The bigger things are the no-server and personal projects--and in my case, my personal project is (at least conceptually) based on the no-server project (which was, as it sounds, just client-side Angular app.)
This was my no-server project. It's a markdown editor with realtime preview,
based on Codemirror, and very unoriginal. The most
interesting thing about it is its vi-like keybindings (which weren't even
working at the time of no-server projects). The editor is available on
Github and installable straight from
npm i markvi && cd
node_modules/markvi && npm start). I kind of like it, how it is, just storing
your docs in localstorage, but having decided to scratch the
Tinder-for-the-unemployed idea (only partially because it's already been
done), I figured since I've got no other ideas and
don't like the idea of just building an Amazon clone (if someone really needs
that before making a hire, it wouldn't take an absurd amount of time, anyway),
I'd go further with Markvi, and maybe actually finish the damn thing.
As it stands, it's not broken, and now it saves to my (live and hosted!)
database document store. I still can't get behind calling Mongo a
database, since it lacks all of the things that make databases useful (and, to
be fair, all of the things that make them horrible). Anyhow, it meets the
requirements for the personal project (MEAN stack, two collections, one ref,
preferably not totally broken), but my cohort seems to be full of very
ambitious people. I knew that we, as a group, would be significantly more
impressive than the previous cohort, but I didn't quite expect to get to a
point where I'd feel... slow.
So, this is where I'm headed. A native(ish) desktop app, built on either Electron (formerly Atom-Shell) or NW.js (formerly Node-Webkit). That end of things currently is halfway okay--posts post, the database stores them, but for whatever reason (the reason is Angular. I blame Angular. Or at least UI-Router, which is still a gigantic fuckwad of uselessness when it comes to stability or debugging.) some things just don't render, or render very poorly. It's progress, though! There's a searchable index of all saved posts for a user (who, right now, is only authenticated the old-school way, though it should be as quick as uncommenting a few lines to get Twitter, Github, Google, and Facebook auth working), though the live filtering of tags isn't working as planned, so I've temporarily re-removed that from the app. I'm leaning towards Electron, not exactly out of preference, because I'm neutral to both of them, but just because its popularity (and the popularity of things based on it, like certain editors on the market) means it's a lot easier to find good examples, generators, packagers, and projects I can tear apart and study. I'm also on the fence about which editing component to use... the stable/production version of Markvi is using Codemirror, but the more I play with them, the more I really enjoy the Ace editor. It's heavier, but it feels more solid and dependable (and I'm biased, since I really like my C9 editor/dev box). Seems like a better API and larger community support, too. Codemirror's great, it really is, and obviously the guy who wrote it is absolutely brilliant, but Ace just feels more natural, or something. I don't actually have a great reason for that.
That whole project is also available on Github, under the temporary name of "As Yet Unnamed Blog App". As always, anything that's mine is totally under whatever license you feel like at the moment (the WTFPL and DBAD are personal favourites, and they boil down to the same thing, legally: do whatever the fuck you want, but don't be a dick about it), and if anyone happens to see this and spot some problems in my code (or solutions to getting things to look right in Electron?), I'm definitely open to PRs.
Oh, and yeah, I have heard of Ghost. Hush. Ghost doesn't come in a fully-offline cross-platform binary. Neither does mine, yet, but after the project presentation (Monday! Eeeeeek...), I'd like to switch to a filesystem-based store, instead of Mongo (which takes a minimum of 300 megs for its journal file, unless otherwise specified at the command line), or maybe a MiniMongo or MiniRedis type of thing (which would be a good excuse to actually use Meteor, which I like almost as much as I love Vue)... so if I could cut it down to less than a hundred megs or so total, once wrapped, it could be something... different. I might even use it, who knows. Whether Ace or Codemirror-based, it'd have a lot of room for expansion to other languages (another area where I think Ace would do better than Codemirror, because of Cloud9's support and plugin community).
OKAY. That's all, that's my update. Again, here's my
editor as it currently stands, totally
useable, just click through or
npm i markvi && cd node_modules/markvi && npm
start and... basically, wish me luck. (As usual, this is said knowing exactly
how many people read this blog--I don't even read the thing, so that's exactly
zero). It's really more of a journal so I have some idea of what the hell I
was doing when I messed up X or how I managed to do Y, than anything else.