I'm pretty confident that what we have now works.
I'm thinking about ways to get some of the glitches out of the process. Installing and building a Node app is a bit much, and editing a JSON file to configure the app isn't too great either. What I'd ideally like to do is turn the code over to the LogSeq developers and ask them to build it in to their app. That would involve porting it to Clojure I guess. Failing that, I wonder if there's a way to turn it into a plugin? I don't know anything about LogSeq plugins, so I'd have to turn the project over to someone who does. I'd be happy to answer questions about my JavaScript code, but I think it's pretty straightforward. I didn't take any shortcuts.
I guess to sum it up, if people really want to do this, we should make it easier.
And at least a couple of people were able to install the software and build a blog.
I learned a few things...
I made a mistake in the instructions, I didn't tell the user how to install Node. I just told them to run npm install as if everyone has it installed. One of our best users got stuck there.
I also need to help people debug incorrect JSON. The parser is not error-tolerant (which is good). I must include a pointer to jsonlint.com to give them a way to find out what their errors are.
If there is an error in config.json the app should exit, not proceed. There's no way anything good can come from going ahead at that point.
If someone leaves a trailing slash off a path that should not be the end of things. That's an "error" I can correct for them.
The Markdown functionality was a big win. I am using Drummer to edit this file, and it's mirrored in the LogSeq app, and using Markdown makes it look really nice over there. I now see why LS users expected Markdown support.
We need to get Windows and Linux E/D working asap.
Overall, this is not the way to provide this functionality
It's a bootstrap, so the first rendition is going to be a nightmare for non-technical users. It has to be that way or else we wouldn't know what to do on the second step.
This functionality must be baked into Electric Drummer. It would be nice if it were baked into LogSeq as well. The code is MIT licensed so there is absolutely nothing standing in the way of it. They wouldn't be hard-coding something that only works with Drummer. The API is something that would be easily replicated by other CMSes. Along the way they would of course have to support OPML, which imho is a good thing.
The first two people who tried using the myLogseqBlog tool tried to do something that hadn't occurred to me, they used Markdown in their posts. Makes so much sense that at first I wondered why it hadn't rendered properly. Then I realized that Drummer blogs have to be told to expect Markdown, it doesn't default to that.
So I've now made the change. The question is will it work?
Let us see -- there should be a numbered list under this node.
Indiana
Idaho
Italy
This should be bold and this should be italic.
Here's a link to a story about the Transcontinental Railroad.
Let's try an image just for fun.
Oh yes I'd say the markdown experience is working quite nicely. 😄
Two changes
When myLogseqBlog uploads the file, it takes care of making it public, so the manual step specified in the docs is no longer needed.
The nodes we create are of type markdown, so that the blog processor knows to render the text through a Markdown processor.
How to update
Download the folder as you did when you installed the app.
Unzip the folder, copy mylogseqblog.js from the newly downloaded folder into the previous folder. This preserves your config.json file which will not change.
Continue building as before.
Yes, this is technical
Eventually, hopefully soon, this functionality will either be baked into an easy to use app and all this will happen in a very natural way.
I am a total beginner in LogSeq and an expert in Drummer, and this part would have been very hard for the LS people to do, so I went first.
After a period of using it, we should put our heads together based on the experience and figure out how to package this for maximum ease of use and power.
I'm just glad to have LogSeq users and Drummer users sharing functionality. This is a first step in what I hope will be a long collaboration! ❤️
5:32PM, just for fun, I'm putting an image in the right margin of this post. That's a picture of my favorite character from The Expanse. Their final episode airs this Friday on Amazon Prime. #Where we're at to start.
I have a JavaScript app that turns a Journals folder into a calendar-structured OPML file.
This is the format that Old School uses to build a blog.
Today's project will be to get the stuff I'm writing now to appear on a blog.
And to make updating the blog as easy as possible.
Note that all this assumes that we're
Running Electric Drummer so we have access to local filesystem.
And Electric Drummer for now only runs on the Mac.
Today I'm going to try an experiment.
I'm going to use LogSeq to write a new blog which I will publish through Drummer.
To do this, I will take advantage of the fact that I have Electric Drummer running on the same machine as LogSeq, my desktop Mac.
My code will read the LogSeq journals folder, and infer from the names of the files what date they correspond to. Then I will load each file into a section of a Drummer calendar structure, and then publish the outline as a blog.
I will report here on my progress.
I created a new Twitter account called mylogseqblog and logged on to Drummer using that account.
I found my twitter oauth token and tokensecret by typing localStorage into a headline and pressing cmd-/.
I copied the values of twOauthToken and twOauthTokenSecret to where?
Wrote a script that builds an outline from the contents of my LogSeq journal folder
It works!
Now the question is how to deploy this mofo.
I think to start I'll make a new project called Drummer and LogSeq.
The first thing in that project will be a JavaScript app that turns a Journals folder into an OPML file that's ready to flow through Old School.
Of course Old School will not be the only processor that can use this functionality.
I have code that reads my Journals folder and turns it into an outline ready to be fed to Old School.
Tomorrow I will figure out how to tell OS to do its thing with this outline.
I'm pretty confident that what we have now works.
I'm thinking about ways to get some of the glitches out of the process. Installing and building a Node app is a bit much, and editing a JSON file to configure the app isn't too great either. What I'd ideally like to do is turn the code over to the LogSeq developers and ask them to build it in to their app. That would involve porting it to Clojure I guess. Failing that, I wonder if there's a way to turn it into a plugin? I don't know anything about LogSeq plugins, so I'd have to turn the project over to someone who does. I'd be happy to answer questions about my JavaScript code, but I think it's pretty straightforward. I didn't take any shortcuts.
I guess to sum it up, if people really want to do this, we should make it easier.
And at least a couple of people were able to install the software and build a blog.
I learned a few things...
I made a mistake in the instructions, I didn't tell the user how to install Node. I just told them to run npm install as if everyone has it installed. One of our best users got stuck there.
I also need to help people debug incorrect JSON. The parser is not error-tolerant (which is good). I must include a pointer to jsonlint.com to give them a way to find out what their errors are.
If there is an error in config.json the app should exit, not proceed. There's no way anything good can come from going ahead at that point.
If someone leaves a trailing slash off a path that should not be the end of things. That's an "error" I can correct for them.
The Markdown functionality was a big win. I am using Drummer to edit this file, and it's mirrored in the LogSeq app, and using Markdown makes it look really nice over there. I now see why LS users expected Markdown support.
We need to get Windows and Linux E/D working asap.
Overall, this is not the way to provide this functionality
It's a bootstrap, so the first rendition is going to be a nightmare for non-technical users. It has to be that way or else we wouldn't know what to do on the second step.
This functionality must be baked into Electric Drummer. It would be nice if it were baked into LogSeq as well. The code is MIT licensed so there is absolutely nothing standing in the way of it. They wouldn't be hard-coding something that only works with Drummer. The API is something that would be easily replicated by other CMSes. Along the way they would of course have to support OPML, which imho is a good thing.
The first two people who tried using the myLogseqBlog tool tried to do something that hadn't occurred to me, they used Markdown in their posts. Makes so much sense that at first I wondered why it hadn't rendered properly. Then I realized that Drummer blogs have to be told to expect Markdown, it doesn't default to that.
So I've now made the change. The question is will it work?
Let us see -- there should be a numbered list under this node.
Indiana
Idaho
Italy
This should be bold and this should be italic.
Here's a link to a story about the Transcontinental Railroad.
Let's try an image just for fun.
Oh yes I'd say the markdown experience is working quite nicely. 😄
Two changes
When myLogseqBlog uploads the file, it takes care of making it public, so the manual step specified in the docs is no longer needed.
The nodes we create are of type markdown, so that the blog processor knows to render the text through a Markdown processor.
How to update
Download the folder as you did when you installed the app.
Unzip the folder, copy mylogseqblog.js from the newly downloaded folder into the previous folder. This preserves your config.json file which will not change.
Continue building as before.
Yes, this is technical
Eventually, hopefully soon, this functionality will either be baked into an easy to use app and all this will happen in a very natural way.
I am a total beginner in LogSeq and an expert in Drummer, and this part would have been very hard for the LS people to do, so I went first.
After a period of using it, we should put our heads together based on the experience and figure out how to package this for maximum ease of use and power.
I'm just glad to have LogSeq users and Drummer users sharing functionality. This is a first step in what I hope will be a long collaboration! ❤️
5:32PM, just for fun, I'm putting an image in the right margin of this post. That's a picture of my favorite character from The Expanse. Their final episode airs this Friday on Amazon Prime. #Where we're at to start.
I have a JavaScript app that turns a Journals folder into a calendar-structured OPML file.
This is the format that Old School uses to build a blog.
Today's project will be to get the stuff I'm writing now to appear on a blog.
And to make updating the blog as easy as possible.
Note that all this assumes that we're
Running Electric Drummer so we have access to local filesystem.
And Electric Drummer for now only runs on the Mac.
Today I'm going to try an experiment.
I'm going to use LogSeq to write a new blog which I will publish through Drummer.
To do this, I will take advantage of the fact that I have Electric Drummer running on the same machine as LogSeq, my desktop Mac.
My code will read the LogSeq journals folder, and infer from the names of the files what date they correspond to. Then I will load each file into a section of a Drummer calendar structure, and then publish the outline as a blog.
I will report here on my progress.
I created a new Twitter account called mylogseqblog and logged on to Drummer using that account.
I found my twitter oauth token and tokensecret by typing localStorage into a headline and pressing cmd-/.
I copied the values of twOauthToken and twOauthTokenSecret to where?
Wrote a script that builds an outline from the contents of my LogSeq journal folder
It works!
Now the question is how to deploy this mofo.
I think to start I'll make a new project called Drummer and LogSeq.
The first thing in that project will be a JavaScript app that turns a Journals folder into an OPML file that's ready to flow through Old School.
Of course Old School will not be the only processor that can use this functionality.
I have code that reads my Journals folder and turns it into an outline ready to be fed to Old School.
Tomorrow I will figure out how to tell OS to do its thing with this outline.