3 ms·

Every time I read an article about attempts at non-hierarchical filesystems, I try to figure out how I'd take the huge piles of stuff I generate when I'm drawin

Every time I read an article about attempts at non-hierarchical filesystems, I try to figure out how I'd take the huge piles of stuff I generate when I'm drawing (and publishing) a graphic novel and reorganize it under tags. It's never pretty. Like, okay, sure, I tag everything with the name of the project, that's a no-brainer. But if I just do that then I get the hundreds of files I generate (one per page) mixed up with everything else - web-res renderings of each page, model sheets (and their source files), promotional material, stuff sent to publishers to try and convince them to deal with that part of the process, and the huge mass of files I generate for each book I print (which can be more than one for a single multi-year project): source files tweaked for print, print-res renderings thereof, files for the kickstarter for each book... So I tag all of these attributes too, and imagining putting all these tags on a file as I save it sure is a lot of fun, even if I imagine some sort of save requestor that keeps a list of all my previously-used tags, including ways to filter those - I don't care about any of the tags attached to my music collection or my collection of cartoon porn or my programming projects when I'm working on my comics projects, for instance, so I'd want to quickly narrow it down to just tags found in my art projects, and... Ultimately it just starts to look like a hierarchical structure in my mind, except for the fact that I'm interacting with it by some kind of tag-filtering file browser on top of a huge filesystem that mixes everything together in a non-human-browseable structure.

dgreensp 9y ago
I can strain my imagination and optimistically say that with a suitable UI, the tag-based file browser would be no worse than the hierarchical one — using tags to categorize your files in a not-too-coarse, not-too-fine way, and maybe even having more flexibility in how you organize and browse your files. Do you ever put information in the file name that could be a tag, or make up meaningless file names when there are few enough files related to a particular page, for example? Or put information in names that could be attached as some sort of comment or notes metadata instead? But I think you bring up some important issues when it comes to organizing your stuff. Would it be better if, instead of having kitchen drawers and cabinets, we had a tag-based system, because, you know, the structure of human knowledge and all that? Why should I be forced to put a utensil in zero or one drawers? Actually, the cabinets and drawers system is nice because you have a sense of “place” — you can think intuitively about where an item is, where an item goes; even if where an item goes is a somewhat arbitrary choice, at least you know exactly what decisions have to be made (which drawer and which compartment in the drawer organizer, for example) to put it away. You can also do a traversal through the cabinets and drawers to see what objects are being stored and how they are being organized, and any time you open a cabinet, you are focusing on a different set of objects. Imagine if you had 10 cabinets and 10 objects in each cabinet, but you actually only have 20 objects total. Every time you open a different cabinet, you see a different 10 of those 20 items. Confusing. I wonder if it would help to have required tags, and exclusive tags. Files with tag X must have tags of type A, B, and C, and may have other nonessential tags. X would be something like, “is a project file for some project,” and A could be the type of project name tags, and so on.

dgreensp 9y ago

I can strain my imagination and optimistically say that with a suitable UI, the tag-based file browser would be no worse than the hierarchical one — using tags to categorize your files in a not-too-coarse, not-too-fine way, and maybe even having more flexibility in how you organize and browse your files. Do you ever put information in the file name that could be a tag, or make up meaningless file names when there are few enough files related to a particular page, for example? Or put information in names that could be attached as some sort of comment or notes metadata instead? But I think you bring up some important issues when it comes to organizing your stuff. Would it be better if, instead of having kitchen drawers and cabinets, we had a tag-based system, because, you know, the structure of human knowledge and all that? Why should I be forced to put a utensil in zero or one drawers? Actually, the cabinets and drawers system is nice because you have a sense of “place” — you can think intuitively about where an item is, where an item goes; even if where an item goes is a somewhat arbitrary choice, at least you know exactly what decisions have to be made (which drawer and which compartment in the drawer organizer, for example) to put it away. You can also do a traversal through the cabinets and drawers to see what objects are being stored and how they are being organized, and any time you open a cabinet, you are focusing on a different set of objects. Imagine if you had 10 cabinets and 10 objects in each cabinet, but you actually only have 20 objects total. Every time you open a different cabinet, you see a different 10 of those 20 items. Confusing. I wonder if it would help to have required tags, and exclusive tags. Files with tag X must have tags of type A, B, and C, and may have other nonessential tags. X would be something like, “is a project file for some project,” and A could be the type of project name tags, and so on.