In 2010? I reworked my Comet-based realtime game-in-a-web-server library, RailsGame.
I fully rewrote, updated and released a webapp called WantMyJob.com.
I wrote a circuit simulator game called Shanna's Pizza to start teaching my two-year-old the basics of programming. I also started doing Snap Circuits with her so she'd have a physical basis to understand the circuit simulator game. Also, had a second kid.
I wrote a Rails 3 tutorial, including an app with logins and nested resources, called RefactorIt!
I finally put together a real programming portfolio, on the web and everything (angelbob.com/portfolio).
I read up on Bayesian Statistics and Data Analysis (thanks, O'Reilly!).
I blogged and blogged and blogged.
I reworked, resurrected and took screen shots of a scary amount of my old work, partly for the portfolio. This sort of blurs with converting my old Rails stuff to Rails 3, but I did some of both.
I continued to work on my day job, of course. And gave a JRuby presentation for Engine Yard based on what we do there.
Did you know you get more done if you write it down somewhere (like a blog and portfolio) and tell people about it? True story.
Did you know you get more done if you write it down somewhere (like a blog and portfolio) and tell people about it? True story.
Do you find that? I'm the opposite, once I "announce" something all desire and motivation for completion goes out of the window. For me I think it's about the appreciation, so when I explain the idea and love it, why complete it.
It varies for me. Sometimes it's just an idea and I want to share it. Maybe I'll never do it, maybe I want to get feedback on it. Some pretty cool stuff came out of just "What-If" like Mongrel2 and Librelist.
Other times I get infected with an idea and have to crank out a prototype to see if it'd work. When I'm done, I still try to throw it up there and see what people think. Usually the process of writing about it actually helps me work out if it's worth it or not as well.
Indeed writing and interacting about something levels you up. I'm not (yet) a blogger but recently I have started interacting on a few social coding networks (stackoverflow, github, HN) and I feel more productive.
PS. Is RailsGame open source ? It looks interesting!
I wrote a rant about how much I hate people that kick back and put their feet up on the table.
Oh, nope, that was Zed, too!
(But seriously, Mongrel 2 is an amazing work of software engineering. In addition to being a good web server, it's also a good example of how nice C is if you forget how they did it in the 70s and write it like it's 2010.)
Wrote my own web server. Which runs my own web framework. Which runs my new webapp a friend and I wrote yesterday. Writing a simple web stack isn't hard. Writing a full featured, secure, and reliable stack is more difficult.
I mean we are mostly all people who make stuff here. I am slightly offended by your attitude that ZS is some how showing us up and we need to prove what we did this year.
Launched wildlifenearyou.com, built a new commenting system for guardian.co.uk (with a great team), ran a marathon, got married, went on a six month honeymoon through France, Spain, Morocco and Egypt, co-founded and launched lanyrd.com with my wife (from an apartment in Casablanca), learnt to scuba dive, and we're now on a self-drive safari in South Africa. 2010 has been /awesome/.
Hey, I understand the guardian comment system uses Pluck and we're looking to integrate pluck ourselves at the ABC in Australia. I don't suppose I could pick your brains for any kind of implementation tricks and information you can provide can I?
You say "got married" = hosted a fantastic wedding, complete with huge conga lines, traditional Jewish chair dance, a wedding "cake" of cheese, otters, owls, eagles, and more!
Since we're all bragging, I finished my MSCS at Stanford with over a 4.0 GPA, wrote a game that was a finalist in the CS248 competition, had my first paper published, TAed CS210, launched my startup, learned Django and CoffeeScript and JavaScript, wrote a charting library in CS/Raphael, started learning how to be someone's boss, and moved across the country and forced myself to make friends. All in all I feel good about the year.
Over a 4.0 GPA? Man, are grades that badly inflated now that they have to add extra points to the GPA to make people seem better? No knock on you at all, but just curious what the scale is.
At my uni (University of Montreal), A was worth 4.0 and A+ was worth 4.3. I don't think it comes from grade inflation, though. Just a different convention.
Well, an A+ is worth 4.3, so if you're some kind of superhuman I suppose you could get a 4.3 overall, though it's not really practical when you're taking 3 classes a quarter and being a TA or RA. I managed to get three of those, and honestly I think I earned them - I worked extremely hard throughout my MS and that occasionally resulted in very strong results. There are some classes with inflation-compatible grading schemes (Pat Hanrahan's 148 comes to mind, in which, when I took it, 7/8 projects offered 25% extra credit each, and each project was 10% of one's grade, meaning the maximum grade for the course was 117.5%...though I didn't get an A+ in that course, since I didn't do well on the exams), but they're the minority.
The problem is, without knowing the cap on "bonus points," I can't know how far under 4.0 your GPA might have been without the bonus. Maybe your GPA was "only" a 3.7 before adding 1 bonus point? It's ambiguous at best, deceptive at worst.
On the other hand, if a school doesn't give extra points for more advanced classes, someone who took the hardest curriculum possible and did pretty well might have a lower GPA than someone who took the easiest curriculum possible.
* Kicked out more bad guys out of compromised networks than usual.
* Grew my main business, had our first year where I've not had a single worry about running out of money (touch wood) in the middle of a serious recession.
* Started building academic and vendor relationships (never thought I'd do that) to develop new technologies and services.
* Did more vulnerability research this year than I have in the last two but haven't published anything public yet.
* Moved to Sales, Marketing, now back to Technical for the time being.
* Took on a business partner and mentor
* Bid for two research grants
* Built up pipeline, wrote more proposals than I have in a long time.
* Seriously damaged my knee in a freak dancing accident.
* Grew a Mo to support Prostate Cancer charities (and raised a few bob too)
* Started a proper non-security side project (http://www.minklinks.com/) - failed to get it into beta for HN November but it will be out before year end.
* Did a whole load of pentesting.
That's it so far, still more things to come before year end though. Hope I pull them off.
* I'm still helping out Python newcomers on Pylons, Tornado, and recently Flask mailing list.
* I fix bugs on a couple of gems.
* I start blogging again.
* I start learning http://keplerproject.github.com/orbit not too long ago after answering my own question "I wish there's a hybrid between ruby+python and faster than both."
Well, I made a new cloud computing platform using my new federated REST database and my new message queue and my new Subversion sharding layer and my new unit testing library, all in a language I didn't know when I started. Then I made a web site and management console for it that runs on it. And a network API and a command-line tool because you need those.
I love the natural style of page conversation on the example page. As a Java developer, I wish that this style were possible on the servlet platform:
local login_page = Tir.view("login.html")
local login_form = Tir.form { 'login_name', 'password'}
local function login(web, req)
local params = login_form:parse(req)
repeat
params = web:prompt(login_page {form=params, logged_in=false})
until login_form:valid(params)
return web:redirect('/')
end
Tir.start {route='/Login', main=login}
Also, +1 to this: "the framework creator shouldn't be shoving stateful/stateless dogma in your face".
The natural style is sweet for things like that. I did the Hype.la signup process with natural style and it meant I didn't have to touch the database the whole time. If they quit half-way through then no DB junk left over.
Where this kind of coroutine stuff sucks though is in anything that's "RESTy". In other parts of Hype.la I had simple Xhr callbacks that just grabbed lists of data, so the coroutines were just annoyingly in the way and half the time they were just a big loop with an if-statement for the URLs.
That's where the evented came in. In that way I just map actions to callbacks and done.
Then I ran into a few simple things that didn't work in evented and were still stateless, like the search interface, and so I use the stateless handler for that. Nice and snappy with no extra junk.
Another advantage of all this is that, if you have a part of your app that needs to be resistant to restarts, then you can craft it evented or stateless. In other parts, restarts aren't a big deal so you can do natural. And since everything is just a bunch of little processes, it scales out nice and doesn't have as much of the "gotta-copy-the-world-for-users" that Seaside upgrades do.
This one really surprised me: "Finally, there's not much for deployment. I run my apps in GNU screen and that's about it." Seriously? Is he that lucky, serves that little traffic, or does he run with redundant power, network, server and 24/7 monitoring in an inaccessible bunker?
I haven't been able to run even my toy web apps without something like monit - due to both things under my control and outside of it.
Hype.la is the only real Tir app, with the Mongrel2 BBS demo being a "proto-Tir" app. There's hardly anyone on it yet since I didn't really "launch" it. Because of that, I just run it in screen and I'll worry about the deployment scenario later. Also, the Tir start command actually works kind of like a little monit/daemontools, so if a handler dies it gets restarted. It's pretty robust already.
And yeah I need some monit soon, but so far it's been running fine for the small scale.
On my hobby projects I run stuff in screen, I haven't had a problem and my server uptime is currently above 200 days. Is it being used? Something like 1,000 concurrent users for most of the day, so yes.
I regarded this as cowboy, but as it worked for me I haven't needed to re-visit it.
It is cowboy, but there is a time and place for that. If you did this in a large organization, you'd be putting the team at risk. Since it is just you, well, do what you find prudent.
Last time this was posted I was really interested in it. So interested in fact, I built a django-style template library to go with it because I really liked 'extends' and 'block'.
Interesting, I'll take a look at this some more. My only question is, isn't plain old Lua already like django's templates. What specifically are you adding that embedded lua doesn't just give you?
`Template inheritance¶
The most powerful -- and thus the most complex -- part of
Django's template engine is template inheritance. Template
inheritance allows you to build a base "skeleton" template
that contains all the common elements of your site and
defines blocks that child templates can override.`
Having made it in lua myself, I can tell you it is indeed quite complex. ;)
The lua template only has "include" (i.e. copying another template into this template), but not "extends", (i.e. use another template as a skeleton and this template only provides details on how to fill it).
Well, I only did what I needed, but since Lua's included code blocks are also functional, wouldn't it be easier to just allow passing arguments? It'd also make a lot more sense. I find Django/Jinja inheritance concepts really confusing.
I'll take a look at the code and see what I can come up with.
If you wanted, say, your about page to have an extra image next to the logo (as an exception to all the other pages), then you'd have to rewrite another header file to include, whereas if you used template inheritance you could just override that logo part without having another almost identical header file.
It may be confusing if you're new to it... but so are OOP, Lambda calculus, and others. ;)
EDIT: The code isn't that clean... Apologies in advance.
Not with the same notion of inheritance that django templates have (with multiple blocks that can be overridden in child classes). There is one project someone attempted but never finished.
Ha, yeah I had to put that in. I find working in the valley is kind of like working with a bunch of guys before Knuth and Dijkstra established structured programming. It's like they say, "Yo, I don't do if-statements man. I use gotos only. That logic and structure crap just slows me down and busts up my creative chi."
As for hype.la, stay tuned. I gotta fix up the ad blocks and improve the UI a bunch, but I'm really aiming for it to be for FLOSS and other artists to do ads without feeling like sleeze. Definitely something I want.
Very cool. I was just looking for other Seaside-style web frameworks today. The "natural style" mode in Tir is just like the main way to build web apps in Seaside.
Aside from Lua and Smalltalk, what other languages could host this kind of framework? I know Ruby 1.9 has Wee, but that doesn't seem to be "ready for prime time."
In its early stages it was continuation-based, but now it's just regular ol' closures. You can still use continuations rather easily though; some of it is modeled on the PLT stack (send/suspend, etc.).
Code written in the continuation monad (i.e., simple imperative style) can be translated to an Automaton arrow and fit into a Haskell on a Horse application. This approach has the advantage of type-safety and greater composability.
The main feature that is required is the ability to callcc and/or coroutine yield from any point in the stack. If you look at Wee, the user of the framework has to callcc all over the place to make it work. In Tir I have web:prompt(), which itself does the yielding without anyone having to care.
The significance of this is that Tir can provide one API for all three styles without special effort from the user.
What did you do this year?