Today, Mark Bernstein suggested that someone look at the issue of linking with Tinderbox aliases.
I whipped together a quick email...
This *has* been an annoyance. I just figured I was doing something wrong, that if I ever really really needed to link to aliases, I would be able to read the manual and figure it out.
My guess is that link types could be very useful here. I could see where I might want to have the links point to things in a number of ways.
Of course, you could approach it at least two ways.
One way would be to add a new set of link types denoting behaviour, so you would get another dropdown when you create a link (not just a link to an alias, or a link to a note that has an alias. You need to be able to deal with the potential that an alias might be created)
You could also do something like this: each individual link gets its own behaviour defined, and a link prototype sytem could be set up, just as with notes. That way, I could preset some of the types of links. So, if I create a link called "it does not follow" for logic, I might have it only link to the individual note or alias. There is little need for the link to go anywhere else, so I can safely preset those settings (although they would be overridable, as with prototypes for Notes). But if I create a link type like "a relevant source", I might want it to take me also to the other places. If I go back to the original note, I might find related sources. If I look at the other aliases, I might find other things that quote or refer to the source. Thus, I might make "a relevant source" point to all aliases as well as the original note.
Key to this discussion is also the idea of backlinks. My email is already getting too long, so I won't go into this.
Another key issue is the idea of transclusion. Tinderbox doesn't fully support the idea of transclusion except in exporting. I imagine that a lot of issues get raised in this area that are similar to the question of aliases. Perhaps if Tinderbox merged the idea of aliases and transclusion (I have put thought to this, but not nearly enough), some of these issues would be resolved or at least clarified.
I sent thse comments to Mark Bernstein, who replied with a few thoughts...
But the reason link types were essentially abandoned for more than a decade was Randall Trigg's work, which showed that if you asked people to choose types when they made links, they would work hard to defer that choice. In essence, everyone winds up choosing "Type: Thing" and "LinkType: GenericLink" nearly all the time.
To my thoughts about the topics being rather related to the issue of transclusion, he said:
Also, similar to versioning in hypertext systems.
I have a few more thought-through ideas on the topic, but I want to think through them a bit more before I post them tomorrow.
But there are enough of us bloggers who use Tinderbox to come up with a really good way to answer questions like this. We only use the cathedral method of research because it is so expensive for people to get together and talk to each other. Why not use our blogging to discuss and decide on the answer?
Odd that writing technology makes extended, fair, thoughtful dialog and discussion possible, after ages of people derriding writing, we have come up with a way to discuss in real-time in writing, with ways of accessing more information more quickly, more ubiquitously than memory could ever serve. Plato's objections, if they ever had any weight, certainly cannot stick any more.
We live in exciting times. The very nature of human thought is changing before our minds.