On Tue, Apr 5, 2016 at 8:12 AM, Dominik Schrempf <dominik.schre...@gmail.com> wrote: > Hi John, > > thanks for your long answer. I could take out a lot of it. I think I > did not describe my problem well although all of you seemed to guess it! > Just a short rephrase with vocabulary from your posts: > > I want a logbook/journal where I always append text about new stuff to > the end. Only, when I work on a task that has arisen before, I file > this text under this task. This text should belong to the task only. > Text after the task should belong to the top level (but this is not > possible as you pointed out). > > The thing is, I can't really do this with headings only, because I often > append text after a task in the same minute and it does not make sense > to create a new headline then. I now use drawers to achieve this. That > was not intuitive to me but it seems to work well! I can now fold the > task-specific text under the tasks. If I expand all text but the > drawers, I get the view that I wanted (at least within Emacs). > > The setup would be perfect if I could file tasks without headlines, kind > of (is this clear?). This may correspond to what you called "inline > tasks" (because I want my tasks to show up next to where I take my > notes). > > E.g.: > ------------------------------ > Top level entries. This is all very interesting (A). > > ** TODO A task bla bla. > :DESCRIPTION: > We have to do this, because... > And then, ten days later we finished it using... > :END: > > This is the top level again and text here may be written directly after > I wrote (A) from above and filed the task. > ------------------------------ > > Do I make myself clear now?
I think so, and thus I can really only think of a couple options if you need to have the structure you describe (aside from your leaning toward drawers): - inline tasks: they're just sort of "meta data" within whatever you're currently typing, so that makes sense. If I understand correctly, you're treating the task as just a "this is an action amidst these notes," but the notes might continue on after that task is documented. - regular headlines: I *still* think this could work and may be wrong, but think you mentally dislike the org hierarchy more than anything being "wrong," per se, about how it's organized. Said another way, you want the task to appear chronologically where it arose, even though it doesn't *need* to appear there... With a simply modification of making a headline per day, all would seem to be resolved, no? #+begin_example * 2016 April ** 2016 April 05 *** Thing 1 This is all very interesting. I only belongs to logs for 2016 April. *** TODO a task This is in the contents of a task that came up on April 5th. It only belongs to the day's log. *** Thing 2 This is also very interesting, and includes a task belonging to these notes. Here are more notes **** TODO a task Here is a task which belongs to Thing 2, April 5th, and April 2016. I inserted it at the very end of Thing 2 so that none of the non-todo based stuff about Thing 2 belongs to this task #+end_example Is it the levels of hierarchy that bother you about this approach? I may have misread your initial request, but I wrestled with this initially with org, finally accepting that it's all about headlines. They're far more compatible with agenda, more meta data (e.g. drawers, tags, etc.) can be attached, and re-arranging is easier. Moving to drawers is up to you. I personally dislike accessing/editing them. Search this manual page for "drawer" and then search it for "headline" and "subtree"... - http://orgmode.org/manual/Structure-editing.html I use C-c C-[n/p/U] a lot, which gets you to the next headline. It just seems like extra work to arrow down and TAB on the drawer when the same can be accomplished in a headline. Looking at your example, I've started to do something similar with some of my headlines when I get new information or revisit things. I work in R&D, and am often doing follow-up experiments or continuations of some previous work. I used to go through the effort of linking between related headlines, i.e. creating a main headline do document the work when I started, then a headline in perhaps another month's tree with the continuation. I'd =C-c l= on the old headline and then =C-c C-l= to insert a link like: #+begin_example * Some description [2016-04-04 Mon] Today I re-did [[id:2745364e-8692-43ff-9d3c-89bdbf265d5e][this experiment]] with such and such modifications. #+end_example Now for the sake of ease I often just refind the original and do: #+begin_example * Some description [2016-03-21 Mon] Did an experiment today about blah... [2016-04-04 Mon] This is the experiment I did today after the observations from above: tables, data, code, whatever... #+end_example The mental block, at least I've found, was in my own mind... Keep in mind the above are just examples; they'd be within some top level headline (which I've called "Journals") and a sub heading for the month (e.g. "2016 March"). Should the experiment(s) belong to the 2016 March headline or 2016 April? Are they really the same work done at different times or genuinely separate efforts? Before that I was pretty obsessed with data being associated with a specific project, and I maintained time-stamped journals within many separate top-level project headlines. This is probably made clear in the mailing list links I provided; I wanted the notes in Org to match my mental model, which was that the work was done for a *project* but also at a specific *time*. Finally I just gave up and adopted the time-based model. For me, it came down to asking myself how I would actually *use* Org mode. Maybe 10-25% of the times I pop open Emacs I'm adding content. The rest is almost entirely to access data stored in Org through C-s or C-a s followed by visiting the headline. Since I'm often after several bits of data about the same general thing, I'm find with leaving the initial work and all follow-ups under the same date tree (generally when it was initiated). This way, I can just scroll through the paragraphs and "re-live" the order of events. In large part, this was when the US moved away from a first to invent intellectual property system. I used to have to document my daily work, so chronological order was really important and thus printing/witnessing efforts made more sense chronologically as to not forget new work hiding under a past headline that I'd already printed. Not the US is first to file so I no longer have to maintain that sort of documentation. Alright, going to wrap up that novel! Good luck in finding a solution that works for you. This might seem like a nightmare to you, but you could always try a few different ways, each for a week or month and then compare how they work in terms of content generation as well as how you end up accessing/using said content? Just an idea! Best regards, John > Thanks, > Dominik > > On Thu, Mar 31 2016, John Hendy wrote: >> On Thu, Mar 31, 2016 at 3:12 AM, Dominik Schrempf >> <dominik.schre...@gmail.com> wrote: >>> Hello, >>> >> >> [snip] >> >>> A possible example: >>> >>> * February 2016 >>> February 10th: Some text and stuff in February 2016. >>> >>> ** TODO A task to be done. Filed on February 10th. >>> E.g., February 24th: Some text that should belong to the task only. >>> I could not work on this task before February 18th. >>> >>> February 18th: Some more text belonging to February 2016 and not to the >>> task. >>> >>> * March 2016 >> >> Like Eric, I'm a little confused of what you would want instead. The >> above is great for what currently happens, but could you do a similar >> example of what you want? You ask if this "feature" exists, but I'm >> not sure what it would be... all I can envision as a modification to >> above is: >> >> * Feb 2016 >> >> Feb 10th: blah blah blah notes >> >> Feb 18th: blah blah notes >> >> ** TODO filed feb 10th, but *done* on 2/24 >> >>> And so on. Maybe this feature does already exist, but I am not aware of >>> it. I know that especially upon export, this is hard to realize, >>> because all text always belongs to the previous headline. But maybe it >>> is worth thinking about it because at least to me this would be highly >>> useful (e.g., having different styles in HTML export for the text under >>> the task and the text of the top level, the time line). >> >> I've wrestled with this a lot myself, at least if I put this in the >> bucket of "what's the *best* way to organize an org file." To expand >> on Nick's comments, something can only be in one hierarchy at a time, >> and everything afterward will live in that parent/child, unless you >> start a new sibling. The downside is you can't "escape" a current >> sub-heading to return to "just the parent heading" again. I've not >> quite wrestled with that, but moreso the desire to have one thing live >> in several places at the same time. I posted some structure questions >> when I migrated from TiddlyWiki in 2010; you could take a look at >> these threads if you're interested: >> - https://lists.gnu.org/archive/html/emacs-orgmode/2010-03/msg00390.html >> - http://lists.gnu.org/archive/html/emacs-orgmode/2011-07/msg01173.html >> >> If your example is accurate, why not make everything it's own >> headline? The notes from 2/18 wouldn't, then, "belong" to the todo >> filed on 2/10 and completed on 2/24. You'd just have: >> >> ** Feb 10 >> blah blah >> ** TODO Feb 10 something >> Notes about task >> ** Feb 18 >> blah blah >> >> It seems the core of your issue is not being certain on whether or not >> you want the TODO to be represented in the date tree according to >> creation or completion. That, or you don't like that you have to >> decouple the todo itself and your notes about it, which would lead to >> separate entries, one for the todo on 2/10 and one for the notes about >> what you did to complete it on 2/24. Are any of those accurate? I >> think clarification would be helpful if I've missed what you're >> wrestling with. >> >> I've taken to a pure datetree for notes, with inline todos for >> anything that comes up in the context of something else (and which I >> want to keep in that context). So: >> >> * Meeting about blah :tag: >> [2016-03-31 Thu] >> >> Notes here about thing >> >> ************* TODO some task >> >> Notes I did about this todo >> ************* END >> >> Otherwise, I have a separate tree just for tasks where I don't care if >> they're decoupled from their context. It's just a headline called >> "Tasks" which is my dumping ground for todos. My actual org file looks >> like this: >> >> * Tasks >> ** todo something 1 >> ** todo something 2 >> >> * Journals >> ** 2016 March >> *** Something >> [2016-03-31 Thu] >> >> Notes >> >> *** Something else >> [2016-03-30 Wed] >> >> Blah >> >> For the tasks, I often just delete them as I don't care, but for ones >> where I've noted progress about them, I use C-c C-x A to archive them >> in a subtree of "Tasks." Then I could search for the info in them down >> the road if I want. I've also started just adding time stamps and >> updates to other month's headlines if the activity is a continuation >> of when it started. So maybe: >> >> * Journals >> ** 2016 Feb >> *** Experiment for projA >> [2016-02-10 Wed] >> >> Set up this experiment today... >> >> [2016-03-02 Wed] >> Ran a modified version of this experiment today... (and so on) >> >> Hope that helps a little... I love thinking about org file strategies, >> so please keep the thoughts coming if you'd like to discuss more! >> >> John >> >>> >>> Thanks and best wishes, >>> Dominik >>>