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 This forum has now been replaced by a new forum at http://forum.eastgate.com and no further posting or member registration is allowed.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 4 
Send Topic Print
Tinderbox for Textual Analysis (Read 111537 times)
Charles Turner
Full Member
*
Offline



Posts: 180
New York, USA
Re: Tinderbox for Textual Analysis
Reply #30 - Jul 03rd, 2009, 6:29pm
 
Not a criticism of Tinderbox, but it's fundamental, "atomic" if you will, element is the note. Textual analysis, OTOH, wants all of Tbox's functionality to work at the level of the word. (Or maybe phrase, but that could consist of one word, correct?)

So, making Tinderbox really effective at this finer granularity would be a major revision to the design or paradigm of the program.

You could break the text into phrase-sized notes. That would enable the application of Tbox's current functionality, but seriously challenge anyone's ability to mentally reintegrate the text into its whole.

Happy 4th! Charles
Back to top
 
 
WWW   IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Tinderbox for Textual Analysis
Reply #31 - Jul 4th, 2009, 6:32am
 
@Charles.  I'd agree with your analysis that TB's designed atomic level is the note rather than the word and that altering the level is likely a bigger (software) designed change.

@Jean.
Quote:
One of the reasons that this is hard-ish in Tinderbox is because the source text is extended:  it's important for the analyst to be able to look at it whole.  Otherwise the source could probably just be decomposed into chunks, and the chunks managed in various ways that Tinderbox already supports very well.

If you want to see a whole text - or passage of text, why not use Nakakoji view? With a little preplanning and appropriate templates, the Nakaoji view can be used as a text viewer. OK, there's no markup (bolding, etc.) but you can use it to view and read many different notes (and/or other information). I suspect if there was a demonstrated demand/need for a text (re)view that might be comparatively easy to add compared to some of the other stated needs.

If you need attributes at word and sentence/phrase level, you could always explode each phrase into words and make the latter children of the phrase note, thus giving you the ability to tag at word level.

Gettysburg Address? Yes, it's not unknown this side of the Pond!
Back to top
 
 

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



Posts: 97

Re: Tinderbox for Textual Analysis
Reply #32 - Jul 4th, 2009, 5:01pm
 
Hi Jean and Charles!

Jean, it's interesting that in your first post in this thread you express reservations about incorporating textual analysis functionality within Tinderbox. But in your second, you ably demonstrate (1) why this would be exceptionally useful within a tool like Tinderbox, and (2) why it is doable (barely) within Tinderbox right now, and would be much better with the sort of ideas I'm considering.

In the coming week or so on http://loryn.me I'll certainly re-examine Textual Analysis Example 1 in light of Mark Anderson's comments about the footnoting tool, and Jean's comments in light of her suggested workarounds within Tinderbox and the potential and limitations of using multiple specialised tools within a textual analysis process.

MarkA, please don't misunderstand: I do know how much effort would be involved in overlay visualisations. It is a substantial extension of Tinderbox's functionality. Whether or not Mark B decides to implement something like it within Tinderbox, or within a related product, will depend on his strategy with respect to managing the ongoing design and development of Tinderbox. OTOH, I suspect the concept of textspans (syntagm lenses) and concordance display agents would be relatively trivial given the programmatic framework within Tinderbox.
Back to top
 
 
  IP Logged
Loryn
Full Member
*
Offline



Posts: 97

Re: Tinderbox for Textual Analysis
Reply #33 - Jul 4th, 2009, 5:03pm
 
Charles

You've got it in one:

Quote:
Not a criticism of Tinderbox, but it's fundamental, "atomic" if you will, element is the note. Textual analysis, OTOH, wants all of Tbox's functionality to work at the level of the word. (Or maybe phrase, but that could consist of one word, correct?)


That's exactly what I'm shooting for within this piece of research.
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Tinderbox for Textual Analysis
Reply #34 - Jul 6th, 2009, 7:45am
 
There's nothing to stop you entering one word per note. Tinderbox makes no assumptions as to what you put in the note (or study about that note). It can as easily be a word as a whole essay. I make the point as some comments thus far appear to assume otherwise (presumably simply because of not having tried). At some point there's an issue of scale. The Gettysburg address would be fine but I'm guessing the whole Koran as note-per-word might push the limit of a single TBX.

I've investigated Exploding to word level, since I'd effectively suggested doing as much in earlier posts.  Seems TB can't do this. I tried a space character as a delimiter and also \s. The former gives now exploded result and the other varied but unexpected results. So an Explode split to word level is probably one more thing on the want list for the type of textual analysis being discussed.

For now, to split to word-per-note, I put the text into BBEdit (or TextWranger) and replaced all spaces with tabs and split on that.  Doing this using the Gettysbug address threw up some interesting typographic curve balls for processing. One was the use of ellipses - two at the beginning of paragraph 3; in fact they're faux-ellipses in my source, being just 3 periods in a row, which have no whitespace at the containing word boundaries. The other punctuation form were em-dashes, there being four of these in paragraph 3 of the address. So you'll need to find each of those punctuation form and add a tab after the punctuation mark before doing your Explode-on-tabs, which then works perfectly (I used v4.7.0).
Back to top
 
 

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



Posts: 180
New York, USA
Re: Tinderbox for Textual Analysis
Reply #35 - Jul 6th, 2009, 9:55am
 
Hi Mark A.-

Yes, it's technically possible for Tinderbox to support this use, but here's where design and (poor) human cognitive abilities get in the way. Once the text has been dis-integrated into notes, most humans would have difficulty re-integrating the original text.

We somehow desire a presentation that we can mentally dis- and re- integrate at will, through a shift of attention.

Tinderbox could get closer to this ideal by putting the "tagged" phrases into the note Name or Display Name, so that the text could be read as a whole in an Outline view.

Again, no criticism of Tbox (or even a request for features). I'm just playing out a gedanken experiment...

Best, Charles
Back to top
 
 
WWW   IP Logged
Jean Goodwin
Full Member
*
Offline



Posts: 136
North Carolina
Re: Tinderbox for Textual Analysis
Reply #36 - Jul 6th, 2009, 11:47am
 
Everything that Charles has said, plus:  The source text is often going to be "dis-integrated" in several competing ways (overlapping and/or embedded "tags"), with different lengths (from single word or even part of word to string of sentences).  That's going to be cumbersome (although not impossible) to manage just by "exploding" the source text and then adding attributes to the new child notes.

As Charles also said, this is just an interesting thought-experiment.  It's been helpful to remind me of the existence of the footnote tool (Charles--that tool will automatically put the "tagged" phrase into the footnote's Name).
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Tinderbox for Textual Analysis
Reply #37 - Jul 6th, 2009, 1:34pm
 
A problem with thought-experiments relating to actual tools is it allows us to keep shifting to goalposts to find new things we "can't do" which discourages anyone to investigate how much we actually can do!

Quote:
Once the text has been dis-integrated into notes, most humans would have difficulty re-integrating the original text.

But, there's nothing to stop one getting at the text!  As pointed out we can make the note title be the same as the $Text - either directly ($Name=$Text) or via use of $DisplayExpression.

Just because we've split the source text down to words and phrases doesn't mean we have to delete the source text - indeed Explode leaves the original in situ. Want only the text from para 3 or chapter 4? Then simply set up text template to allow you to show all the constuent per-word notes in sequence notes in a Nakakoji view. Before we go round the buoy again over 'syntagem lense' info I know that sort of info can't be shown - but that's not the point here - you can reassemble and view 'split' text (assuming you don't dump the punctuation (which Loryn's original example doesn't do.

Lest my comments be misunderstood, I'm not arguing for or against anything but simply trying to correct mis-presumptions about what TB "can't do". I'm not suggesting TB can carry out all of the things in Loryn's original text analysis needs but there are a number of constituent parts - like viewing snippets of text in aggregate - that are possible. The best way to see that the case is simply to try it out. If unsure let's explore that in a new thread.
Back to top
 
« Last Edit: Jul 6th, 2009, 1:35pm by Mark Anderson »  

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



Posts: 97

Re: Tinderbox for Textual Analysis
Reply #38 - Jul 6th, 2009, 4:06pm
 
Hey MarkA

I really appreciate your contributing to this discussion ideas as to how the current Tinderbox might possibly achieve the effect we're after. After all, my analytical example focuses on how I *have* used Tinderbox to achieve what I'm after. And with your input, I've learned some tricks I could've used to make that more effective.

WRT moving the goalposts. I think that's why one needs to use a series of worked, concrete examples (usage scenarios). That gives us a set of solid use cases to focus the discussion.

I'm planning on updating the description of Textual Example 1 to incorporate your suggestions on the footnoting tool in order to better represent the present capabilities of Tinderbox. And why didn't I use that in the first place? Simply because I knew *nothing* about it.
Back to top
 
 
  IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Tinderbox for Textual Analysis
Reply #39 - Jul 6th, 2009, 6:11pm
 
If I can help with working an example, let me know - happy to talk off-list if more appropriate (e.g. to work finer details).

Meanwhile I'm looking at some functions to help with data splitting and when that's done I'll look at the reverse. Following Jean 's mention of it, I'm using the Gettysburg Address (text as from Wikipedia) as my start text. Then we can look at the reverse - re-creating (sections) of source from the broken down text. What I hope to achieve is to have a TBX with some examples others can then use to apply to their own data. Nothing to show yet as I'm wrestling with command lines - getting Unix CLs to detect things like em-dashes. As ever the last 5% takes the most time. I'll share when I've something usable for others.


Back to top
 
 

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



Posts: 180
New York, USA
Re: Tinderbox for Textual Analysis
Reply #40 - Jul 7th, 2009, 7:42am
 
Quote:
A problem with thought-experiments relating to actual tools is it allows us to keep shifting to goalposts to find new things we "can't do" which discourages anyone to investigate how much we actually can do!

Hey Mark A.- I believe I'm the only person to demonstrate a working text tagging system for Tbox Notes. I've actually "done it" in Tinderbox. I don't think I'm "moving the goalposts."

Sure you're not "re-inventing the wheel"? :-)

Best, Charles
Back to top
 
 
WWW   IP Logged
Jean Goodwin
Full Member
*
Offline



Posts: 136
North Carolina
Re: Tinderbox for Textual Analysis
Reply #41 - Jul 7th, 2009, 8:09am
 
Hi, MarkA:  I'm very sorry if I'm appearing to be too hard on Tinderbox--that certainly isn't my intention.  I <3 Tinderbox.

Here's a reality check:  I'm on track this summer to write 4.5 articles for publication, and one major grant application. All of them will be drafted in Tinderbox:  from earliest random note collection through brainstorming to preliminary outlining, pre-writing and formal drafting.

Three of the articles involve what's being called here "textual analysis," so it's cool (to me) to imagine the many ways Tinderbox could help with that, too. This discussion has produced 4 approaches so far. Charles' is the most elaborate--as I understand it, it involves tagging-plus-Nakakoji view of the tags.  There's also Loryn's ideas; my idea about using the Footnote tool;  and MarkA's idea about using Explode.

Still, in my experience, students and professors struggle more with drafting than with analysis.  It's in the idea-and-communication-development phase that many good projects fail to be completed.

So to me, a clearer understanding of how Tinderbox can support  analysis would add a fillip, but only a fillip, to Tinderbox's basic functionality as the tool for my notes.
Back to top
 
 
  IP Logged
Charles Turner
Full Member
*
Offline



Posts: 180
New York, USA
Re: Tinderbox for Textual Analysis
Reply #42 - Jul 7th, 2009, 11:18am
 
Quote:
Still, in my experience, students and professors struggle more with drafting than with analysis.  It's in the idea-and-communication-development phase that many good projects fail to be completed.

Sage words, Jean!
Back to top
 
 
WWW   IP Logged
Mark Anderson
YaBB Administrator
*
Offline

User - not staff!

Posts: 5689
Southsea, UK
Re: Tinderbox for Textual Analysis
Reply #43 - Jul 7th, 2009, 11:29am
 
This isn't the whole process, but as Jean - if not others - are stuck into the process already, I've uploaded a TBX that is an exploration of 2 facets of our discussion. Breaking text down into word-notes and investigating how to view text that's been split. The file, made in v4.7.0, is here:
http://www.acrobatfaq.com/tbdemos/Gettysburg_Address_Text_Analysis.zip (see footnote**)

May I suggest that any questions arising from the 'mechinal', TB functional side, of what l demo in the file is taken froward in a thread of it's own so this can continue at the more conceptual level. From a (alter) reader's perspective threads get hard to follow when the theoretical and practical get too intermingled.

By the same token, if there's another aspect of this process that you can't figure in TB, please ask in a separate thread covering the procedure in question; we can cross-reference back to here as required. By all means contact me off list  with your "I got this far and then got stuck" TBXs if you don't want to post them more publicly.

Doing the TBX above was another pleasant reminder of how much you can to in TB with some lateral thinking. I'm aware the things I've shown aren't single-click-and-go but for that to be the case the whole process would need to be defined and locked down which would likely constrain some people's way of working. If there is interest in this sort of analysis in TB, I can certainly see some areas that likely can be made easier.

~~~~~~~
** new file uploaded (8:30pm UK time). Mark B quickly bowled out the problem.  I'd overlooked the fact I'd some Rules running using a '=' rather than a '|=' operator. The rules are resource  intensive and with = running all the time. I've amended the rules and added some notes to those already in the demo re using |= operators to ensure rules don't run unnecessarily. If you downloaded the file in the 4 or so hours since the original link was posted please download the new file (same filename).
Back to top
 
« Last Edit: Jul 7th, 2009, 3:33pm by Mark Anderson »  

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



Posts: 52
California
Re: Tinderbox for Textual Analysis
Reply #44 - Jul 7th, 2009, 4:38pm
 
Howdy. Although I have just downloaded Tinderbox and am considering buying it, I have been intrigued by this thread. My area of study is Biblical studies, exegesis, and history (with secondary interest in linguistics).

I would like to ask a question about this approach and whether I have misunderstood what has been proposed. I know that when approaching the study of a Biblical text, examining individual words, while interesting, can be misleading in terms of understanding the word in context. In other words, individual words considered individually may cover up the intended meaning of the author.

On the other hand, it seems that it might be better to look at relationships between words/phrases. Thus, it might be more helpful (in my analysis, for example) to see when a Pauline use of λεγω is following by an accusative or dative, or is εις used for purpose anywhere in the Pauline writings.

Is this the kind of analysis you are looking at, or has my old age overtaken me again.



... from an old codger who in college once learned the cutting edge technology of Fortran IV on an IBM 360, when it was the latest and greatest. Roll Eyes
Back to top
 
 
WWW   IP Logged
Pages: 1 2 3 4 
Send Topic Print