Hi Puming,
this all sounds like good points to move the project forward and I would very much appreciate any of your contributions. Indeed the only reasons why I didn't commence to go further with this yet, is that so far it was good enough for its current purpose and that the other projects drain all my development resources. But since recently I have reserved some hours per week for various open source projects, so I can try and help a bit here and there (but some others, like DDOX and vibenews unfortunately also need to fit into that schedule).
To your suggestions, I would add one additional thing that has a relatively high priority: The comment system needs to get a spam filter (e.g. using antispam) and a better editing UI, as well as proper e-mail notifications for commenters. Such a comment system could then also be nicely reused for the new Phobos documentation instead of Disqus.
But by checking out the project I found that some of the things are lacking for a personal blog engine as compared to ghost:
- no style. I would be nice to have a default style.
Agreed, this is probably the most notable thing that is currently missing. A possibility to switch between different styles using the management page would also be nice.
- sparse documentation.
- no binary download.
We finally have a good CI server running, which I could use now to automatically upload binaries for various platforms (similar to DUB). This is something that I could start to tackle over the next weeks, possibly together with a little home page (e.g. at vibelog.vibed.org).
- dependency on mongodb. I'd also like that posts could be saved into files (like jekyll & github blogs) or simple dbs like sqlite.
Sounds like a good idea. Together with this it would also be good to use the userman library, which is also used for the DUB registry and for the forum, to manage user credentials and make that DB engine independent as well. This would be something that I could look into as well.
- syntax highlighting and other tools are not incorporated
What I'm using now is JS based syntax highlighting with google-code-prettify. This requires a little text filter to be installed (blogSettings.textFilters ~= html => html.replace("<code><pre>", "<code><pre class=\"prettyprint\">")), but Brian Schott's planned std.d.lexer looks like it would be perfect for server side highlighting and I'm also planning to incorporate that for DDOX and the new Phobos documentation, so that would be a low hanging fruit.
I know the situation may due to you are focused on vibe.d/dub and other higher priorities, so I'd like to contribute to it.
I've add bootstrap and a theme similar to ghost the other day, you can view a demo on http://zhaopuming.cn
Looks very nice and clean. If you want to contribute this or a similar variant, I would be happy to add it.
What is your vision of this project? Do you want it to become a useful blog tool for not only your vibe.d site, but also a general purpose blog engine?
In fact, it's already used for a couple of small blogs and while there will always be a strong focus on the embeddable property, I'd definitely like to improve it as a stand alone system, too.
Here is what come into my mind first:
- bootstrap based themes. It should have a default theme, and people could contribute their themes. By doing that I'd like to change the class names and html styles into a more bootstrap flavored one.
I'd be OK with that, I'll have to adjust a couple of styles, but it's always good to settle for a kind of standard.
- A better editor. I love markdown, so it would be nice to have a markdown editor with live preview. similar to stackedit
Good point. That should be very easy to implement and would especially help a lot for people who are not fluent with Markdown. This is also a good idea for writing forum posts, at least on large screens.
make the database layer independent on databases. First I'd like to add plain .md file support.
Move more logic into the front end. (Yes, ajax and all those crazy javascript stuff)
Sounds good, it should just still stay fully functional with JS disabled. But there are definitely a lot of places where the UI could greatly profit from some JS and AJAX.
What do you think? I'd like that it stays within the line of vibelog, so I don't want to do a fork that would go too far away.
Looks like all of your ideas are very much in line with mine, so except for the DB stuff, which I'd like to tackle in conjunction with the userman modifications, everything of the mentioned things should be good to go.
If you pre-approve what should be done first, I'll start doing it.
I think it has the same potential as ghost/nodejs. Or even wordpress.
Still a lot of work ahead, but it would definitely be nice to have something for D in that area. Especially when we can devote some time to do some serious performance tuning, it could have a good buzz factor.