Re: Leo for organizing notes? [Comments Item 4]
On Saturday, February 8, 2020 at 2:15:42 PM UTC-5, andyjim wrote: > > Seems this issue needs a lot of thought. Niklas Luhmann's zettels had > numerical ID numbers, without textual clues as to their content. And it > was a paper system. And he certainly didn't work by remembering filenames > (he had 75,000 zettels) > > I understand your wanting individual text-based filenames, in order to be > forward-compatible with an uncertain future. I get that and agree with the > principle. But it appears to me that while that idea is forward-compatible > it's not current-compatible with a software-based zettelkasten. How do we > resolve this? Well, you suggested an optional, user-entered title as the > UID. What if the system could (optionally) generate a separate file using > the zettel title as the filename? The reason I say 'optionally' generate > that file is that in my case I do intend to use titles, but they won't be > unique. I might use the same title for a hundred different zettels, some on > entirely different subjects. > I think we're pretty much on the same page here. There seem to be three similar but different things in play: 1. Unique IDs for each zettel; 2. Titles; 3. File names for a putative move to to another system or even (temporarily, we hope!) a manual system. 1. IDs are easy. I happen to prefer ids that are more or less readable, but that's not all that important. They are helpful when debugging, for example. What I've suggested has worked well for me before: make an id out of the title, with modifications if that title is already in use. If a title is missing, make up an arbitrary ID. Luhmann used unique indexing expressions, mainly so he could find things again, and refer to them from other zettels. A hierarchical system like the one he devised helps in keeping related zettels - like child notes - quick to find. As you say, since we expect to keep relationships in the system, we don't need indexing terms to play that role. 2. Titles are good for quick visual recognition, and as a memory aid. They can also be the basis for all kinds of later analysis, such as generating clusters, or connecting different thoughts because they have something similar in the titles. The system can check to see if a title is unique, and modify it or just let you know if it's being duplicated. I suggested making a title out of the first line of the zettel, if the system sees that there is no title. Titles wouldn't be used by the system to refer to anything, so one could always change a title later. 3. File names. I do think that for export, it would be a good idea to use the title in the file name - assuming that the export method generated one file per zettel. That assumption could be revisited. For example. you could have all the zettels in a single file, separated by some distinctive mark. But *if* it were to be one file per zettel, then the system could make sure it didn't duplicated any by modifying the filename to avoid a name collision. E.g., a duplicate "U.S. Navy in WW2.txt" filename could become "U.S. Navy in WW2_02.txt". I'm with you that any and all of the metadata, including id and title, should be optional when the zettel is typed in. Let the system work it out if the information isn't included. The only item that can't be changed later is the id. -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/leo-editor/3a1e53fe-09e7-495b-b20f-a016a16a6233%40googlegroups.com.
Re: Leo for organizing notes? [Comments Item 4]
Seems this issue needs a lot of thought. Niklas Luhmann's zettels had numerical ID numbers, without textual clues as to their content. And it was a paper system. And he certainly didn't work by remembering filenames (he had 75,000 zettels). I don't think it could have worked with textual, meaning-based filenames---not for that big a system (and I have the same problem, not on his scale but I will at least be in the 10s of thousands I think). How did he do it? Well, we know he had an indexing system, though we don't know much about it. And we know that his zettel ID numbering system itself created 'clusters', so the zettel IDs served as a sort of internal indexing system, which was in turn connected with his external indexing system (I'm assuming his indexing was external to the zettelkasten itself. The software zettelkasten systems I've seen simply generate a time-based numerical UID. Fine for the system, but gibberish to the user, so they do not employ Luhmann's type of indexing system. Maybe we need to figure out how to employ those principles. Maybe some sort of indexing system is what I'm fishing for when I talk about 'mapping'. I'm going to be thinking about this. I understand your wanting individual text-based filenames, in order to be forward-compatible with an uncertain future. I get that and agree with the principle. But it appears to me that while that idea is forward-compatible it's not current-compatible with a software-based zettelkasten. How do we resolve this? Well, you suggested an optional, user-entered title as the UID. What if the system could (optionally) generate a separate file using the zettel title as the filename? The reason I say 'optionally' generate that file is that in my case I do intend to use titles, but they won't be unique. I might use the same title for a hundred different zettels, some on entirely different subjects. Hard to conceive? Well here's an example for a PIM system: a file called Books of Interest. Now suppose my PIM system covers a number of different categories, in each one of which I want a list of books of interest. In my system, which is a thoughts system, not a PIM system, there will be thoughts, concepts, ideas, themes that occur in many different contexts, hence the high likelihood of identical zettel titles in many contexts. That's why I need much more than just a simple filename in order to locate any given zettel in its full context, to say nothing of my inability to remember thousands of files. This is the magic of the zettelkasten system. So in my thinking, using a zettel's title for its UID won't work in this kind of system (maybe I should say in *my* use case of this kind of system). But, having said that, I certainly can buy into what I'm calling your forward-compatibility principle. So how about going with the best of both worlds, and provide an option to generate files by zettel title? They could be loaded into an archive as both a second copy of the files, and as a hedge against an unknown future. Or if you want, an *option *to use zettel titles as UIDs instead of having the system generate time-based UIDs (Let the system do it: I wouldn't want to type those in either). -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/leo-editor/a2a01618-5d6b-48a8-b0a9-a4f830973e77%40googlegroups.com.
Re: Leo for organizing notes? [Comments Item 4]
On Thursday, February 6, 2020 at 10:59:37 PM UTC-5, Thomas Passin wrote: > > Andy, thanks for your comments. I will give my reaction to them in a > series of posts, one per item. > Continuing my reactions to your comments - 4. Unique Identification @andyjim: "I propose YYMMDDHHMMSS as a UID format. I further propose that each section of a zettel use the UID of the zettel, with an added digit as an identifier for that section." Here I will disagree. Based on a lot of experience with my bookmarks system and other experiments with topic maps, I prefer identifiers that convey some information. To the software, they would still be arbitrary. Suppose you end up with a lot of zettel files but without the software. You want to find some notes relating to (Zettel Design Notes"). Wouldn't you rather scan the file listings for file names like "zet-design-1", "zet-design-2", and so forth, compared with "2020031412034522"? I would. And they are much easier to type, too It's true that the software can display a list of titles no matter what the identifiers are, and it's also true that computer-generated identifiers are easier to guarantee that they are unique (sorry for the awkward sentence structure there!). But my thinking here is dominated by the desire to have a system that can be transported to other software or, in the last resort, still be useful manually. After all, we're talking about possibly years of work stored in the zettelkasten. It *has* to remain usable. Also, if I were typing a note, I wouldn't want to have to type in identifiers that look like that. It's too hard and finicky. My concept at this point is that the user types in an optional ID and an optional title. If he doesn't include an ID, the system tries to make one from the title. If there were no title, the system would make an id from the first line of the text of the zettel. If somehow the zettel had been filed without any text - maybe as a placeholder to be revisited later - then the system would make up an arbitrary ID, possibly exactly what you suggest. The zettel would still contain the time, just not as the ID. -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/leo-editor/c90a7904-88c6-46ca-bb66-6ba29f074c9f%40googlegroups.com.