This is my focus protocol. Whenever I find myself having trouble trying started on a task, I create a new desktop and open windows related to the task only. DnD on. Pick a next step. Execute.
I've found mostly the opposite. Some well arranged windows are quite a nice anchor, I'm working on what's there in front of me. It's like bowling with bumpers in place, instead of the ball going in the gutter, the structure keeps it in the lane. I've found it necessary to devote time to cleaning and clearing windows, and sometimes I forget what's going on, and as I'm closing out the windows because I forgot what was going on, oh! there's this half finished thing that I actually really want finished.
What am I working on, what's in progress? The work space is the map. The terrain is changing as the task progresses, and so must the map, but the map is useful, even if it takes a bit of redrawing here and there.
The desktops (multiple, 3-7) are the map of the work. Part of the work is keeping the map accurate, not wadding it up and throwing it in the trash.
I suppose different things work for different people, but I started with the suggestion here and came around to skillful use of space as the work map itself.
Cleaning and updating are continuous, not a 'big bang' clear-the-desks event, mostly. But if it's not continuous, the big bang is probably better.
Some spots are problem spots, like digital notebooks, desktop icons. When I notice a problem spot, I create a recurring task to remove one X per week, or in some of the worst cases, one X per day. I have a rule of clearing out the oldest two days of email each day. I miss some days if I'm busy, but on average rate out = rate in, because I will always catch up within a day or two applying the rule that the oldest two days of email need eviction (make a task out of it, archive it, whatever) every day. Rate out = rate in
Alas, I think it is far more likely that there is no secret to any of this. Different strokes will arrive at the same place for a lot of people. All the more true for things that are even remotely creative in nature.
Not to say that routine and form can't get results. It is hilarious how much of the current fascination with LLM writing can be summarized by "actually filling out a routine template will satisfy a ton of requirements." People that are surprised with how well some output works, but would have scoffed at filling out a lot of boilerplate in previous technologies.
So, yes, try it. But do not become attached to it. If it works, rejoice in that. But do not count on it always working. If it stops for a time, feel free to leave it for a time.
Every day I work on my main project, I clear my desk completely, take out a small notepad and write my overriding goal, and my next step goal, and then make a list of tasks for that next step goal.
Then I work.
Writing the major goal every day is important to not let sub-goals overshadow it. Writing the immediate goal every day is important because together the two goals create a very clear direction of action with a clear next step.
I have my screen mounted on the wall, and have side end-tables for pens, papers and notes I need, etc. so my desk is absolutely clear.
My desk is a half circle, but not that deep, because that optimizes the usefulness of the surface for work (not storage).
I believe we clutter our workspaces because we suck at keeping iterations short. We always want to add one more feature, tweak one more thing, etc.
Eventually, some external pressure (boss, client, IM, whatever) causes us to open a second context simultaneously. Then it happens with a third, a fourth, etc
This is happening because the world is expecting shorter and shorter time to results due to better tooling in the last 10 years, but most have not figured out that all the LLMs and agents in the world won’t shorten the loop, only the person using them can do that.
I find that for any given problem, if I don’t see results in 30 minutes, it’s time to stop that problem and likely reshape it. If I don’t actually get the result in 90-120 minutes, I’m doing something wrong.
Every morning I close all work browser tabs from prior day. 99% of them I don't need again/can just reopen if I need. The 1% I'll note on a todo list or keep open somewhere.
Meh. This just sounds like all the interface theory stuff we users have to deal with, where useful things are removed in favor of a 'clean' and empty interface that makes you work harder to get your actual work done.
Then there's Adobe who remove features to add feature and justify it's next version; or clones it into a separate product so they can justify it's next subscription rise; or moves it into a different product so they can justify it's subscription expansion.
The end of every night should start with an empty page
and then start every morning with an empty page
It's pretty simple.
As a data hoarder something like onetab is amazing, there is still a lot of room for improvement though in browser ergonomics, session resets that force you to log back in and refind your place, it's nice to see some tools like data bricks that will at least let you reauth in a new tab.
So I'm one of the people who shuts down their computer at the end of the day and starts fresh. I don't use any type of session resume at the OS level or in the browser, I don't like it.
But! I've learned to harness the power of ending the day without complete closure. I stop work when I know the next step I'm about to do fully. Then the next day it's completely obvious what to start with, and I'm back in the flow without procrastinating as much.
It took some attempts to get comfortable with this; NOT finishing can be kind of excruciating if you're not build for work/life separation. But once I learned to delay my gratification in this regard, I found it set me up for many other things that require daily habits. I also balk much less at "this will take daaaaays" scenarios in general. I'm more comfortable now with things that stretch over longer periods.
What am I working on, what's in progress? The work space is the map. The terrain is changing as the task progresses, and so must the map, but the map is useful, even if it takes a bit of redrawing here and there.
The desktops (multiple, 3-7) are the map of the work. Part of the work is keeping the map accurate, not wadding it up and throwing it in the trash.
I suppose different things work for different people, but I started with the suggestion here and came around to skillful use of space as the work map itself.
Cleaning and updating are continuous, not a 'big bang' clear-the-desks event, mostly. But if it's not continuous, the big bang is probably better.
Some spots are problem spots, like digital notebooks, desktop icons. When I notice a problem spot, I create a recurring task to remove one X per week, or in some of the worst cases, one X per day. I have a rule of clearing out the oldest two days of email each day. I miss some days if I'm busy, but on average rate out = rate in, because I will always catch up within a day or two applying the rule that the oldest two days of email need eviction (make a task out of it, archive it, whatever) every day. Rate out = rate in
Not to say that routine and form can't get results. It is hilarious how much of the current fascination with LLM writing can be summarized by "actually filling out a routine template will satisfy a ton of requirements." People that are surprised with how well some output works, but would have scoffed at filling out a lot of boilerplate in previous technologies.
So, yes, try it. But do not become attached to it. If it works, rejoice in that. But do not count on it always working. If it stops for a time, feel free to leave it for a time.
Then I work.
Writing the major goal every day is important to not let sub-goals overshadow it. Writing the immediate goal every day is important because together the two goals create a very clear direction of action with a clear next step.
I have my screen mounted on the wall, and have side end-tables for pens, papers and notes I need, etc. so my desk is absolutely clear.
My desk is a half circle, but not that deep, because that optimizes the usefulness of the surface for work (not storage).
Eventually, some external pressure (boss, client, IM, whatever) causes us to open a second context simultaneously. Then it happens with a third, a fourth, etc
This is happening because the world is expecting shorter and shorter time to results due to better tooling in the last 10 years, but most have not figured out that all the LLMs and agents in the world won’t shorten the loop, only the person using them can do that.
I find that for any given problem, if I don’t see results in 30 minutes, it’s time to stop that problem and likely reshape it. If I don’t actually get the result in 90-120 minutes, I’m doing something wrong.
and then start every morning with an empty page
It's pretty simple.
As a data hoarder something like onetab is amazing, there is still a lot of room for improvement though in browser ergonomics, session resets that force you to log back in and refind your place, it's nice to see some tools like data bricks that will at least let you reauth in a new tab.
But! I've learned to harness the power of ending the day without complete closure. I stop work when I know the next step I'm about to do fully. Then the next day it's completely obvious what to start with, and I'm back in the flow without procrastinating as much.
It took some attempts to get comfortable with this; NOT finishing can be kind of excruciating if you're not build for work/life separation. But once I learned to delay my gratification in this regard, I found it set me up for many other things that require daily habits. I also balk much less at "this will take daaaaays" scenarios in general. I'm more comfortable now with things that stretch over longer periods.