Forum Posts Feed
Software Request: Open Source Personal Content Repository
< Next Topic | Back to topic list | Previous Topic >
Posted by Franz Grieser
Sep 13, 2018 at 12:23 PM
Re: TagSpaces
>I know that the Pro edition uses a different mechanism but
>the free trial doesn’t.
My fault: The trial supports the sidecar files for tagging without interfering with the file names.
Posted by Amontillado
Sep 13, 2018 at 01:13 PM
I use a timestamp prefix on file names, particularly in my personal records/bookkeeping DevonThink database, but I’ve gotten on a plain text kick (again) lately, particularly for documentation and creative writing. Of course, I repeat myself. In the marketing department, documentation is creative writing.
DevonThink is a hard-to-break habit. It keeps files as files, so if DT vaporizes the files aren’t lost. Tags and replications, maybe, but not the files themselves, and I’ve made an interesting discovery.
Imagine a novel in a number of separate files in a group, comprising the chapters and scenes in your work.
If you set a group to “unsorted,” you can drag and drop the order of files in the group. Conveniently, the order you set is persistent. You can play with a manual sort, switch to sort by name (or any criteria), and then when you switch back to unsorted you get your manual sort again. Newly added files go at the end of your manual sort.
So, a collection of chapter and scene files can be safely sorted into the correct order. “Compiling,” as Scrivener-speak would have it, works pretty easily with Pandoc or something similar.
Pick a view that shows the folder list and the files within folders and switch to unsorted to get your custom order. Click on a file and use Command-A to select them all. Hit Command-C to copy them.
Now bring up a terminal window. When you hit Command-V at the command line prompt, you get the list of full path names to all your files, with spaces properly backslash-escaped.
To compile a bunch of markdown files in the right order to docx: pandoc -s -o gatsby.docx -f markdown -t docx Command-V
Pretty cool. Those pandoc options are -s (create complete file, not a fragment), -o (output to gatsby.docx), -f (from markdown syntax) -t (to docx format). Command-V pasted the list of source files into the command line. Hitting enter did the rest.
A pure plain text solution would be nice. I kind of like DT’s replicants, though. Like symlinks without worry of which is the file and which is the reference.
Posted by nathanb
Sep 13, 2018 at 03:59 PM
Simon wrote:
>I have come to the conclusion that my main problem is that I don’t stick with something long enough.
The rest of your post tends to disagree with you! I get what you are saying though. Your core Devonthink method is impressive and I’m jealous and impressed by your archive. This is my favorite kind of description, of what someone has actually done for years. That’s what separates concepts, of which all of us here are drawn to regardless, and viable action. Thank you for your post. I’ve found it helpful.
Posted by nathanb
Sep 13, 2018 at 04:52 PM
Franz Grieser wrote:
>You also mentioned TagSpaces in the thread. Do you (still) use it? I
>find the concept appealing but wouldn’t want TagSpaces to mess with my
>files’ names. I know that the Pro edition uses a different mechanism but
>the free trial doesn’t.
Ha! No I don’t actually use it. I’ve tested it out a few times and it is one of my favorite options to possibly transition to. I like it’s open nature and because it embeds everything in the filename you don’t have to worry about a particular installation breaking or being tied to one workstation. I wasn’t put off by it extending filenames, I’m quite used to butchering filenames with Z-codes anyway and actually like that seeing intentional tagging/z-coding following the actual name is itself information that tells me that file is special enough in some way to have warranted this extra forced layer of metadata. I mean, we have tens of thousands of files littering our hard-drives…er…storage mediums. Only of which a few hundred/thousand enough are personally relevant enough to be ongoing ‘reference’ material. I’m going to want to add some kind of marker to differentiate the signal from the noise.
The reason I don’t use tagspaces is because my zettel-stamping of some of my files/folders has proven ‘good enough’ organization. This is mostly because I’ve given up on straightening out my directories and just consider that a ‘back-end’ to my info system. Like 3% of my files are ones I seek out and open directly on a regular basis by drilling down into the folders. Basically my current hot projects. I rely on my note systems (currently OneNote and Notion) and task manager (currently todoist) to be my index to most my stuff.
There is of course a lot of value in having a well-organized file system like some other people here have described (my hat’s-off to them!). It’s been on my to-do list for years now. I’ve started it like 30 times and instantly run into roadblocks of deciding whether this tax document goes under ‘financial’, ‘2017’, ‘archives’ etc. Or if this manual to my mower should be associated with the receipt and therefore by month bought or in some ‘house-maintenance’ folder. So I just think in circles for a few minutes, get frustrated at the inherent limitations of single-hierarchy filing, give up and do something else. If I ever decide to finally suck it up and spend an entire weekend whipping my files in shape, it’s definitely going to be with tagspaces…because that’s the only tool I’m aware of that allows me to organize via multiple categories and doesn’t lock me into a proprietary system.
I don’t like the idea of relying on simple search in lieu of intentional organizing, but that’s my reality. However, it turns out that blind search and my scattered z-codes have proven to be reliable. I can usually find what I need, when I need it. So I don’t know if it’d be worth the efficiency/clarity gain to intentionally tag/file ALL of it and keep it that way. I’m starting to come to peace with most of my digital mess remaining a mess, with the occasional z-code ensuring findability of the neat stuff. Then selectively putting structure on a select few subsets of things. I’m finding Notion.so to be quite helpful with that. It lets me dump-notes away in my normal randomish One-Note fashion and occasionally embed real databases to add structure on-demand. That’s perfect for things like reading lists, product research, software research etc.
Posted by Donovan
Sep 21, 2018 at 07:05 AM
I think solutions for the entire concept described in the original post by ‘thought’ is probably becoming needed far more than simple outliners. We live in a world of mixed-media, and the need for instant access, far more than many of us simply need to file away writing in their proper tree structure (or whatever). Bringing together all of this content under one piece of software is probably not going to ever be a solution because of the enormity of the project. Because of this, I think we should be looking to regular file/folder structures in our preferred operating system and simply find a pleasant-to-look-at “front-end” that would serve as the Google of our local content - whatever digital content that might be.
Think about the internet and how it is organized. Right, it’s not. Some things for some content certainly are (Wikipedia springs to mind), but on the whole it’s just “out there.” It didn’t take long to discover that Yahoo’s Directory of the World Wide Web was a mess and what was needed was what is still the “front-end” of the web today—Google.com. We simply search and find (usually). But, we can bring intelligent naming, tagging, etc. to the locally stored content and bring some sense of order and near certainty that the front end search will being it up in milliseconds.
I like Schmid’s Notebooks maybe for the simple reason that it relies on the regular file system and brings things in a pleasant fashion to the notebook.
The original post described a need we see more and more.
Think on this:
- All content in regular files and folders
- Intelligently named for easy searching
- Super-tagging or meta-names
- Anybody who has used the free search program “Everything’ knows how fast that local content, of all kinds, can be cataloged and pulled up - as you type! ( https://www.voidtools.com ) Amazing software.
- Maybe there is a gleam of a solution that is as simple as utilizing our preferred tools to create our content, while using a lightning fast front-end (like ‘Everything’) to find and retrieve ALL of our content.
- Maybe changing the mindset from trying to finely tune the organization of our content in Yahoo-like directories of old, to a Google for our own stuff.
I know the above doesn’t address every issue described in the OP. What it *could* do is get us thinking about a change of mindset and think about how we can work within a search system (and allow for tabbed searching to bring up multiple pieces of content) while being able to connect one piece of local mixed-media content to the other. It may sound half-baked because, well, it is. But, I leave it here as food for thought. In other words, it’s an idea to a problem, it seems more and more of us are having, which is how we retrieve all of the mixed-media content we collect from simple text files, to images, HTML, video, presentations, on and on, without it all being in some proprietary enormous database. Not a “Room of our Own” but a “Google of our Own.”