[Journey]

Note: when I say "evolves" in the prior post it's meant to be said with a smile on my face. I forgot to include the smiley however so here it is :-)

[joelande]

Originally Posted by Journey

One thing I like about the GTD methodology is the concept of an in basket, in which tasks are unloaded from the mind and dumped into it. If I can have only one context / activity, I end up spending a lot of time trying to think of what my context hierarchy could look like, and when I am viewing things in Context mode I would end up having to scan all the items because an item that should be in multiple contexts won't show up under every context where it should be.

Multiple contexts allows me the simplicity of "offloading" items from my mind.

Nice catch -

Now there is a a good GTD-point for all of those By-the-GTD-bible doctriners out there.

I too think allowing tags or multiple contexts would help clear the mind, which is a very important GTD concept -arguably the most important (besides actually getting things done!)

[ptone]

Some good discussion here.

I'd like to make a couple points - the UI for adding additional contexts could be as simple as a comma.

There is a big difference between hierarchy (very powerful and useful) and relational (Many to one, one to Many).

A single action should exist in only one place in a nice normalized OF DB - duplicating it in multiple locations in the hierarchy is no good.

For those that argue that there is a logical single context approach to your data, I don't argue that. But the fact is that the human brain is not always a rigorous logical implement.

To continue with our somewhat simplistic example. At the time I create an action, the most logical context for it might be phone. But were I to create the same action at a different time I might think to put it under Sheryl. Now when I want to review my actions by context, I have to do a mental pause and consider which context I put it in.

An imperfect analogy. To perform an action in many applications, I have a choice, I can:

Select and item from a menu
Perform a keyboard shortcut
Use a contextual menu
Use a toolbar

Now, any ONE of those may be logical, but does that mean that the UI should be constricted to only use one of them? The great thing about this approach is that it casts a wider net that can catch peoples intentions through multiple parallel approaches, rather than require that everyone remembers the one way to do something.

-P

[TheSteveW]

Great thread...

My comments about tagging had to do with the way my (ADD encumbered) mind works. I think (few) contexts are great for a rigid approach to GTD and if that works for you that's great.

Personally though I'd like the flexibility to quickly ask or add a quick note about, for example "OS X dev problems" without having to do a search. Since it's more of a hobby than a core element of my day to day work it doesn't make sense to create a context for it. If I did that I'd have hundreds of contexts - which then of course defeats the purpose of GTD as would the need to do a search. If I put it under a "hobby" context - again - I've got lots of threads in my brain. I want a tool to help me focus quickly.

Not every moment of every day is or should be (for me) driven by what I'm supposed to be doing. Sometimes I just want to "wander". Multiple contexts and/or tagging could be easily implemented as an option. If you don't want it simply turn it off in prefs and you have rigid GTD.

Of course this gets to the issue of diluting the basic functionality but how hard/disruptive would it really be to add a tag or multiple context filter to a context view? Would it fundamentally alter the tool or would it on the other hand possibly open it up to a broader audience?

Great thread and am enjoying reading and learning from all the different perspectives...

[al_f]

Originally Posted by joelande

I too think allowing tags or multiple contexts would help clear the mind, which is a very important GTD concept -arguably the most important (besides actually getting things done!)

Ah, but I disagree: to me if you put things in more than 1 context you probably haven't thought it through fully, as you haven't defined what is essential to doing the task. I prefer to do that thinking at processing time, so when I'm working I know for sure that I can do everything that's listed under the context that I'm in.

An example of how multiple context might not work is this: say you work with Sheryl and need to talk to her about a project. If you're taking a multiple-context approach you could put that in the 3 contexts office, sheryl and calls (assuming you don't need think you need to talk to her face-to-face). You bump into her in the cafeteria at lunchtime and check your Sheryl list: however, you then realise that actually you need to refer to some project materials in your office files during the discussion. So, the office context is actually the only one that applies in this case as you physically need to be there to have the discussion. OK, that's maybe a bit contrived, but for me GTD is all about making those kind of decisions when you're planning, not when you're working.

Like it or not, despite the people saying it isn't, at heart OF clearly is a GTD app: the project/context model is from GTD and OF is based on Kinkless GTD, which was explicitly a GTD app. While you might be able to use it for a non-GTD system, I think GTDers are the app's main target audience and that's where I think the feature set should be focussed.

Originally Posted by TheSteveW

Personally though I'd like the flexibility to quickly ask or add a quick note about, for example "OS X dev problems" without having to do a search. Since it's more of a hobby than a core element of my day to day work it doesn't make sense to create a context for it. If I did that I'd have hundreds of contexts - which then of course defeats the purpose of GTD as would the need to do a search. If I put it under a "hobby" context - again - I've got lots of threads in my brain. I want a tool to help me focus quickly.

I suppose again it depends on how you define a context, but for me "OS X dev problems" would be a folder in planning view (or a parallel project) and hobby would definitely be a folder in planning view. The actions in the projects in those folders would then be placed in contexts depending on the resources you need to get them done (computer, garage, hardware store/errands etc.). Would that work better for you?

[Volker]

Originally Posted by curt.clifton

For example, I could buy a new shirt on-line or in a store. But I'm never simultaneously in my Web and Errands contexts, so I have to choose one.)

I think, this is a good example to express my idea, that not the tool is insufficient, but the definition of contexts may be incomplete.

Lets take your example where the shirt-task is assigned to Web and Errands. If I'm in a certain context, e.g. I'm standing on a street full of shops, I want to see all possible tasks for this context, say Errands. In your case of double assigning, I must take into account, that not all of Errand-activities are allowed in my current context. I have to look if the activity is assigned to further contexts, which forbid to do that task right now. This is in contrast to the GTD philosophy, which wants to keep your system simple: just look and decide - not to discuss, what is correct.

Therefore Errands seems to be the wrong context for the shirt-task. Imho it should be Web. Because if I'm sitting before a computer with internet connection, I can browse all my open Web activities and voila, there is my shirt-task.

If the Web context is too widly defined for your needs, e.g. you only want to shop online, if you're sure, your account has enough money, than simply define a context Web-Errands besides Web and Errands. The definition for the contexts would be:

- Errands: activities for physical shopping or similar activites
- Web: web browsing, e-mail, information collecton, reading news etc. but without shopping
- Web-Errands: Webshopping

I use two contexts for Errands. As I'm traveling a lot for my job, I'm using Errands_anywhere and Errands_homecity. The first context is defined for things, which are to big to transport with me in train or plain. So if I'm standing in a shop of a foreign city, my PDA presents me the correct activities only and doesn't bother me with unnecessary ones.

Conclusion:
The correct and complete definiton of a context structure helps to keep things - the decision process in certain contexts - simple.

[mcoad]

This has become a really good thread, with some of the best and most instructive discussion on the forum. But even as the examples and usage scenarios multiply, the conclusion seems to get more obvious. Some users prefer the rigid single-context method, and find this better implementation of GTD. Others prefer multiple contexts, and find this implements GTD better (this, incidentally, is also the conclusion drawn by other GTD app developers: Things with its tags, Life Balance with its ability to place any number of contexts within any other context). So, obvious conclusion: allow multiple contexts, and those who find life better without them simply won’t use them, without the app being affected in any other way, while for the rest of us things will ease up greatly.

If only other feature issues were so simple.

[ptone]

Originally Posted by al_f

Ah, but I disagree: to me if you put things in more than 1 context you probably haven't thought it through fully, as you haven't defined what is essential to doing the task.

Two responses to this: In a busy day there may not always be time to so fully think out an action in the planning process. That may be the ideal, the goal to strive for, but not always the reality. If I'm trying to get through the inbox, process a ton of items, and I have to spend twice as long on each one to decide what is the perfect and true context, then I'm getting less done.

Also, you assume that every action has a single context that is necessary, while I would argue that some actions have no single necessary resource, but several that are sufficient. For those actions that have a essential resource requirement, then use a single context, but for those where there is more than one EQUALLY SUFFICIENT situation - multiple contexts increase your efficiency because you don't have to do the extra work (at planning or reviewing) to figure out which of the equivalent contexts it went into. There are uses of multiple contexts that would be flawed and lazy - but we all really do sometimes have flawed and lazy brains, so having a tool that works with us is better. (this is not to encourage lazy or sloppiness, just to acknowledge that what seems the "right" context at one time may not be apparent another)

Originally Posted by al_f

...but for me GTD is all about making those kind of decisions when you're planning, not when you're working.

Many situations like the one you listed may require a single context. But one can be quite effective and more efficient if you didn't HAVE to always make that decision. Add an action, add 2 contexts. Search one of those contexts later (the one that occurs to you) see the action, do it. Done. No need to consider on both ends which is the truest context.

Originally Posted by al_f

While you might be able to use it for a non-GTD system, I think GTDers are the app's main target audience and that's where I think the feature set should be focussed.

Other than perhaps dev resource time and some UI issues - I have yet to see any of the pro-single context voices articulate why allowing multiple contexts would negatively affect their ability to choose to use single contexts. It sounds mostly like religious differences.

-P

[mcoad]

Originally Posted by ptone

Other than perhaps dev resource time and some UI issues - I have yet to see any of the pro-single context voices articulate why allowing multiple contexts would negatively affect their ability to choose to use single contexts. It sounds mostly like religious differences.

-P

Exactly. But it would be good to hear from Ken or someone else at Omni on this. You’ve been very silent on this for a while, guys. Here’s hoping this is because you’re still open on the matter, watching and waiting...

[eronel]

Originally Posted by ptone

Two responses to this: In a busy day there may not always be time to so fully think out an action in the planning process. That may be the ideal, the goal to strive for, but not always the reality. If I'm trying to get through the inbox, process a ton of items, and I have to spend twice as long on each one to decide what is the perfect and true context, then I'm getting less done.

Exactly. That is the appeal of the tagging/multiple context approach.

... some actions have no single necessary resource, but several that are sufficient. For those actions that have a essential resource requirement, then use a single context, but for those where there is more than one EQUALLY SUFFICIENT situation - multiple contexts increase your efficiency because you don't have to do the extra work (at planning or reviewing) to figure out which of the equivalent contexts it went into.

Well said. That is the drawback of the single context approach. I find it especially bothersome for tasks related to people and shopping. Most of them can be done online, via telephone, or in person, unless there's something special about the situation or thing. But often I personally find that when I am not clear about how I want to accomplish something, I'm not really clear about what I want to accomplish. So for now I am sticking with single contexts.

Other than perhaps dev resource time and some UI issues ... It sounds mostly like religious differences.

You're right. It's both. But while dev resources and UI issues are not unimportant, they are likely to be decided by the OmniFolks. Religious differences can be fought out for centuries :-) .