Ask HN: Simple tools/hack for small project management

Sure, here's an example from 2 days ago.

During the day, my wip.md file looks like this https://gist.githubusercontent.com/hboon/cd7ffa9ed94ce3a8bc8...

I abuse the Markdown format:

    * "projects" marked with headers, with a prefix :: so I can search for it

    * I mark "tasks" as done with an *

    * I mark uncompleted tasks with @@

    * When I start to work on a project, I add a line near the top with the start time and when I finish, I mark the end time too, using a vim macro.

I use MacVim all the time, so my wip.md is always open as the first tab. And I bind cmd+1 to it.

At the end of the day:

I run a vim macro which runs a script to generate the total time for each project ending up with: https://gist.githubusercontent.com/hboon/b8770dea077f67664d9...

I then go through it manually and delete the tasks that are still marked as @@ just keeping those with an asterisk. So this effectively becomes my journal for that day. I keep a short note under the Notes: line. For eg. I might say that I didn't work because I went out with my family, or if I'm sick. I delete those mark with an asterisk in wip.md, add more or tweak the remaining ones and I have the file ready for the next day. I do my daily review at this time, so the macro which I mentioned actually opens a few other files including one called on-deck which is basically where I can grab things over from to add to wip.md. Over the years, I sometimes end up with wip.md growing too much, making it quite inhibiting. I try to keep the wip.md to ~1 page so it's more manageable.

I'm ashamed to share the script which I wrote a few years ago :P But it's simple, it just goes through each line looking for a regex match for:

    11:10 PM - 11:21 PM Mimic

I have a config file that looks like this:

    [timelog]

    projects=Everyday,Marketing,Mimic,etc

    ignoreInTotal=Plan,Review,Photos,Expenses,HomeAdmin,Paperwork

    timelogDirectory=/Users/hboon/Dropbox/backup/timelog/

So it can figure out the project name from the regex and only include the relevant ones. Then compute a total for each project, for AM/PM, as well as a grand total for the day.

You can see the result in that second gist link.

I think I got the idea from John Carmack's .plan files https://github.com/ESWAT/john-carmack-plan-archive/blob/mast.... My original version back in 2009 only had this:

    11:57 AM - 12:17 PM	Code: Check conversation view problems

    12:19 PM - 12:23 PM	Code: Crash when timeline->avatar->followers->back (not home, probably), probably problem with ContactsPicker

It just evolved from there.

Replied above.

I keep all paid work in a work.org file and clock in/out of headings there. The basic structure is

* Projects * Customer X

** TODO issue 42 breakage SCHEDULED:

* Customer Y

** subproject foo

** TODO fix slow frobnication SCHEDULED:

* Ideas

* Misc

When I get an e-mail or something about a new task, I use org-capture to turn it into a TODO with a link back to the e-mail and a default SCHEDULED value, and I hit C-c C-w to refile it directly into customers/subprojects (otherwise they end up under Misc and I'll refile later). I always clock into the task I'm doing, and I use the clocktable (limited to a time range) to figure out how long I've worked on what for my invoices. Since I switch tasks a lot, and didn't like how the recent-tasks-lists was emptied on restarting Emacs, I ended up writing a little helper package at [link redacted] to let me quickly clock in to recent tasks (also lets you navigate to recent tasks if you use ivy).

I do something similar to user terminalcommand, where I schedule tasks in my agenda, so when I open my work-agenda, the stuff that needs doing is always there. I might leave "sometime-maybe" tasks/ideas unscheduled (like "try shiny new library"), but anything people are actually asking me fix will be scheduled. It's really easy to push things towards the future in the agenda, so it's not really overwhelming, even though my work.org file is now at 11677 lines (and I see my work.org_archive is at 6403 – maybe I should archive more things, but it's not like I notice the size).

For collaboration, I use whatever other people use (trello, wikis, meeting notes in etherpad checked into SVN).

Ditto with Trello, and not just for software. I love the 'email to board' feature where I can forward supplier quotes/docs straight to the board from my inbox, so I can find them later!

I hope Atlassian look after it

ugh, my line breaks got lost. Trying again since I can't edit:

    * Projects

    ** Customer X
    *** TODO issue 42 breakage
        SCHEDULED: 

    ** Customer Y
    *** Fooproject
    **** TODO fix slow frobnication
         SCHEDULED: 

    * Ideas

    * Misc