Forgive what will likely appear as a rudimentary problem, but I tend to learn by examples and I’m trying to wrap my head around this one…
In researching things like grants, I’d like to see the people and the organizations connected… Often times, if I look at XCorp Community Chest and see FamiliarPerson, I start to wander off looking for the other places and where I met them. While it is a fascinating and often interesting trip through many rabbit holes, I need to be tidier in my notes so I can get back to whatever I started on before trying to remember where I saw a face before and being distracted to see that they know InterestingPerson2 that I went to college with, etc…
I’d like to be able to take notes of Grant (amount), by Org, to Org, notes
For organizations: Groups Name, type, people related (by type of relation, i.e. director, employee, volunteer, etc)
People: connections to orgs and groups, notes, roles,
From there, I wonder if I could see a map that clusters people around affinity – UnknownPerson connects to InterestingPerson by way of NewCommunityChest, etc…
I know it’s possible, but I stammer at getting it set up! any advice greatly appreciated… I suspect I don’t do prototypes correctly.
Rather than starting with a map, which requires some preparation to get going, I would start with Attribute Browser and use it to explore interesting queries.
I assume your notes include attributes and/or links with values that can be held in common across multiple notes to aid the exploration.
Short of magic, any software is going to need something to go on. As you’ve noted, prototypes are yous friend. They enable:
identification of all notes of a common type (e.g. people, projects, etc). This is even more useful if the notes of a type aren’t all in one part of the document.
visual identification of notes of a type: consistent colour, shape, badge, flags, etc.
consistent Key Attributes - these will likely vary by type
they assist in emergent structure, things that are of a similar type are different: people, vs orgs, vs products, etc.
Storing strands of data in attributes (which don’t have to be Key Attributes†) either as well as in note $Text or as well as it, can help a lot as you start exploring structure. Why? Because it makes it much easier to query things. Otherwise any query becomes a regex ‘contains’ query. Also, once a an attribute has range of values it makes use of Attribute Browser (as @PaulWalters has rightly already suggested above) easier. Plus, you can use the values() action code to collect a list of unique attribute values. If you pass that to a note’s text and explode it, you have per-value notes. I use this technique a lot to do things like extract—via values()—all the authors for some papers/objects/whatever, and split them out into pre-author notes which I can then link back to the relevant per-paper/projects/etc. notes. This works well as the amount of information grows.
Also as your dataset grows, keeping it all on a single map (i.e. in a single container) becomes a chore. This is where hyperbolic view can help. It only shows notes with (in-document) links but shows them regardless of where they are in the document.
† A common misconception amongst some users is the attributes only exist if they are shown in a note’s Key Attributes. Wrong! Key Attributes are simply those attributes—system or user—that you choose to display for a given note when it is selected, i.e. in the text pane. Inclusion in Key Attributes has no effect on an attribute’s value. If you set Key Attributes for a prototype it formalises this display for notes using that prototype as they inherit the latter’s Key Attributes.
I’d start this task in Map view, though as @PaulWalters points out, Attribute Browser is going to be handy too.
I’d begin by choosing a couple of Grants and making notes about each. I’d like a distinctive visual appearance the the Grants: perhaps they’re all light green. I’d make a prototype at this point that would define properties that all Grants share.
Then, I’d make notes about the Organization behind each grant. Again, I’d link Organizations to Grants, and give the Organizations a distinctive appearance. What I’d be looking for at this point is what sort of information I really want to record about organizations: do I mostly just want the name, or a few formal details, or do I need 8x10 glossy pictures with a paragraph typed on the back of each one.
I’d arrange these however made sense at the moment, and rearrange as needed. Do you have a bunch of different contacts at the NSF? Maybe we need to divide Organizations into Programs? Or, do we want to have a good way to emphasize that Prof. Higginbotham is on the board at XCorp and also the VP at Acme Industries?
The point here is not to Get It Right the first time, but to (a) record stuff right now, so this gets big fast, and (b) learn what is really useful — and what questions you actually want to ask right now.
Thank you for the encouragement and the advice. Here’s a stumper for me: how do I do the links?
Let’s take Prof. Higgenbotham — I can make a “People” prototype note, and give it the values for {name, notes, etc} but what is the best way to link him to Organizations Xcorp and Acme? Is it a note that describes his relationship to that org? If so how to connect it?
If it is just dragging links to them, then how to add the information about his connection to the organization {VP, board director} etc?
I’d start by (a) just dragging a link from the person to the organization. If the nature of the relationship is significant and not obvious, I’d just describe that in the text.
That might be plenty. Or, maybe you’ll find that you want more information. I remember a restaurant reviewer who kept copious notes about each visit, but found she sometimes forgot to note down things that would ALWAYS be needed — phone number/credit cards accepted/reservation policy, stuff like that. And those missing bits were always found a couple of hours before press time: that made it good to add structure that always would capture those required details.
On the other hand, if you ask for too much metadata all the time, sooner or later you’ll be in too much of a hurry and then you’ll stop collecting it consistently, or at all.
Okay – I think I got the gist enough to move past paralysis here.
On the links, though, I’m considering using an intermediary note… So Note (organization prototyope):Acme Corp --> Note:Employees --> Note (VP, year began, etc) --> Note: Person, Professor Higgenbotham
That will alllow me to capture information about the link and aggregate commonalities …
but for Note (organization): XCorp --> Board (any notes about composition, terms, etc) —> Note: Director, years, notes) —> Professor Higgenbotham
I’ll try it! I just went through the steps to update to 8.7 so I’m going to try and avoid reading the release notes and playing LOL
I’m just coming back to say thanks. I’ve moved along quite a bit and learned a lot along the way by forcing to do it.
I am finding a lot of practical things I don’t know how to do or find frustrating trying to figure out.
-how can i marquee select a bunch of notes on the map view to move them/arrange them?
-is there a magic step involved in getting attributes pasted into note text automatically (AI) recognizes? Ex: if I drag a link from safari into the map, or paste one, then the Net/URL -attribute is automatically populated with the value. HOWEVER, it I explode a note that has a list of links, the exploded notes do not have URL auto detected How to have that happen for 50+ exploded notes?
The detection of the URL in the first case is because you are dragging information into the app. Tinderbox has to make sense of it, finds a URL and stored if. By comparison, ‘Explode’ is designed to split large sections of text based on a given delimiter. There is no reason store scan the text. What if a new note has 2, or more URLs, which is used. It’s easy for us as a user to idealise a ‘simple’ scenario, but as the latter questions show it is far less simple than imagined.
That said, if you could provide some representative examples of the exploded notes I’m sure it might be possible to do something to help automate ‘extracting’ URLs.
The bolded text is a hyperlink. I want to keep organization name. I would love to know how to automagically extract the address to attribute values
Scenario 1: If I highlight the whole block of text for ORG1 in safari, and drag that text to the map canvas of tinderbox, the result is that I get a note with all the text in the note. Title is ORGANIZATION1 and the URL attribute has the hyperlink value already set. I like this result.
Scenario 2: I highlight a list like that in safari, copy it, paste it into a new note in tinderbox. I get a next whose text is the list. The bold hyperlinks are present. Then, I run explode and set the delimiter to \n\n And tell explode to set the title to the first sentence. The result is that I get exploded notes for each list item (expected behavior, I like it). The title of each note is the respective ORGANIZATION name (I like that). The text of the result is the original item — the first line in the text remains ORGANIZATION as a hyperlink. (I like that). But, the URL attribute has not been set as it automagically happened in scenario 1.
Q1: what procedure should I run on the exploded notes to cause the automatic setting of the URL from the ORGANIZATION hyperlink (if present)?
Q2: is there a similar procedure that will detect the address and set attributes and map coordinates, etc?
This could be problematic. Not 100% sure (without seeing the doc) but it sounds like you have RTF links in your text. If so, this creates a further problem as such links aren’t Tinderbox web links but are smart links and currently the latter can’t be accessed via export or action code. I say this as those systems can’t then ‘read’ the URL.
I’m not in a position to try now, but perhaps AppleScript could read the notes in turn, read the URL target of the link in each note’s $Text and set that as the $URL value for the note (and make $URL a Key Attribute too).
They come from web pages (Html) and I can’t think of any steps that would have made them into RTF links. Unless tinderbox does that to note text, but it seems like it keeps what it gets.
So, maybe it’s just easier for me to copy/paste each of 50 items as new notes. (Scenario 1).
I think I’m surprised I can’t trigger that same magic later.
That leaves Q2 - I’m guessing this would be applescript to find that info. I’ve never played with it, but can applescript export the autodetected address? I get the grey box around the addresses so some system is an address
any advice greatly appreciated… I suspect I don’t do prototypes correctly.
I just went through the steps to update to 8.7 so I’m going to try and avoid reading the release notes and playing LOL