Recently, JFM was wondering about drafts. Do you use drafts? I don’t. My brain doesn’t register my own typos and missing words. Sometimes a kind soul like Björn Buckwalter edits my wiki and fixes some of my typos. Thank you so much! So what that means is that I publish early and I publish often.
As soon as most of what I want to say is there, I save it, and by saving it I also publish it. And then I reread it. It looks different when I read it outside of my regular editing buffer. Perhaps it’s the line wrapping, who knows. And I often return to posts after a day or two, and then if it was interesting, I’ll return even later. I’ll edit the main page if that’s what needs changes, and I’ll add comments if I found new material, sometimes just links and excerpts.
For a while, I was worried about the feed. These days, I try to use the following strategy: the first edit is a “major” edit and gets noticed in the default feed and is what you’ll see in the default list of Recent Changes. Future edits of my own pages are marked as “minor” edits, excluding them from the default feed, and from Recent Changes. There are ways to find them, and to subscribe to them, but I think for casual readers that’s not too important. Any comments I leave always count as a major change, but the default feed doesn’t list comment pages.
Back when I thought I needed drafts I implemented a system I felt was “simpler”: the wiki simply offered a single draft file per user as determined by your username (which isn’t protected). But then I ended up never using it. And I don’t edit text files locally with the intention of saving them “eventually”. I’m typing this here, without saving it, hoping that my editor makes automatic backups, and then I’ll post it, immediately.
JFM mentioned how hard it was to work on multiple drafts in a folder of essentially independent files using a version control software that uses change sets. That makes sense to me. I guess what I would do is use RCS, the predecessor of CVS. It’s a single file version control system. The only strange thing is that per default, there are no files in your working directory. It’s empty, except for the invisible “.RCS” directory. You use “rcs checkout FILE” to check out a read-only copy of FILE. Now you have a file to read. You use “rcs checkout -l FILE” to lock the file (important on a multi-user system). Now you can make changes to the file and when you’re happy, use “rcs checkin FILE” to commit the file. You’ll be asked for a commit message. And then the file is gone again. You’re free to checkout a read-only copy again, of course.
The commands “ci” and “co” do the same thing with less typing.
It’s weird to see that this super old version control system is still available on my system. Good times! 😀
Anyway, for the longest time I kept my config files using RCS. Their changes are all independent of each other, after all. But these days I use git for everything. It’s just less of a mental burden, I guess.
Strategy for writing drafts in a Gemlog?
#Programming #Wikis
(Please contact me if you want to remove your comment.)
⁂
As I’ve mentioned before (in the comments to Blogging Spirit) I think one of the coolest things about your blog is that it is a wiki. The coolest is your thoughtful and sincere writing, of course! 🙂
– Björn Buckwalter 2020-12-20 13:57 UTC
---
Thank you for the kind words! 😀
– Alex 2020-12-20 15:10 UTC