Hot damn! I'd never have expected him to find his way there, but it seems like it would be a good fit. A smart ambitious guy who hates bullshit, and one of the few companies that it's impossible to accuse of bullshitting. I wish them both luck.
Sure, you might prefer jamming with other musicians. But Fret War really is for guitarists who're keen on developing technical proficiency,, composition and improvisation skills. To be technically proficient, you always have to strive to be better than others, not necessarily "beat them".
The information collected is actually pretty interesting. I assume voting will have something to do with sounding good in addition to speed and accuracy. Besides making a cool site, he just crowd-sourced music production (everything is released with a CC license).
I didn't find it hard at all, just not very interesting, so wasn't motivated to learn it (no offense). I think having more open-ended jam tracks would work better.
Alright, prove it. Send in your submission. If it was so easy you should be able to send it in and compete, regardless of whether it's interesting or not.
Nope, not rails (troll). It's a web.py application combined with a Lamson application. The frontend is run through nginx and I use some heavy caching tricks to make it fast as possible.
The primary design for hosting costs (scalability) are:
2) Javascript determines the login state of a user and then displays user state links.
3) Generation of the site is "out of order" for convenience of generation/caching, and then javascript knits it back together. We'll explore more of this later.
4) Caching for about 5 minutes on most pages, depending on usage.
6) Gzip compression (should do precompression really) and other tuneups.
7) Using Lamson to accept posts via email so that they're processed off-line and statically generated. This is also really easy for people to use and easy to setup, so triple win.
It's currently not close to a production setup, but I'm going for making it easy to host on a small server to keep costs low. Part of that is just not using the framework when possible.
For statistics, right now the Nginx server is getting a modest 4-10 req/sec for html, but the backend is only serving about 1 request a minute thanks to all the above. Also the site really doesn't accept uploaded content via HTTP except say for photos.
Wow, someone else using web.py, and that too for a "real" project! I always feel like a n00b for using such a stripped-down "framework" but after spending some time trying Pylons and Django, I found them both to be too "heavy-weight" for me.
I like web.py for this because it's tiny, fits HTTP real well (try doing REST with Django), and it's "just a library" so it's easy to mesh with Lamson. Also, I was bored.
It's not asshole-ish at all, it was just direct and made the point much clearer with the proper amount of urgency. This bug crashes sites consistently and is easily exploited. That's serious enough to warrant an immediate fix and release with an apology for the design. That's what I've done on my projects, and that's what I expect.
Now, the question is why do you demand humility from me, but not from them?
Seriously? A bug report that identifies a DOS condition gets flagged as "won't fix" -- and Zed's followup explains, with details, why it should be fixed, and what the fix should be.
I tried, and I couldn't find anything asshole-ish about the bug report. It looks like good work to me.
We actually keep the sessions very light, and if I could I'd get rid of them entirely.
For caching I'm just using nginx and generating as much content to disk as possible. It is much simpler that way, and the goal is to keep things simple and cheap so I can bootstrap the site.
Chances are he used django for it. He has blog posts on his new found love for django and his ratings code is in python. According to the Lamson (his project) FAQ, Lamson was influenced by django's features and structure.
