Feature idea: Automatic clocking

Feature idea: Automatic clocking

[PT]

I just started using clocking and it seems really useful. It
occured me it could also be done automatically for certain tasks
which are performed in the org buffer.

For example, I work on some text which I keep in an org subtree,
the branches of the subtree hold the chapters, etc.

If the main subtree which is the root of the document has a CLOCK
property (put there by a previous manual clocking) and also an
AUTOCLOCK or similar property then it could monitor if I modify
the text within the subtree and start the clock automatically. If
I stop modifying the subtree then after a while (say, 30 seconds,
configurable) it would stop the clock automatically.

So for subtrees explicitly marked for automatic clocking the user
wouldn't have to start/stop the clock manually at all, org could
do it itself.

What do you think?

Re: Feature idea: Automatic clocking

[Bernt Hansen]

PT <spamfilteraccount@gmail.com> writes:

> I just started using clocking and it seems really useful. It
> occured me it could also be done automatically for certain tasks
> which are performed in the org buffer.
>
> For example, I work on some text which I keep in an org subtree,
> the branches of the subtree hold the chapters, etc.
>
> If the main subtree which is the root of the document has a CLOCK
> property (put there by a previous manual clocking) and also an
> AUTOCLOCK or similar property then it could monitor if I modify
> the text within the subtree and start the clock automatically. If
> I stop modifying the subtree then after a while (say, 30 seconds,
> configurable) it would stop the clock automatically.
>
> So for subtrees explicitly marked for automatic clocking the user
> wouldn't have to start/stop the clock manually at all, org could
> do it itself.
>
> What do you think?

Hi PT,

I've been using org-mode clocking since 2006-08-29 Tue 11:44 and I am
skeptical about how useful this would really be in the general case.

Most of my tasks involve *thinking* not just typing so stopping the
clock when I'm working on solving a problem would be bad.  I also clock
tasks while working on another machine which org-mode knows nothing
about so stopping the clock due to inactivity isn't appropriate.

I don't like the idea of automatic clocking for a number of reasons:

  - It lets you be sloppy about starting and stopping the clock -- which
    means the clock won't be running for some task you are working on
    (say one that is not marked for automatic clocking).  This means
    you're going to work on stuff and not have it clocked when you need
    it to be at some point.  I bill based on clock time and it needs to
    be correct.

  - Clocking stuff in and out rigorously is a good habit to learn if
    clock data is really important to you.  Automatic clocking defeats
    this goal.

  - If you're clocking some important project task and you happen to
    touch the task marked for automatic clocking you'll clock out the
    project task and clock in the new task... and a short time later the
    clock stops when you move back to the project task but you're still
    really working on that original project task.

Clocking the right task usually takes more intelligence than just what
part of an org-file changes.

I have org-clock-out-when-done set to nil so that org-mode does not stop
the clock when a task is marked DONE.  This makes me responsible for
when the clock starts and stops for all tasks - I clock in and out for
everything that matters.  I change the clock when I switch tasks and I
think it's really hard to get that right automatically.

So there's my two cents :)

Regards,
Bernt

Re: Feature idea: Automatic clocking

[Samuel Wales]

If something like this were implemented, I would definitely try it, as
a way to see where I am wasting effort etc. in the org hierarchy.  I
might even use it all the time, if it works.

However, I would probably want it to apply to all headlines above a
certain level, and measure all activity below each of those.

And I would want the clock information stored in a separate, central
location, instead of in each headline.

It is like profiling code.

-- 
Myalgic encephalomyelitis causes death (Jason et al. 2006)
and severe suffering.  Conflicts of interest are destroying
research. What people "know" is wrong. Silence = death.
http://www.meactionuk.org.uk/What_Is_ME_What_Is_CFS.htm

RE: Feature idea: Automatic clocking

[Nuxoll, Andrew]

[-- Attachment #1: Type: text/plain, Size: 2450 bytes --]

So, I dig the idea of automatic clocking.  I've been trying to use the manual clock feature and I haven't been able to train myself to do it very reliably.  I think the big problem is that I don't have a big stake in it.  I'm not billing different accounts for my time.  The only one who sees the clock data is me.  So, the motivation is low.

Here's an alternative implementation of autoclocking:
1.  You specify what hours you're typically working on tasks.  For example, if your org list was used it work this might be 9am-5pm except for lunch break: 12pm-1pm.  Alternatively, if it was your at home org list then it would be 6pm-10pm and also 7am-8am.  
2.  Whenever you mark a TODO item as DONE, the org-mode assumes that's all you've been working on since the last time you marked something DONE.

Obviously this is flawed in a number of ways, but it is easy to implement and I think you could get a reasonable measure of what you're spending your time on.  For those of us, like me, who don't have a strong motivation to clock our time that might be plenty sufficient.  The fact that automatic CLOCK: entries would appear when I marked an item as DONE would motivate me to go and edit the entry to be more accurate if it wasn't.  

I would really value a feature like this.

:AMN:




-----Original Message-----
Date: Fri, 21 Aug 2009 18:35:27 +0000 (UTC)
From: PT <spamfilteraccount@gmail.com>
Subject: [Orgmode] Feature idea: Automatic clocking
To: emacs-orgmode@gnu.org
Message-ID: <loom.20090821T202641-395@post.gmane.org>
Content-Type: text/plain; charset=us-ascii

I just started using clocking and it seems really useful. It
occured me it could also be done automatically for certain tasks
which are performed in the org buffer.

For example, I work on some text which I keep in an org subtree,
the branches of the subtree hold the chapters, etc.

If the main subtree which is the root of the document has a CLOCK
property (put there by a previous manual clocking) and also an
AUTOCLOCK or similar property then it could monitor if I modify
the text within the subtree and start the clock automatically. If
I stop modifying the subtree then after a while (say, 30 seconds,
configurable) it would stop the clock automatically.

So for subtrees explicitly marked for automatic clocking the user
wouldn't have to start/stop the clock manually at all, org could
do it itself.

What do you think?


[-- Attachment #2: winmail.dat --]
[-- Type: application/ms-tnef, Size: 4067 bytes --]

[-- Attachment #3: Type: text/plain, Size: 204 bytes --]

_______________________________________________
Emacs-orgmode mailing list
Remember: use `Reply All' to send replies to the list.
Emacs-orgmode@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: Feature idea: Automatic clocking

[Bastien]

Here is another approach to this auto-clocking question.

Maybe what you really want to track is not how much time you spend on a
specific task but rather how you distribute your time over the tasks.

In this case, you don't need to explicitely clock in and out.  You can
act as if a clock was always running¹ and then you register any action
you take on your Org file.  "Working on this X task" would just be an
action among others like "switching task X from TODO state to NEXT",
"closing task X", "Archiving task X", "Moving task X under Tree T", 
etc.

Thus, instead of automating the aggregation of clocked-in tasks, you
consider yourself always clocked in and you automate the logging of
stuff you do (from which you can extract clock information later.)

I guess this idea come closer to what Samuel had in mind when he talked
about "profiling".

¹ Actually, astronomers from this list could tell us that the universe
  is such a running clock in itself :)

-- 
 Bastien

Re: Re: Feature idea: Automatic clocking

[Carsten Dominik]

Hi

I am also very skeptical to automate something like this.  This is
a typical are where it is all about the discipline to establish a habit.
May experience is that if you cannot find the discipline to do it,
an automatic process might does *something* for you, but you are
not going to make any use of the collected data.


:-(

- Carsten

On Aug 21, 2009, at 8:26 PM, Bernt Hansen wrote:

> PT <spamfilteraccount@gmail.com> writes:
>
>> I just started using clocking and it seems really useful. It
>> occured me it could also be done automatically for certain tasks
>> which are performed in the org buffer.
>>
>> For example, I work on some text which I keep in an org subtree,
>> the branches of the subtree hold the chapters, etc.
>>
>> If the main subtree which is the root of the document has a CLOCK
>> property (put there by a previous manual clocking) and also an
>> AUTOCLOCK or similar property then it could monitor if I modify
>> the text within the subtree and start the clock automatically. If
>> I stop modifying the subtree then after a while (say, 30 seconds,
>> configurable) it would stop the clock automatically.
>>
>> So for subtrees explicitly marked for automatic clocking the user
>> wouldn't have to start/stop the clock manually at all, org could
>> do it itself.
>>
>> What do you think?
>
> Hi PT,
>
> I've been using org-mode clocking since 2006-08-29 Tue 11:44 and I am
> skeptical about how useful this would really be in the general case.
>
> Most of my tasks involve *thinking* not just typing so stopping the
> clock when I'm working on solving a problem would be bad.  I also  
> clock
> tasks while working on another machine which org-mode knows nothing
> about so stopping the clock due to inactivity isn't appropriate.
>
> I don't like the idea of automatic clocking for a number of reasons:
>
>  - It lets you be sloppy about starting and stopping the clock --  
> which
>    means the clock won't be running for some task you are working on
>    (say one that is not marked for automatic clocking).  This means
>    you're going to work on stuff and not have it clocked when you need
>    it to be at some point.  I bill based on clock time and it needs to
>    be correct.
>
>  - Clocking stuff in and out rigorously is a good habit to learn if
>    clock data is really important to you.  Automatic clocking defeats
>    this goal.
>
>  - If you're clocking some important project task and you happen to
>    touch the task marked for automatic clocking you'll clock out the
>    project task and clock in the new task... and a short time later  
> the
>    clock stops when you move back to the project task but you're still
>    really working on that original project task.
>
> Clocking the right task usually takes more intelligence than just what
> part of an org-file changes.
>
> I have org-clock-out-when-done set to nil so that org-mode does not  
> stop
> the clock when a task is marked DONE.  This makes me responsible for
> when the clock starts and stops for all tasks - I clock in and out for
> everything that matters.  I change the clock when I switch tasks and I
> think it's really hard to get that right automatically.
>
> So there's my two cents :)
>
> Regards,
> Bernt
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Remember: use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode