Welcome, Guest. Please Login
Tinderbox
  News:
IMPORTANT MESSAGE! This forum has now been replaced by a new forum at http://forum.eastgate.com and no further posting or member registration is allowed. The forum is still accessible via read-only access for reference purposes. If you wish to discuss content here, please use the new forum. N.B. - posting in the new forum requires a fresh registration in the new forum (sorry - member data can't be ported).
  HomeHelpSearchLogin  
 
Pages: 1 2 3 
Send Topic Print
Setting up TB 6 for qualitative research project (Read 8268 times)
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #15 - Feb 19th, 2016, 9:24pm
 

Got it.  Here's a decent example of one such file: https://www.dropbox.com/s/ze5y0sregobkpu6/Tinderbox%20example.rtf?dl=0


It's a bit more prefabricated than other files (which are just disorganized text files).  This one actually has an order, naming players, events in chronological order, important quotes that related to those events, etc.  Anyway, it does provide a good framework for how I'd like to import those core components into Tinderbox, and use them accordingly.

Thanks in advance for your help with this!
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #16 - Feb 20th, 2016, 1:11pm
 
As v6 doesn't allow drag-drop form one window (or tab) to another, Tom webster's process can't be done as in previous versions. But many of the same types can be used.

From this I'd say you want a people prototype with name rank/title and unit/formation at key attributes; perhaps also the $StartDate and $EndDate for their arrival in a unit and leaving it. Another prototype could be entities such as location/organisation - or you could have separate prototypes for each. Thus people (notes for each of Rivers, Shorthill), locations (Slovenia, Army bases, ...), organisations (Ishkabar), etc.

Now your exploded notes can reference these in various ways. Your (exploded) notes might has key attributes with user-created Set-types $People, $Locations. So the first item would have a $People value of "Rivers;Shorthill". As a set it de-dupes so each person only exists once (or not at all) in the listing for each note. You can also use the built-in $Tags to store keywords like 'transaction', 'contract', 'relationship', etc.

With this set up, it is possible to link a person-prototyped note to all notes using that note's $Name as a value in $People. As documented elsewhere in the forum, it's also easy to take any attribute (single or multi value) and the use values() to make a list of the unique values in the $text of a note and explode that into per-value notes.  these can then be linked to occurrences of that value and for notes on that subject.

I hope that gives a few ideas.
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #17 - Feb 21st, 2016, 12:58am
 
Thank you so much for this!  It's very helpful...

So, it seems like the first steps you're suggesting involve creating prototypes with key attributes for names of person, names of unit -- both with with date parameters -- as well as for locations and organizations.  And then I would create values for key attributes (i.e., names that are associated with $People values), correct?

Then, as I understand it, after I explode my note, Tinderbox will recognize the key attributes from the exploded note (and the values created, therein), and then...what happens?

If I understand this properly, I'll ask some questions about how to complete these steps. I'm still trying to grasp TB's overall process as well as mechanics themselves.

Thanks so much again for your help...
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #18 - Feb 21st, 2016, 3:46am
 
Quote:
Then, as I understand it, after I explode my note, Tinderbox will recognize the key attributes from the exploded note (and the values created, therein), and then...what happens?


Not quite, but you may create prototype for the exploded source notes. The Explode process lets you set a prototype for the newly exploded notes. Exactly what attributes you want in the prototype's Key Attributes (KA) is up to you.

Notes don't auto-detect anything but agents and rules, etc., can. My rough idea was you:
  • Set up prototypes for things you know you'll need (like the main players in the story, etc.) and for the source notes.
  • Import/explode the source notes - or an example set of them.  Look at whether your prototypes and their KA configuration covers what you need. Add additional attributes, prototypes, etc. as required.
  • If the last step resulted in a bit of mess, consider throwing away the imported notes, strip the doc back the prototypes and start over. Sometimes this approach of an exploratory test at outset makes for cleaner work after, even if it seems like extra work at outset.
  • Start you analysis.
  • Once you've some extracted data - i.e. notes/attribute values on your source data, you can look at detecting things.  for instance any notes about individuals might want to link to notes where they are mentioned - or perhaps where they are a significant character (or both, via different link types).
Don't worry if there are some tasks later in the work you don't yet understand.  You'll find that easier to see when you're working on your own data - not least as you'll see any odd edges peculiar to your data than won't be in generic examples. If stuck, just ask here. Just as with your RTF example doc, examples really help others help you as they give a concrete reference example. Good luck!
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #19 - Feb 22nd, 2016, 11:45am
 

Thanks so much again for this...it's very helpful in many ways.

Forgive me if my questions seem a bit remedial.  I've read quite a bit about TB, and have tried my hand at it a few times, but I do find the first set up a bit challenging.  I'm trying to improve, and this process has been very helpful.  

The thing is, some of the instruction I've followed in terms of documents and screencasts seem dated and/or reference folder versions of TB --- and I'm using TB 6.  So, are there links to instructions for setting up Prototypes, and their key attributes, that one could suggest for TB 6?  Also, as I was considering the set up for my prototype's KA's, I was thinking about the example I posted and how certain units and people had multiple start/end dates - for example, when they served multiple tours.  How would one resolve this in a prototype?  Is it possible to set up KA's in such a way so that there are multiple start/end dates?  Or should one create create seperate note - prototypes per each start/end date range?

Finally, a quick clarifying question.  In your last reply you wrote:

"Start your analysis...
Once you've some extracted data...you can look at detecting things..."

Just so I'm clear on the exploded notes process, are you saying that after one sets up prototypes (and refines them after a few passes), I will, at first, be manually assigning them to prototypes?  Also, at one point can one transition to using agents to automate some of this?  Thanks again for your tremendous help.
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #20 - Feb 22nd, 2016, 12:26pm
 
I think you would help yourself by taking a step back and teach yourself these various tasks (prototypes, KAs, Explode) outside the main work flow. Otherwise you'll get loss as to the cause if something unexpected happens.

Prototypes. To make a note a prototype, simply tick the box on the Properties Inspector's Prototype sub-tab. That one task designates that a note is now a prototype.

Configuring the KAs for prototype is exactly the same as for any other note. Inheritance via prototypes is explained a bit in my tutorials here and here. Indeed, you might find this listing of my Tinderbox walkthroughs of use (N.B. - they cover a mix of versions, some are for v5).

As for Explode, I suggest you experiment in a new throwaway file until you're happy you understand the process.

Quote:
Also, at one point can one transition to using agents to automate some of this?

It depends what you mean by automate.  There is no magic feature that just finds stuff. You have to tell Tinderbox what to look for. The point of the user attributes, prototypes, KAs, etc. is to enable you to easily 'annotate' your imported notes, and notes you make about those notes, such that you can then more effectively use Tinderbox's tools. I think you'll understand the mechanics of the latter once you've made yourself familiar with the basics of attributes & prototypes.

Do keep asking questions  Smiley
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #21 - Feb 23rd, 2016, 12:03pm
 

Thanks very much for this. Those links are terrific...  (By the way, it seems you sent me the same links for your "tutorials here and here"...and I'd love to see the other link you meant to send, if it's not too troublesome.)

I hear you about taking a step back and teach myself these various tasks outside the main work flow.  That's exactly what I'm trying to do right now, so we're on the same page.  To that end...

The links you send that involve prototypes and KA's mostly involve set ups for the appearance.  That's fine, and I suppose that's helpful for creating a distinguishable form when roughing out the map view (correct?).  Right now, I'm trying to understand the nuts and bolts of setting up KA's that have to do with their relevant info as it pertains to my research - i.e., name, unit(s), duration of service, deployment location, etc.  Could you direct me to links that would provide instruction on how to set up that information within prototypes and KA's?  I find the concepts of strings and booleans a bit confusing, so it would be helpful to learn what they mean and therein which options I ought to select.  

Also, as part of that, there are some cases in which a person has served on several deployments, and so I'd need to fill out several durations of service, deployment locations, etc.  How would I do that w/ KA's?

Thanks very much once again...
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #22 - Feb 23rd, 2016, 1:28pm
 
Ah, sorry about the paste error. Try this on Inheritance and on Prototype-based attribute inheritance. They're essentially different cross-cuts on the same thing. As long as one makes sense and you 'get' the basic concept of Tinderbox inheritance you'll be forging ahead. It seems complex at first sight but is powerful and elegant and very supporting if experimentation (mistakes!).

KAs just show - or allow input of - data you choose, so they don't 'do' anything per se. Also they aren't set in stone.  Whilst doing initial reading of source texts you might want one KA arrangement, then you might reconfigure them for a review phase and again for writing export. If you're comfy with prototypes, that sort of thing is a snap as you'll change the KAs in the prototype rather than in lots of your actual notes. Plus you can toggle a notes KAs as hidden/shown either to avoid distraction or just to see more text.

Under the hood, Tinderbox is all notes and attributes. Many are built-in (but few people need all those for any one doc) and user-attributes can be added as needed. Like form boxes print on an old file card, all cards have all boxes whether you use them or not. In other words, once you've added an attribute any note whether new or existing now 'has' that attribute. KAs are about those things you want to see up front.  The others can be interacted with via the Get Info (Cmd+Opt+I) pop-over or the Inspector.

Quote:
Also, as part of that, there are some cases in which a person has served on several deployments, and so I'd need to fill out several durations of service, deployment locations, etc.  How would I do that w/ KA's?


Hopefully from above, you'll see it's that attributes that will capture this. The KAs are just those attributes you've chosen to foreground for a particular note or notes.

So, people serve on 'deployments'. assuming you only ever serve once on any given deployment, a Set-type attribute might be best.  this will store one or more text ('string') values per note. If for some reason the note needs to sort multiple values that duplicate in the same list, then use a List.

To add your data capture, if you know all (or even just some) values in advance for your user attributes (Ranks, posting dates, formations) you can seed those values in advance
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #23 - Feb 24th, 2016, 11:16am
 

Ok, I think I'm starting to understand how things work a bit more...sort of.  

First, to back up a minute... When I first started this query, I was asking about ways in which one could use the process outlined in "Processing Qualitative Data with Tinderbox" video.   Further down in the thread your wrote, "If your data essentially has embedded  attribute-type data and you don't control the source, consider splitting the input via explode and then using agent actions, or rules, to mine this effectively tabular data to populate user attributes."  (BTW, I wasn't clear about what you mean about "and you don't control the source" Could you quickly explain that? - thanks!)  

Anyway, my understanding is that one started with creating prototypes and KA's, and then creating agents that would recognized and import/connect to the exploded notes, accordingly.  Is that correct?  I just want to be totally clear about that.  

You also suggested that I first "make and apply prototypes," and then "explode imported text and apply a prototype to the notes created."  So, I have a general understanding of how to set up a prototype and KA's, and I've tried setting some of them up (see below), but I'm not entirely clear on how one would take an exploded note, and then graph / absorb sections of the exploded contents with prototypes and their respective KA's.  I've been trying to imagine how that might work as I've been setting up prototypes and KA's, and considering how I could align the values in them with the example file I uploaded.  I feel like some essential step in this process, that involves selecting a note, and then applying a prototype to it (that contains / reflects the notes values, expressed in the prototype's KA's). If I've properly understood that process, can you clarify how I can apply it?  I've had a hard time locating & selecting prototypes that aren't "built-in," and that seems to be an essential part of the process you've outlined.  Thank you!

Ok, a few mechanical questions...

-First, I've been trying to set up a prototype for people - or "sources" - based on some of the values I mentioned before: names of people, names of unit, locations, dates of deployment, etc.  At first, I thought: There's already a built-in Prototype for "Person" -- maybe I just use it, and amend it to reflect my KA's...but I'm guessing that's not advisable, correct?

-I've tried setting up a prototype through the various approaches in the links you sent: both through the inspector panel and the Text panel that appears to the right of the outline view (that allow for the selection of KA's, and then the arrangement of those values in the "list" view).  I'm a bit confused by some of the settings.  For instance, when I'm in the Properties Inspector pane, the prototype check box is checked, but the drop down bar reads "none" even though I'm working on a particular prototype.  Is that right?  I ask because I feel that when I make changes to a prototype, they're not always saved.  How can I ensure that changes to prototypes are properly saved?  

-Also, the default setting under Quickstamp (a pane whose function I don't yet understand) reads "General."  Because I'm creating a prototype for characters, should I select "Person" instead?  I'm not sure if / how that will affect the settings and KA's I desire...  Furthermore, it seems to me that I need to got to the Appearance Inspector->Badge->and select "Person" to get the "Person" badge (instead of selecting Person from Quickstamp section), correct?

-Right now, the KA's I've set up for my first (attempted) prototype are: Name; FullName; Organization (which seems where "unit" would go); Badge = person; StartDate & EndDate (for dates of deployment, ostensibly). I also included: TimelineBand (though I don't know what it does yet); Searchable; Created & Modified.  So, does this seem to capture some of the KA's I need for the values I'm trying to create (again, names of people, names of unit, locations, dates of deployment, etc. )?  Or should I use different KA's?  Again (forgive me for saying this once more) there are more than one deployment per character, so I expect that might involve some kind of string KA set up, correct?  If so, can you please advise me how I'd do that?

I've got more questions, but I think that's a good place to start!  Thank you again for all of your tremendous help.  I feel like I'm getting there, in fits and starts...
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #24 - Feb 24th, 2016, 12:13pm
 
OK, there a lot of questions that I'm going to split out into separate threads otherwise this is going to get unreadable quite fast. That's not a rebuke! I think it's worth continuing the overall research theme here and doing the individual "how do I...?" question in separate threads. I'll start one that but there's a lot of writing and I'll need to shift location here before I'm done (go home!).
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #25 - Feb 24th, 2016, 12:36pm
 
Quote:
"and you don't control the source"

My apologies, I'm pointing out that if you write the source notes or make the database export, etc., that is your starting data you can control it's structure. Thus you might create output that makes it easy to find and extract data you want to put in attributes etc. If someone else is creating the you'll have to take what you're given.

Tinderbox isn't a magic box finding as yet undefined terms. You need tell the app to look for things and what they should be. As the hardest/slowest way of searching is on the text of notes using regular expressions (or just text terms) it makes sense to abstract some data to attributes. For instance, you might create a user Set-type called 'People'. As you review your notes, you add the names of any characters in your story to that note's $People attribute (here I'm using the '$' prefix as shorthand to indicate I'm talking about an attribute - which may or may not yet exist). Now to find a character called Smith, instead of doing $Text.contains("Smith") we can query $People=="Smith". Don't worry about the meaning of that code here. The point is that there is more than one way to do things, no single way is the right way, different ways suit different personal work styles, some ways are more processor intensive than others (the last begins to matter if your project becomes big - 000s of notes as opposed to a few 10s of them).

Quote:
then creating agents that would recognized and import/connect to the exploded notes, accordingly


Possibly, depending on what you set up (given my comments above). At the early stage you'll be doing some experiments to refine how you categories/label things (attributes, prototypes, etc.). Some folk are happy with little formalisation and don't move much beyond what you might do on paper. Others structure their data to allow quite complex querying of relations ships. The latter makes it easier to ask, "Did X and Y ever serve together in location Z?" without simply reading lots of text notes.
Back to top
 
« Last Edit: Feb 26th, 2016, 3:48am by Mark Anderson »  

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #26 - Feb 24th, 2016, 1:04pm
 
Quote:
You also suggested that I first "make and apply prototypes," and then "explode imported text and apply a prototype to the notes created."  So, I have a general understanding of how to set up a prototype and KA's, and I've tried setting some of them up (see below), but I'm not entirely clear on how one would take an exploded note, and then graph / absorb sections of the exploded contents with prototypes and their respective KA's.  I've been trying to imagine how that might work as I've been setting up prototypes and KA's, and considering how I could align the values in them with the example file I uploaded.  I feel like some essential step in this process, that involves selecting a note, and then applying a prototype to it (that contains / reflects the notes values, expressed in the prototype's KA's). If I've properly understood that process, can you clarify how I can apply it?  I've had a hard time locating & selecting prototypes that aren't "built-in," and that seems to be an essential part of the process you've outlined.  Thank you!

Tinderbox's strength is you don't have to get everything right, so don't worry if you've spotted every possible attribute, prototype you might need. Indeed, unless it's obviously needed, I'd leave it until later.  Unlike a database, Tinderbox makes it easy to add structure like a new attribute mid-process. At worst you may need to run a query to find existing notes that might need to have that attribute field out but it's not hard.
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Setting up TB 6 for qualitative research project
Reply #27 - Feb 24th, 2016, 3:14pm
 
Quote:
-Right now, the KA's I've set up for my first (attempted) prototype are:

It might be useful to say what it's called. Note titles ($name) are case sensitive.  One naming convention that can be helpful is to put some other character at the start of the name so it's clear as to purpose.  I tend to use a 'p' prefix, e.g. "pPerson". Others use an underscore, asterisk, etc.

Quote:
Name; FullName; Organization (which seems where "unit" would go); Badge = person; StartDate & EndDate (for dates of deployment, ostensibly).

Unless you're going to change the badge for that prototype regularly, there's no point in adding $Badge to the KAs. Having set a badge, it shows in maps, outline etc, so you don't need it adding an extra row to your KA. Adding/removing an attribute from KA has no effect on that attribute. as the term 'key attribute' implies, those listing in the text window are those you've chosen as being 'key' for that note (or group of notes via a prototype).

Quote:
I also included: TimelineBand (though I don't know what it does yet);

Given you can add it back at any time, why put in things you've no clear idea of a need for? That's not meant rudely. Consider: there ought to be a rationale for adding KAs otherwise they are visual clutter. Plus you can always, easily, change it later.

[url]Searchable; Created & Modified.[/url]
I'd pose the same question as above re $Searchable. Do you envisage constantly wanting to swap all notes (using the prototype) out of searches? I suspect not. $Created never changes (it's when the note [sic] was first created. $Modified is read-only and calculated on the fly and shows when the note's $Text or user attributes or some system attributes were last modified. I can't see you needing that information in a KA. I'd suggest a test you apply; if you can't explain to someone why you absolutely need a particular attribute as a KA, you probably don't need it.

Quote:
So, does this seem to capture some of the KA's I need for the values I'm trying to create (again, names of people, names of unit, locations, dates of deployment, etc. )? Or should I use different KA's?
 
At present I think you're still trying to guess the 'right' see of attributes. If you find you keep wanting to look at a feature of a note and it's not a KA then you should add it (if necessary making a new attribute).

Quote:
Again (forgive me for saying this once more) there are more than one deployment per character, so I expect that might involve some kind of string KA set up, correct?  If so, can you please advise me how I'd do that?

A String data type attribute is not a good choice for a list. Strings are single-value. Editing or replacing the current value, results in just the new values. Multi-value attributes - lists of values - in Tinderbox are Lists or Sets. Sets have the useful feature of not accepting duplicate values but otherwise broadly behave like lists. For a $Deployment attribute, I'd go with a set as the same person can't be (I'm assuming!) on same deployment more that once. By that I mean they can't go to Croatia for the whole of 1986 more that once (not without time travel). They might however go the the same place at a different time (e.g. return to Croatia in 1988), though I'd assume that would be considered a different deployment.

I don't want to overload this answer, but if you need to change a String to List or Set - or vice versa, it is possible but the detail of that will just confuse this juncture.
Back to top
 
 

--
Mark Anderson
TB user and Wiki Gardener
aTbRef v6
(TB consulting - email me)
WWW shoantel   IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #28 - Feb 25th, 2016, 10:26pm
 
Quote:
It might be useful to say what it's called. Note titles ($name) are case sensitive.  One naming convention that can be helpful is to put some other character at the start of the name so it's clear as to purpose.  I tend to use a 'p' prefix, e.g. "pPerson". Others use an underscore, asterisk, etc.


Thanks for this! But I'm afraid I'm not grasping where these settings for KA's are located.  I've hit the "+" sign to Add Key Attributes in the right pane, beside the outline view, and have gone through all of the categories, but haven't seen Note titles ($name).  Is that (along with "pPerson") a KA that you've customized through another KA setup?

Quote:
Unless you're going to change the badge for that prototype regularly, there's no point in adding $Badge to the KAs. Having set a badge, it shows in maps, outline etc, so you don't need it adding an extra row to your KA. Adding/removing an attribute from KA has no effect on that attribute. as the term 'key attribute' implies, those listing in the text window are those you've chosen as being 'key' for that note (or group of notes via a prototype).


I see.  Thanks.  I thought it might be helpful to have an easily identifiable icon associated with a note in the outline view - esp. for people/characters. Or, if I'm understanding your properly, you're saying that it's possible to set a badge for such prototypes, but there are better, more advisable ways to go about doing so. Is that correct?


Quote:
Quote:
I also included: TimelineBand (though I don't know what it does yet);


Quote:
Given you can add it back at any time, why put in things you've no clear idea of a need for? That's not meant rudely. Consider: there ought to be a rationale for adding KAs otherwise they are visual clutter. Plus you can always, easily, change it later.


Totally fair point.  I suppose I was just experimenting to see how I could best achieve include a time element for a person/character note. Trying to figure it out w/ trial and error -- but I really do appreciate your point!

Quote:
[url]Searchable; Created & Modified.[/url]
I'd pose the same question as above re $Searchable. Do you envisage constantly wanting to swap all notes (using the prototype) out of searches? I suspect not. $Created never changes (it's when the note [sic] was first created. $Modified is read-only and calculated on the fly and shows when the note's $Text or user attributes or some system attributes were last modified. I can't see you needing that information in a KA. I'd suggest a test you apply; if you can't explain to someone why you absolutely need a particular attribute as a KA, you probably don't need it.


All totally great points.  Thanks for making them! I don't know about modified, though. It can be useful to have that info to determine the last time I updated info.


Quote:
Quote:
So, does this seem to capture some of the KA's I need for the values I'm trying to create (again, names of people, names of unit, locations, dates of deployment, etc. )? Or should I use different KA's?


Quote:
At present I think you're still trying to guess the 'right' see of attributes. If you find you keep wanting to look at a feature of a note and it's not a KA then you should add it (if necessary making a new attribute).


I assume you mean I should not add it!  I hear you.  And yes, you're absolutely right: I am guessing my way through the KA's, trying to figure out which would be the right ones to suit my purposes.  I feel like I have a conceptual understanding of how TB ideally works (and would serve my purposes), but man...it isn't the most accessible, straightforward interface!  Hence the guessing! And why I'm so deeply grateful for your help and guidance...


Quote:
Quote:
Again (forgive me for saying this once more) there are more than one deployment per character, so I expect that might involve some kind of string KA set up, correct?  If so, can you please advise me how I'd do that?


Quote:
A String data type attribute is not a good choice for a list. Strings are single-value. Editing or replacing the current value, results in just the new values. Multi-value attributes - lists of values - in Tinderbox are Lists or Sets. Sets have the useful feature of not accepting duplicate values but otherwise broadly behave like lists. For a $Deployment attribute, I'd go with a set as the same person can't be (I'm assuming!) on same deployment more that once. By that I mean they can't go to Croatia for the whole of 1986 more that once (not without time travel). They might however go the the same place at a different time (e.g. return to Croatia in 1988), though I'd assume that would be considered a different deployment.

I don't want to overload this answer, but if you need to change a String to List or Set - or vice versa, it is possible but the detail of that will just confuse this juncture.


Ok, I think I generally follow you. Could you walk me through how I'd set up something like a $Deployment attribute?  That might help me figure it out more clearly.  And could you give me an example of how Strings would be used (perhaps in the context of the document I uploaded)?  That might clarify what you mean a bit more...  Thank you!
Back to top
 
 
  IP Logged
NZT-48
Full Member
*
Offline



Posts: 24

Re: Setting up TB 6 for qualitative research project
Reply #29 - Feb 25th, 2016, 10:38pm
 
Quote:
Tinderbox isn't a magic box finding as yet undefined terms. You need tell the app to look for things and what they should be. As the hardest/slowest way of searching is on the text of notes using regular expressions (or just text terms) it makes sense to abstract some data to attributes. For instance, you might create a user Set-type called 'People'. As you review your notes, you add the names of any characters in your story to that note's $People attribute (here I'm using the '$' prefix as shorthand to indicate I'm talking about an attribute - which may or may not yet exist). Now to find a character called Smith, instead of doing $Text.contains("Smith") we can query $People=="Smith". Don't worry about the meaning of that code here. The point is that there is more than one way to do things, no single way is the right way, different ways suit different personal work styles, some ways are more processor intensive than others (the last begins to matter if your project becomes big - 000s of notes as opposed to a few 10s of them).

Quote:
then creating agents that would recognized and import/connect to the exploded notes, accordingly

Possibly, depending on what you set up (given my comments above). At the early stage you'll be doing some experiments to refine how you categories/label things (attributes, prototypes, etc.). Some folk are happy with little forma;libation and don't move much beyond what you might do on paper. Others structure their data to allow quite complex querying of relations ships. The latter makes it easier to ask, "Did X and Y ever serve together in location Z?" without simply reading lots of text notes.


I really appreciate everything that you've written here in trying to answer my questions.  Truly.  I feel that I'm getting a closer understanding of how TB might work in achieving the "Processing Qualitative Data with Tinderbox" video -- without the drag and drop feature (which frankly made it so much of this process seem so much easier!).  But honestly, it's hard to imagine the mechanics you've described.  I feel I need to somehow get to the point where I can set up a framework -- through prototypes, KAs, and agents (and whatever else you might suggest) -- and try making it work.  In that way, I might have a more concrete understanding of how things might work.  Make sense?  Or would you suggest another course?  Please be honest.  And thank you...again!
Back to top
 
 
  IP Logged
Pages: 1 2 3 
Send Topic Print