How to use the linking productively

[srdiamond]

The discussion has benn possibly excessively abstract, mine included, in that no one has mentioned the actual nature of the links people find helpful to forge. In an ordinary outline, links are parts of hierarchies. In an outline with cloning, they are part of multiple hierarchies. But UR offers still more flexibility than that, and I wonder whether this flexibility is or is not an asset.

UR encourages links not only based on headings that the user creates for some classificatory purposes, but links between bare items. These links can even be circular, so that an item can be both parent and child of the other, in different respects, one should hope.

Is this actually useful, except for a rare situation? When you make all your linkages in UR, are you apt to see any kind of patterns emerge? What have users learned by linking anarchically? Or are the masses of linkages more confusing than the original data?

[PureMoxie]

Stephen,

I've been using the links to provide context for information separately from storage. For example, I may store all project correspondence in a folder called 'Correspondence' below the project item, but if a message relates to a particular task, I can copy and paste it below the task as well.

Similarly, I can have contacts that show up under a 'Clients' item and also under the individual projects they relate to.

I realize these are all fairly standard uses of cloning/linking. The UR linking approach seems open enough that it could be used in some creative ways.

One limitation to the general approach is that you have to remember to manually create links to build relationships.

[srdiamond]

Let me see if this discussion might go forward better around some more specific issues in linking.

I don't know whether it makes sense to test the limits of UR's liking capability. Something that's technically possible isn't necessarily conceptuably meaningful. Of course no one is forced to use all the possibilities afforded by logical linking.

For instance, you can link documents to folders or other documents (among other possibilities), but nothing stops the user from setting as a rule to himself to forego linking documents to documents. The user could conceivably create the kind of a priori folder hierarchy, per Idea!, to implement facet analytic concepts in knowledge management. [The user might then choose between the two programs on other criteria, such as in which the ergonomics of linking is better, which has the superior writing environment (UR), or which has the better external linking (Idea!). Or which allows you to use the linkages directly to find documents with Boolean operations prefigured (Idea!).]

PureMoxie does not indicate he links documents to documents. I wonder if others find document to document linkages more than trivially useful.

[PureMoxie]

Stephen,

Since I have used UR for about three weeks, so far I haven't run into the need to link documents much with each other (if, by documents, you mean included/linked files of the Office/PDF variety). But I'm sure I will use this in the future.

Let's imagine I am writing a proposal. Each draft of the proposal could be a document and also have as children links to other documents that I refer to when building the proposal (spreadsheets, etc.). Each item, whether a document or otherwise, can "live" with the other items it relates to, one or many times.

I may have misunderstood your point if you are referring to documents in a more general sense (any written text item). If you intended this more expansive meaning of document, then I do link these together quite freely.

The only trouble I have with the basic approach of linking manually is that I must remember what sort of structures I am trying to build. Saved search items relieve some of this burden. Still, manual linking is not quite to the level of "assignment" in Zoot, where you build in some initial structure and then it's very easy to "link" items to certain concepts without needing to navigate around a tree.

[srdiamond]

Yes, that's an interesting use I hadn't considered. I had been thinking of UR as a database, not a "writing tool." But a hierarchy of documents could be part of a process of modular writing, as in programs devoted to such a process like ndx Cards, Miss LonelyNotes, or even Writers Blocks. To turn UR to that end, supplanting these other programs, I would await UR's upgrading its export capabilities.

One thing UR does offer to help remember what you are trying to build is the notes card attached to every item. I Would suppose one of its purposes is to annotate the nature of the links.

What is very important, it seems to me, is to have some concept of the data structure intended. I can imagine many users including myself connecting away in the hope that some structure will emerge inductively. I think this will only work within some constraints that the user rationally imposes. The difference between the programs UR and Idea!, coming back to what's to me the most useful contrast because of their basic similarity, is that Idea!'s developer is apparently in the knowledge management field and seems to have built constraints into the program to avoid creating useless structures. UR, if I can speculate, was written by a crack programmer who provided the means to implement whatever structure the user desires.

To make the need for system concrete, consider even something so trivial as the relationship between a document and its sources. Which is parent and which is child. Logically, it makes sense either way. Does it make a practical difference which way one does it? I don't know, at least not without fiddling, although it occurs to me that if the hierarchical relationship is something other than set inclusion--in other words, the hierarchy is used to model some relationship not inherently hierarchical--then the mapping is arbitrary and equivalent, whichever side is called parent and which child. The point here is that the user must make a conscious choice about even simple matters, although fortunately, many of these choices, once made, will seem natural and inevitable.

Originally posted by PureMoxie, quoted selectively herein

Let's imagine I am writing a proposal. Each draft of the proposal could be a document and also have as children links to other documents that I refer to when building the proposal (spreadsheets, etc.). Each item, whether a document or otherwise, can "live" with the other items it relates to, one or many times.

The only trouble I have with the basic approach of linking manually is that I must remember what sort of structures I am trying to build. Saved search items relieve some of this burden. Still, manual linking is not quite to the level of "assignment" in Zoot, where you build in some initial structure and then it's very easy to "link" items to certain concepts without needing to navigate around a tree.