything
else) doing it for me.
For me, boosting a parent's urgency/importance should leave the
relative ordering of the tasks under that parent exactly as it was.
All the tasks should move up the overall list together but keep their
ordering as before.
activity down into more steps and to a deeper level doesn't
automatically make those individual tasks more important if you boost
their common parent. If they are more important/urgent, I can go
in and make adjustments at that level, I don't want MLO )or anything
else) doing it for me.
For me, boosting a parent's urgency/importance should leave the
relative ordering of the tasks under that parent exactly as it was.
All the tasks should move up the overall list together but keep their
ordering as before.
scoobie
Jul 16, 2009, 2:41:53 PM7/16/09
to MyLifeOrganized
Bob,
What's your take on being able to do all this in an iphone sized
processor and screen?
Do you think its possible?
> ...
>
> read more »
What's your take on being able to do all this in an iphone sized
processor and screen?
Do you think its possible?
>
> read more »
Vallon, Justin
Jul 16, 2009, 5:06:51 PM7/16/09
to myLifeO...@googlegroups.com
Bob,
You might have considered this, but if you are dynamically computing the
scoring, you could try a cached-value approach, where you cache the
calculation of the score, and some operations (adjusting inputs,
reparenting, modifying weighting of parent) would invalidate the
calculation. If you are lazy about recalculating the score, then it
would be no slower than now on display (with a speedup after being
computed once), for some additional cost when you have to invalidate the
tree (when the root is modified or reparented).
Of course, this is a space-for-time tradeoff, and the devices are
memory-constrained.
-Justin
-----Original Message-----
From: myLifeO...@googlegroups.com
[mailto:myLifeO...@googlegroups.com] On Behalf Of scoobie
Sent: Thursday, July 16, 2009 2:42 PM
To: MyLifeOrganized
Subject: [MLO] Re: Computed-Score Priority and Start Dates
ratz
Jul 16, 2009, 10:31:52 PM7/16/09
to MyLifeOrganized
Actually there isn't only a fraction of the users that report
struggles because only a fraction of the license sold lead to someone
posting hear in the forums. What the silent majority has for
experiences we can only guess, and unfortunately if you are on this
list you are a cogg in this highly statistically schewd crowd. So
welcome to the funny farm as they say :) :)
>
> What do you have or would you advise to have as the top level tasks and what
> principles would you use in terms of how you organise tasks under this. Is
> there a template that works well with CSA?
>
I did 3/4 of the templates so here's as authoritative answer for you:
CSA: All the GTD ones.
Heriarchial:
Traditional FranlinkCovey
FlyLady
MLO Demo
Do it Tomorrow - don't know I assume Heirarchial
On Jul 16, 2:53 am, "Richard Collings" <r...@rcollings.co.uk> wrote:
> Thanks again for the detailed reply. Sadly I have tried both schemes and
> can't make either work for me. And there is a steady stream (trickle?) of
> other people posting similar comments.
>
> Although the last time, I tried CSA I was also using the Weekly Goal which
> from what you have said, distorts the behaviour significantly.
>
> I would definitely put myself in the 'too many tasks' category so perhaps I
> should go back and try it again
>
I would suggest using the "reset all tasks to normal urgency and
importance" button and starting over ranking things as needed. This
button exists for 2 reason; if you come from the heirachial method we
recommend you reset and start over. Secondly people get confused and
then I say "press RESET" you didn't understand the premise. (((yes
someday I need to write a tutorial, but this darn thing started life
as a power tool for the geeks on this list, it's not my fault Andrey
built a great app that happens to sell really well...)))
My #1 piece of advice is never change the Importance and Urgency of a
TASK based on what you see in the Todo List; Only do that from the
Outline when you are looking at the whole picture. If I could have the
sliders disabled in the Todo List view; I would in a heart beat. You'd
all whine, complain and hold your breath; but everyone would have far
better results. If you think the todo list is in the wrong order, then
there is the problem it is in the Outline not the TASK that's in the
wrong place. If you can't resist tweaking individual tasks in the todo
list don't use this method. Seriously don't do it
I have "triaged" a number of users files over the years and without
exception if it doesn't work it's because they put bad data in and
then try to game the system. The program gives them an exact result of
garbage and they are surprised by the result in-spite of feeding it
garbage. The biggest problem is they really refuse to rank items IN
RESPECT to the parent only. You have to do that or it won't work. I
find lots of people are ranking the siblings against each other; that
is wrong.
Look we could go on and on about this for hours of examples but, there
is only so much time in the day;
If you have the following outline below (yes it's way too short). If
we exclude everything except importance for the moment.
Project A
++Task A
++++SubTask A
++++++SubSubTask A
++++++++SubSubSubTask A
++++SubTask B
++++++SubSubTask B
++++++++SubSubSubTask B
Project B
++Task B
++++SubTask C
++++SubTask D
++Task C
++++SubTask E
++++++SubSubTask C
++++++++SubSubSubTask C
The Only Items that must have there importance set are the ones with *
Project A*
++Task A
++++SubTask A*
++++++SubSubTask A
++++++++SubSubSubTask A
++++SubTask B*
++++++SubSubTask B
++++++++SubSubSubTask B
Project B*
++Task B*
++++SubTask C*
++++SubTask D*
++Task C*
++++SubTask E
++++++SubSubTask C
++++++++SubSubSubTask C
That's it; if you set the importance of those you get a valid result.
The mistake people make is how they do that. Let's look at 2 cases.
Case (1) (SubTask C & SubTask D)
The mistake here is to say SubTask C is more important than SubTask D
So let's set C high and D low. That is in not correct. You have to
Decide how important SubTask C is to Task B, and how important SubTask
D is to Task B and you only have to do that because more than 1
SubTask X exists in Task B. Those are different questions completely.
If you don't understand that keep reading that sentence until you do.
Still here? really? Awesome, here's the same idea in a concrete
example:
Think of it like cleaning a room. If you make a list of ten things to
do to clean the room, let's assume 5 of those are probably very
important to cleaning the room; if you don't get them done the room
isn't clean. The other 5 are optional if you don't do them the room is
still clean and you can call it good enough. The first 5 are then very
important to the Parent and the other 5 are not. If only the first 5
are present and the other 5 don't exist then the critical 5 are
"normal" importance to the parent. The first 5 are very because the
noncritical 5 EXISIT; but NOT RELATIVE to them. You rank then IN
RESPECT to the PARENT. What that means in our example is: If the only
the second five existed then they would all be "normal" importance IN
RESPECT the parent, and the Parent "clean the room" would in all
likelihood be less important IN RESPECT to its parent IF it had
sibling items within that parent. Ok backup and read this paragraph
again 3 more times; it's a very very hard concept to grasp, but if you
are struggling if you can grok it; you might get yourself over the
hurdle. What are you waiting for go back and read it again.
Case (2) Why is is subsubsubTask C more important than SubTask D; I
must change it.... NOPE. SubSubSubTask C will be more important if
what you said of above it computes it to be that way. if it's wrong to
your intuition go to the outline and review your outline the problem
is elsewhere because SubSubSubTask C has no siblings and therefore
should be NORMAL importance. (Cavet if SubSubSubTask C had previous
siblings it might be something other than normal). In the virgin case
above though there are 3 tasks that must be completed to get back up
the tree to Task C Therefore If you make Task C important or then all
the tasks below it have to get our of the way to get back up to it;
the more there are the more important they become because the road
block is thicker.
Another good trick if you want to learn that the system works; it to
only use Dates for a while. If you are in a new outline set only due
dates for awhile and watch the results. If REV3 was deployed I would
suggest both start and due dates. If you have an existing outline you
can save a copy and then use the "reset importance" button; and look
at what your due date date is telling you. Most people can get their
due dates right.
> I am not sure that I understand the math. I tried Googling for GLOB sorter
> and couldn't find anything. If the CSA is based on a more widely used set
> of theories I would be interested to read a bit more - do you have any
> references?
Ah sorry GLOB sorter is a term from Chaotic Mathematics and Number
Theory, I think one of my profs made it up 25 years ago; some very odd
stuff. If you chart natural scientific data; things tend to cluster in
storm cells are points of interest where near by points are similar in
the characteristic being modeled. Many, nonlinear chaotic oscillating
function do that. What the algorithm is doing is taking a finite set
of data; and trying to figure out which things in the outline are
important and urgent; It's basically mathematically doing covey's 4
quadrants for you. Most everything the algorithm does can be found in
a Nonlinear mathematics text book; or a good computational math book
would do, I doubt you'll find them combined in one spot. I use
http://www-cs-faculty.stanford.edu/~knuth/gkp.html for most fancy
things I do. The best self regulating qsort with an embedded shell
sort can be derived from that book; but you have to know where to
look.
>
> Would I be right in thinking that what you have as your top level tasks is
> quite important to the CSA. At the moment, I start off with a Home/Work
> split and then split each of these into things like Single Step Actions,
> Daily Routines, etc.
What matters is the first level at which you move the sliders out of
the central position; OR assign a Due Date to an item; From that point
down the algorithm is morphing the data.
>
> And given what you say about the CSA not being suited to sorting things down
> to the level of individual tasks, it maybe points back to the need for a
> layering a manual sort, which I desparately need, on top of this (and I
> believe Andrey is thinking about) - ie: you use CSA to bring the most
> important stuff to the top and then use manual sort to put into an order in
> which you want to tackle things today.
>
Manual sorting on top; I've looked at that twice and all
implementations are ugly due to the need to reset at some point; when
is the right time without loosing the data. It's a real briar patch.
But Andrey may have other thoughts. As for having things in a precise
order for the day? I suggest that people really try and get beyond
that psychological itch; nobody can maintain that in today world. I
know I'm sure that rubs a few people; but I'm old enough now to get
away with it. But I've got a challenge for everyone that disagrees.
Tomorrow take you top ten things you need to do. Write them on an
index card in the order you think you need to do them; then put that
card in a drawer, now right down your top ten things you need to do;
on another card; no read the whole list and pick one; do it; when you
are done cross it off and number item 1. Read read the WHOLE list and
pick the one that feels right and do; when done cross it off and
number it item 2; repeat until done; when you are finished compare the
order of the two cards. If you find that interesting then join the
thread on autofocus. Most of my current energies are being used to see
if MLO can be a platform for that technique.
struggles because only a fraction of the license sold lead to someone
posting hear in the forums. What the silent majority has for
experiences we can only guess, and unfortunately if you are on this
list you are a cogg in this highly statistically schewd crowd. So
welcome to the funny farm as they say :) :)
>
> What do you have or would you advise to have as the top level tasks and what
> principles would you use in terms of how you organise tasks under this. Is
> there a template that works well with CSA?
>
CSA: All the GTD ones.
Heriarchial:
Traditional FranlinkCovey
FlyLady
MLO Demo
Do it Tomorrow - don't know I assume Heirarchial
On Jul 16, 2:53 am, "Richard Collings" <r...@rcollings.co.uk> wrote:
> Thanks again for the detailed reply. Sadly I have tried both schemes and
> can't make either work for me. And there is a steady stream (trickle?) of
> other people posting similar comments.
>
> Although the last time, I tried CSA I was also using the Weekly Goal which
> from what you have said, distorts the behaviour significantly.
>
> I would definitely put myself in the 'too many tasks' category so perhaps I
> should go back and try it again
>
importance" button and starting over ranking things as needed. This
button exists for 2 reason; if you come from the heirachial method we
recommend you reset and start over. Secondly people get confused and
then I say "press RESET" you didn't understand the premise. (((yes
someday I need to write a tutorial, but this darn thing started life
as a power tool for the geeks on this list, it's not my fault Andrey
built a great app that happens to sell really well...)))
My #1 piece of advice is never change the Importance and Urgency of a
TASK based on what you see in the Todo List; Only do that from the
Outline when you are looking at the whole picture. If I could have the
sliders disabled in the Todo List view; I would in a heart beat. You'd
all whine, complain and hold your breath; but everyone would have far
better results. If you think the todo list is in the wrong order, then
there is the problem it is in the Outline not the TASK that's in the
wrong place. If you can't resist tweaking individual tasks in the todo
list don't use this method. Seriously don't do it
I have "triaged" a number of users files over the years and without
exception if it doesn't work it's because they put bad data in and
then try to game the system. The program gives them an exact result of
garbage and they are surprised by the result in-spite of feeding it
garbage. The biggest problem is they really refuse to rank items IN
RESPECT to the parent only. You have to do that or it won't work. I
find lots of people are ranking the siblings against each other; that
is wrong.
Look we could go on and on about this for hours of examples but, there
is only so much time in the day;
If you have the following outline below (yes it's way too short). If
we exclude everything except importance for the moment.
Project A
++Task A
++++SubTask A
++++++SubSubTask A
++++++++SubSubSubTask A
++++SubTask B
++++++SubSubTask B
++++++++SubSubSubTask B
Project B
++Task B
++++SubTask C
++++SubTask D
++Task C
++++SubTask E
++++++SubSubTask C
++++++++SubSubSubTask C
The Only Items that must have there importance set are the ones with *
Project A*
++Task A
++++SubTask A*
++++++SubSubTask A
++++++++SubSubSubTask A
++++SubTask B*
++++++SubSubTask B
++++++++SubSubSubTask B
Project B*
++Task B*
++++SubTask C*
++++SubTask D*
++Task C*
++++SubTask E
++++++SubSubTask C
++++++++SubSubSubTask C
That's it; if you set the importance of those you get a valid result.
The mistake people make is how they do that. Let's look at 2 cases.
Case (1) (SubTask C & SubTask D)
The mistake here is to say SubTask C is more important than SubTask D
So let's set C high and D low. That is in not correct. You have to
Decide how important SubTask C is to Task B, and how important SubTask
D is to Task B and you only have to do that because more than 1
SubTask X exists in Task B. Those are different questions completely.
If you don't understand that keep reading that sentence until you do.
Still here? really? Awesome, here's the same idea in a concrete
example:
Think of it like cleaning a room. If you make a list of ten things to
do to clean the room, let's assume 5 of those are probably very
important to cleaning the room; if you don't get them done the room
isn't clean. The other 5 are optional if you don't do them the room is
still clean and you can call it good enough. The first 5 are then very
important to the Parent and the other 5 are not. If only the first 5
are present and the other 5 don't exist then the critical 5 are
"normal" importance to the parent. The first 5 are very because the
noncritical 5 EXISIT; but NOT RELATIVE to them. You rank then IN
RESPECT to the PARENT. What that means in our example is: If the only
the second five existed then they would all be "normal" importance IN
RESPECT the parent, and the Parent "clean the room" would in all
likelihood be less important IN RESPECT to its parent IF it had
sibling items within that parent. Ok backup and read this paragraph
again 3 more times; it's a very very hard concept to grasp, but if you
are struggling if you can grok it; you might get yourself over the
hurdle. What are you waiting for go back and read it again.
Case (2) Why is is subsubsubTask C more important than SubTask D; I
must change it.... NOPE. SubSubSubTask C will be more important if
what you said of above it computes it to be that way. if it's wrong to
your intuition go to the outline and review your outline the problem
is elsewhere because SubSubSubTask C has no siblings and therefore
should be NORMAL importance. (Cavet if SubSubSubTask C had previous
siblings it might be something other than normal). In the virgin case
above though there are 3 tasks that must be completed to get back up
the tree to Task C Therefore If you make Task C important or then all
the tasks below it have to get our of the way to get back up to it;
the more there are the more important they become because the road
block is thicker.
Another good trick if you want to learn that the system works; it to
only use Dates for a while. If you are in a new outline set only due
dates for awhile and watch the results. If REV3 was deployed I would
suggest both start and due dates. If you have an existing outline you
can save a copy and then use the "reset importance" button; and look
at what your due date date is telling you. Most people can get their
due dates right.
> I am not sure that I understand the math. I tried Googling for GLOB sorter
> and couldn't find anything. If the CSA is based on a more widely used set
> of theories I would be interested to read a bit more - do you have any
> references?
Theory, I think one of my profs made it up 25 years ago; some very odd
stuff. If you chart natural scientific data; things tend to cluster in
storm cells are points of interest where near by points are similar in
the characteristic being modeled. Many, nonlinear chaotic oscillating
function do that. What the algorithm is doing is taking a finite set
of data; and trying to figure out which things in the outline are
important and urgent; It's basically mathematically doing covey's 4
quadrants for you. Most everything the algorithm does can be found in
a Nonlinear mathematics text book; or a good computational math book
would do, I doubt you'll find them combined in one spot. I use
http://www-cs-faculty.stanford.edu/~knuth/gkp.html for most fancy
things I do. The best self regulating qsort with an embedded shell
sort can be derived from that book; but you have to know where to
look.
>
> Would I be right in thinking that what you have as your top level tasks is
> quite important to the CSA. At the moment, I start off with a Home/Work
> split and then split each of these into things like Single Step Actions,
> Daily Routines, etc.
the central position; OR assign a Due Date to an item; From that point
down the algorithm is morphing the data.
>
> And given what you say about the CSA not being suited to sorting things down
> to the level of individual tasks, it maybe points back to the need for a
> layering a manual sort, which I desparately need, on top of this (and I
> believe Andrey is thinking about) - ie: you use CSA to bring the most
> important stuff to the top and then use manual sort to put into an order in
> which you want to tackle things today.
>
implementations are ugly due to the need to reset at some point; when
is the right time without loosing the data. It's a real briar patch.
But Andrey may have other thoughts. As for having things in a precise
order for the day? I suggest that people really try and get beyond
that psychological itch; nobody can maintain that in today world. I
know I'm sure that rubs a few people; but I'm old enough now to get
away with it. But I've got a challenge for everyone that disagrees.
Tomorrow take you top ten things you need to do. Write them on an
index card in the order you think you need to do them; then put that
card in a drawer, now right down your top ten things you need to do;
on another card; no read the whole list and pick one; do it; when you
are done cross it off and number item 1. Read read the WHOLE list and
pick the one that feels right and do; when done cross it off and
number it item 2; repeat until done; when you are finished compare the
order of the two cards. If you find that interesting then join the
thread on autofocus. Most of my current energies are being used to see
if MLO can be a platform for that technique.
ratz
Jul 16, 2009, 10:33:07 PM7/16/09
to MyLifeOrganized
Sounds to me like you are using it the way it should be....by avoiding
obsessing done at the lowest level details. Well done.
obsessing done at the lowest level details. Well done.
ratz
Jul 16, 2009, 10:34:37 PM7/16/09
to MyLifeOrganized
>Thanks again for the detailed reply.
Btw no problem, just don't mind my ton; I'm three days behind on some
obligation because a server when QA-Boom. And I really shouldn't be
spending any time here positng. But I need a distraction every 10
hours or so.
ratz
Jul 17, 2009, 12:17:47 AM7/17/09
to MyLifeOrganized
> I don't really use the urgency slider but here's how I use the
> importance slider for CSA...
> A neutral value (MID-POINT) means that task MUST to be done in order
> to complete its parent. Tasks that are not essential get scored either
> one or two notches below. For the remainder, as all these must be done
> in order to complete the parent, in theory they are all equally
> important for that parent task. However, those tasks that have a
> positive impact on other tasks/goals/aims beyond the parent get scored
> a notch or two higher, depending on how significant this impact might
> be. Eg if I'm writing a few functions for a programme but one of them
> could be really useful elsewhere, then I give it above neutral
> importance. I find that if I use this method, it gives me a reaonably
> consistent scoring logic for importance across tasks. Does this fit
> with your view of how scoring should be used with CSA?
Sorry I missed this question.; that's close but you should only rank
> importance slider for CSA...
> A neutral value (MID-POINT) means that task MUST to be done in order
> to complete its parent. Tasks that are not essential get scored either
> one or two notches below. For the remainder, as all these must be done
> in order to complete the parent, in theory they are all equally
> important for that parent task. However, those tasks that have a
> positive impact on other tasks/goals/aims beyond the parent get scored
> a notch or two higher, depending on how significant this impact might
> be. Eg if I'm writing a few functions for a programme but one of them
> could be really useful elsewhere, then I give it above neutral
> importance. I find that if I use this method, it gives me a reaonably
> consistent scoring logic for importance across tasks. Does this fit
> with your view of how scoring should be used with CSA?
tasks IN RESPECT to their immediate parent that's what the algorithm
expects. It's suppose to relieve you of the burden of thinking about
EVERY task in a GLOBAL context. If a Project contains items that are
important to other projects; i would make that project itself more
important. rather than the tasks within it.
ratz
Jul 17, 2009, 12:25:07 AM7/17/09
to MyLifeOrganized
> If it does, I cannot just get my head around that - all that you have
> done is say - 'all these tasks under these parents are now more
> imporant relative to other tasks elswhere in the hierarchy'. Why
> should altering the priority of the parent, boost lower level tasks
> more than higher level tasks. Surely all the bottom level tasks
> should remain in the same relationship to each other according to the
> urgency/importance settings applied to them and to their immediate
> parents?
almost the same value as their parents. If the sliders are in the
neutral position.
If dates are applied to any task; task below that will have slightly
increasing priority based on depth; the algorithm assumes more tasks
need to be done by the due date so you better get to work. As you
check tasks off, the next task up is less urgent, there's less to do
by the due date; but as the "day and time" move forward the entire
section of the outline becomes more urgent because you are getting
closer to the do date. So both ends push at each other when there are
dates involved.
We once had a user freak out because the priority changed each time he
hit update without changing any tasks. He failed to realize that he
set a due date and time and that he forgot to pause the universe
before hitting update. :) :)
I do think people will have less issues when and if REV 3 of CSA is
available. I did some spreadsheet calcs and I really think this
handles the NULL dates and start dates better. I'm glad I found the
time review it in light of the software changes at large. I really
should have found time before Andrey released 3.0; but that wasn't in
the cards.
ratz
Jul 17, 2009, 12:28:35 AM7/17/09
to MyLifeOrganized
Well I test 70 some iphone todo apps to figure out where to look for
ideas; I can tell you I think most iphone task apps suck because
people aren't thinking iphone; they think desktop app --> iphone. So
I'll be one snob on the team. Example I'd like to see major screen and
view changes handled like tweet deck for the iphone; that page
metophor is very iphone. Don't even get me started on LB for the
iphone or OF for the iphone. OMG overhead and too slow to get anything
done.
> ...
>
> read more »
ideas; I can tell you I think most iphone task apps suck because
people aren't thinking iphone; they think desktop app --> iphone. So
I'll be one snob on the team. Example I'd like to see major screen and
view changes handled like tweet deck for the iphone; that page
metophor is very iphone. Don't even get me started on LB for the
iphone or OF for the iphone. OMG overhead and too slow to get anything
done.
>
> read more »
ratz
Jul 17, 2009, 12:36:52 AM7/17/09
to MyLifeOrganized
Yeah it's not that bad; it's comp-cost equation.
The algorithm is truly recursive. So every IF or SUM that we can
removed is removed time the number of items in the outline. That's why
I don't like adding orthogonal feature. Each option that has to be
test globally mean and extra branch in the algorithm has to be run
through for every node in the task tree.
So you have to think smart. The current rev I was able to identify 5
items we use to calc on the fly because it made sense. I was able to
move those over it the Task OBJECT and handle the updating process in
the gui. That removed something like 12-20 operations from the loop
and that is per Node iteration; that's a ton; We also probably remove
15-20% of the main IF/Then branch with the new matrix and case
statement.
Part of this is that the algorithm got "extend" in version 1.5; and
should have been re-factored at REV2, there wasn't time to it go
bloated. It's now lean and mean and able to grow again; BUT we need to
try a running version of REV3 and have people see how it behaves as
designed before doing anything else.
This is all based on the desktop code of course; I've never seen the
source code for the PPC version; the team takes the desktop and
rewrites it for the PPC so they might have already made many of the
same improvements in the past; they weren't anything special they just
require groking the algorithm and having the time to dedicate to it.
The algorithm is truly recursive. So every IF or SUM that we can
removed is removed time the number of items in the outline. That's why
I don't like adding orthogonal feature. Each option that has to be
test globally mean and extra branch in the algorithm has to be run
through for every node in the task tree.
So you have to think smart. The current rev I was able to identify 5
items we use to calc on the fly because it made sense. I was able to
move those over it the Task OBJECT and handle the updating process in
the gui. That removed something like 12-20 operations from the loop
and that is per Node iteration; that's a ton; We also probably remove
15-20% of the main IF/Then branch with the new matrix and case
statement.
Part of this is that the algorithm got "extend" in version 1.5; and
should have been re-factored at REV2, there wasn't time to it go
bloated. It's now lean and mean and able to grow again; BUT we need to
try a running version of REV3 and have people see how it behaves as
designed before doing anything else.
This is all based on the desktop code of course; I've never seen the
source code for the PPC version; the team takes the desktop and
rewrites it for the PPC so they might have already made many of the
same improvements in the past; they weren't anything special they just
require groking the algorithm and having the time to dedicate to it.
chuckdevee
Aug 26, 2009, 9:31:39 AM8/26/09
to MyLifeOrganized
Ratz/Andrey, would it be possible to tell us when this upgrade is
likely to be released please? thanks
likely to be released please? thanks
Fletcher Kauffman
Apr 5, 2011, 6:21:38 PM4/5/11
to mylifeo...@googlegroups.com, MyLifeOrganized
I know this thread is a bit old, but there seems to be some issue I'm having rearing it's ugly head again.
I'm very interested in this aspect of MLO in particular-- the magic of a program like this is that it has the (potential) ability to answer the question "What should I be doing right now?" without my having to do any thinking.
I have the experience pretty often that (for whatever reason) when my Outline is sufficiently complex, my To Do list winds up incorrect.
I also want to touch on another point about MLO being like a light Project Management piece of software-- it has one "flaw" in this regard, which is that it treats Start and End as the same thing as when the task should appear in the list. I've struggled to understand this for a long time, and I finally resolved to just view it that start (in particular) merely dictates when it should show up on the list.
From that premise, the Start Date should have no impact on urgency or priority (from the aspect that MLO doesn't actually know anything about the task itself-- just how far in advance I needed to know about it).
MLO could make some "smarter" calculations by taking Min/Max times into account to do this, but we're then talking about a more complex Project-Oriented feature set.
I worked in MS Project (and other similar tools) for many, many years and I've been very pleased with MLO getting to close to that experience (especially cutting out all the stuff that makes MS Project unusable for certain things)-- it's when MLO gets right close to the edge of that featureset that it starts to wiggle out.
I'm keenly interested in steering MLO (or an offshoot) toward teams and groups-- a full-fledged, cross-platform, multi-user work-brokering system.
I had started looking at this a few years ago when I asked the question: "What if you used Life Balance at an organizational level?"
Right now, the priority/sorting algorithm is most holding us up-- I am working with two other stakeholders who are non-MLO users (getting slowly acquainted) and the biggest thing I'm having to defend/explain is that we seem to be setting priorities we all agree to in the Outline, but then the To Do list sort does not reflect those priorities.
I know it's been a few years-- any thoughts on this?
pottster
Apr 6, 2011, 2:16:12 AM4/6/11
to mylifeo...@googlegroups.com, MyLifeOrganized
For me, people get too hung up trying to optimize the sorting of tasks when in reality it's the filtering of tasks that is important. Most of the time a "definitive" sorting of tasks is out of date and ignored the moment it's compiled. A rough sorting is usually all that's required within a framework of contextual filtering. There are too many subjective inputs to achieve a rigid, all-inclusive, tightly sequenced to-do list; especially at the corporate level you are talking about. For example, changing priorities, personal energy levels, physical location, dependencies on external factors outide of your control, etc etc. An algorithm is just a tool to help YOU answer the question "what should I do next", a decision aided by filtering out the stuff you don't need to consider right now. That's still a good result. I don't expect automated Project Management anytime soon ;-)
Mario Seixas Sales
Sep 12, 2024, 3:21:54 AM9/12/24
to MyLifeOrganized
holy shit, those 'ratz' comments about CSA are gold
Reply all
Reply to author
Forward
0 new messages