With Priority Inbox, every message in Gmail has two buttons next to it. “Important” and “Not important”. By default, most messages seem to be unimportant. However, as you flag important emails, Gmail will look for patterns and start automatically determining which future messages are important. Important unread emails appear in a special pane at the top of your inbox.

It’s a really simple idea, the sort of thing that makes you wonder why nobody tried it before; we’ve had similar technology in spam filters for eons! Email power users have been sorting their email for important messages via filters for the past however-many years, but Google’s solution is much better from a usability standpoint. How many average users even go to set up a filter in Gmail? I’d be surprised if it were more than maybe one in ten or twenty. Now, the filter comes to them. There needs to be more software that adapts to the user, rather than vice versa.

From the introduction:

…According to the Nobel Prize Committee, the prize was given for Krugman’s work explaining the patterns of international trade and the geographic concentration of wealth, by examining the impact of economies of scale and of consumer preferences for diverse goods and services. Krugman is known in academia for his work on international economics (including trade theory, economic geography, and international finance), liquidity traps and currency crises. According to IDEAS/REPEC (a ranking of Economists by article citations), his work has made him one of the most influential economists in the world, and he is among the 12 most widely cited economists.

Oh, hey. He sounds pretty smart. I wonder what his economic views are?

Economic views [This section requires expansion.] Krugman identifies as a Keynesian and a saltwater economist, and he has criticized the freshwater school on macroeconomics. In the wake of the 2007-2009 financial crisis he has remarked that he is “gravitating towards a Keynes-Fisher-Minsky view of macroeconomics.” Post-Keynesian observers cite commonalities between Krugman’s views and those of the Post-Keynesian school.

That’s it. The shortest section of the whole article. On the subject which the man is the twelfth most quoted. It resorts entirely to labeling classes of economic theory without any description whatsoever as to why the man holds them. I am amused. It’s pretty funny how much better Wikipedia editors are about keeping up with politics than with reading the work of the people they write about.

The idea that THQ is somehow “disrespecting customers” with this kind of rhetoric misunderstands the situation as completely as it is possible to do so. In a literal way, when you purchase a book used, you are not a customer of theirs. If I am purchasing books in order to reward their creators, and to ensure that more of these ingenious codices are produced, I honestly can’t figure out how buying a used book was any better than bootlegging. From the perspective of an author, they are almost certainly synonymous.

It’s exceedingly rare that I purchase a book from a used bookstore these days. I got tired of miscategoized shelves, or being told that they didn’t have a book when they did, or going online to Amazon or eBay and finding fifty copies of the book I was trying to buy listed out like some heathen index of commerce-related rituals. There’s more, besides. At some point in the last few years, I became incredibly uncomfortable with the used books market.

…

I sold my old books for a long time, there’s probably comics somewhere in the archive about it – you can imagine how quickly my cohort and I consume these things. It was sort of like Free Money, and we should have understood from the outset that no such thing exists. You meet one person who writes books for a living, just one, and it becomes very difficult to maintain this virtuous fiction.

I’m sorry, does this seem strange to anyone else? Video games aren’t as unique as people would like to believe.

Used and new games are the exact same product only in theory. In practice, the online and multiplayer community will be much larger and more reliable at release. As the product grows older, multiplayer updates cease and the userbase wanes as players move on to other products. Eventually, sometimes within as little as four or five years, the publisher will kill the multiplayer server. I’d be hesitant to buy the multiplayer componant on a used game a few years old because chances are there’s not much of it left.

This also ignores how, frequently, a pretty large percentage of profit is made very soon after launch day, before copies are used. This is one of the lines we’ve been fed about why DRM is used even through it’s ineffective; it helps prolong the period where a new copy in the legitimate virtual and physical stores is the only version of the game available.

The mainstream game industry makes less sense to me every day. Every time I see something crazy like this with everyone making a fuss, I’m less sure that I’d like to have anything to do with it. Maybe I’ll want a job in it someday and do follow up posts rationalizing everything for the benefit of a future employer. (Hello! ) Maybe I’ll just stick with indie development on the side.

So, yes. Make a complete game in a weekend. Can’t be that hard, right? Cactus does it all the time…

I suppose I should write up a post mordum.

What went well

For my first competition, I’m happy that I even came out with a finished game.

Using a tool I’m intimately familiar with was a good choice. I hadn’t used Game Maker for a real project quite some time, but the important things have barely changed since the 4.x days.

I still like the concept of defeating an enemy, watching it flip upside down, then using it’s overturned belly as a platform. The rhinoceros rabbit was also fun, although the code behind it was a pretty big last minute hack.

What did not

Game Maker does not have an amazing workflow. I was using the “lite” version, and importing images was a pain. It especially didn’t like transparent areas; I was pasting in images from Paint.net and then having to change the background transparency for every single sprite. Additionally, I had to reenter the tileset placement details and finagle with the level editor’s tendency to corrupt the tiled background after major tileset modifications.

My engine going in was a quickie two hour job. As simple as it was, it turned out to have quite a number of bugs which surfaced dramatically when I started trying to implement gameplay. Additionally, I needed to have more than just player code written; a generic enemy object or some simple physics to inherit from would have helped a lot when I started slamming out code in a panic.

I didn’t have a vast amount of time to work on the project. Maybe fifteen hours or so. This might not have been a big deal if my engine had been more complete.

Almost all the level design and gameplay came together in the last two hours. I noticed that people like Notch and Sophie H got something playable with as few game objects as possible as soon as possible. Several other people avoided making games that required much level design at all, a choice a began to envy in my last hour of work.

Takeaway

I’m going to switch to a different platform else next competition; something other than Game Maker. Aside from the annoying workflow, I don’t like excluding people running operating systems other than Windows. I’m not comfortable enough with Unity3D and 3D modeling for it to really be a great choice. If I get really good at it in the meantime while tinkering with StarCats IV, I may consider it. LOVE2D is one possibility; the ease of LUA, all the built in classes, and the ease of packaging things up for distribution makes it very attractive. FlashPunk or Flixel is another choice. Flash may be a dying platform, but right now it runs on just about any desktop operating system and flash game developers get crazy numbers of people playing their stuff. It’s no coincidence that it’s preferable to have a web-based LD entry; the lower the barrier to entry, the more ratings you’ll get. Finally, there’s Javascript. Akihabara looks like a decent framework, and I am using web tech a lot at my current job.

I should also build up a decent engine before the next competition. But that would require planning and foresight. Ha! Maybe I’ll just get good at a development platform instead.

I mean, when you search for a programming question and it embeds the top answer from a related StackOverflow question as the first result? Or suggests pages from the official PHP documentation if it couldn’t find other stuff that was good enough? And it has all those silly !w shortcuts to search other sites that I usually end up setting manually in every browser I use? Yes.

Scroogle, take note. It’s privacy sensitive, but it’s also actually enjoyable to use. Search is more than ten blue links on a page with conspiracy theories and all that.

I have mixed feeling about this. Wave was absolutely fantastic for collaborating on projects. It was a huge hit with my campus ACM group; we used it to outline presentations, plan the setup for a new computer lab, as a makeshift IRC channel, and on a software engineering project or two.

It’s not that we couldn’t have used other tools. Wave could be slow and buggy, after all. But the idea of having this all bundled together in a single tool was just better. It made it really easy to start a project and get everyone on board easily, without having to fool around with some combination of Google Docs and email and IRC.

On the other hand, since I graduated, I’ve used Wave perhaps three times. I don’t have any projects to do, so it’s utility is limited. The general public is in a similar situation and doesn’t have much of a use for Wave either; it’s a really great really specialized tool.

You know how Twitter used to be down all the time, but now it still goes down and is really glitchy to boot, but we love it anyway? That’s pretty much how I feel about Wave.

On the plus side, it’s not like all the functionality will be completely gone. There’s always Analogue and Google Docs. But I’ll miss the integration and the ease of tossing together a group of people together in a virtual room. Without email.

function sendEmail($email, $contents) {
    /* AN EMAIL IS SENT
    BRUSHES MAIL SERVER LIKE LEAF
    SOON USER RECEIVES */

On my Intel X3100 GMA, it’s still unplayable. I hit peaks of 20 FPS. The fog and particle systems really seem to bog things down, dropping the FPS to 3 areas like on the first level. Unfortunately, you can’t disable them; Valve has them marked as “cheats.” Never mind completely disabling the lighting system would be much more of a cheat in this game, and you can already do that.

Launch options: -dxlevel 90 -w 480 -h320

Paste these options into a file. Save it as autoexec.cfg in C:\Program Files\Steam\steamapps\common\alien swarm\swarm\cfg

//Alien Swarm
//Integrated graphics card config.
//http://jamesgecko.com/alienswarm

// Vastly improves performance of Alien Swarm on integrated graphics cards
// Unfortunately, that's not saying much. Expect 20 FPS, tops. Maybe 5 FPS in areas
// with lots of fog.

//Misc
cl_showfps 2 //0 off, 1 fps, 2 smoothed fps.
asw_spinning_stim_cam 0 //disables "picture in picture"
ai_expression_optimization 1 //People don't make faces when you aren't looking at them.
mat_disable_ps_patch 1
mat_vignette_enable 1
r_dopixelvisibility 0
r_drawbatchdecals 0
r_drawdetailprops 0
r_fastzreject -1


//Filtering
mat_disable_bloom 1 //Disables bloom
mat_hdr_level 0 //Disables HDR
mat_grain_enable 0 //Disable film grain
mat_grain_scale_override 1 //Disables film grain. Again.
mat_aaquality 0
mat_antialias 0
mat_bloom 0
mat_bloomscale 0
mat_disable_bloom 1
mat_disable_fancy_blending 1
mat_forceaniso 0

//Models
r_modellodscale 0.10 //reduce quality of models
r_lod 8 //Level of detail. -8 is best, 8 is worst.
mat_max_worldmesh_vertices 0

//Lighting
r_maxdlights 0 //maximum dynamic lights
r_dynamic 0 //No dynamic lighting
r_shadows 0 //No dynamic shadows
r_lightinterp 0
mat_filterlightmaps 0
mat_softwarelighting 1 //1 uses CPU over gfx card.
mat_disable_lightwarp 1
r_shadowrendertotexture 0

//Textures
mat_compressedtextures 1 //Integrated cards are too slow for uncompressed textures
mat_specular 0 //Reduce shiny
mat_showlowresimage 0 //very low quality textures.
mat_mipmaptextures 0 //makes textures look even worse. Character selection screen unusable.
mat_fastnobump 1
mat_picmip 4 //texture detail setting
mat_parallaxmap 0


//Water
r_water_drawreflection 0 //Disables some water effects
r_waterdrawrefraction 0 //Ditto

//datacachesize 128

//Disabling fog/particles
fod_enable 0 //cheat. This would really help with performance
r_drawflecks 0
r_drawrain 0
r_drawparticles 0 //cheat This would really help with performance
r_particle_timescale 0 //Particles aren't animated
r_particle_sim_spike_threshold_ms 0

r_decals 5 //number of visible blood, bullet holes, etc.
r_3dsky 0 //We never see the sky

//Ragdolls
g_ragdoll_maxcount 0
ragdoll_sleepaftertime 0
cl_disable_ragdolls 1 //cheat
cl_ragdoll_collide 0

//Glibs from boxes and such
props_break_max_pieces 0
prop_active_gib_max_fade_time 0
prop_active_gib_limit 0

//Multicore
r_threaded_particles 1 //I've got a core 2 duo. Disable this stuff on a netbook.
r_threaded_renderables 1 
snd_mix_async 1
mat_queue_mode 2

Select element ids from the HTML page in jQuery by using $(‘#theIdIsHere’).functionName(… You need to hash mark in front of the name for ids. If you are selecting a class, you don’t.

This is how buttons work. If jquery click events won’t trigger, remember to double check that you’re using the document ready function.

<html><head>
<script type="text/javascript" src="jquery.js"></script> 
<script type="text/javascript">
$(document).ready(function() {
    $('#my-button').click(function() {
            alert("Yay!");
    });
});
</script></head><body>

<input type="button" id="my-button" value="Click me!" />
</body></html>

This is how you recieve JSON from the server (server -> CLIENT)

$("#hist_button").click(function() {
    $.getJSON(loadUrl,
    function(data) {
    history_id = data.foo;
    alert(history_id);
    });
});`

This is how you send JSON from the server (SERVER -> client)

<?php
    $data = array('history_id' => $x, 'html' => "foo");
    echo json_encode($data);
?>

This is how you send JSON to the server and get the response (server <- CLIENT) All value strings must be in double quotes. Don’t enclose keys in quotes.

$("#hist_button").click(function() {
    $.getJSON(loadUrl, {key: "value"},
    function(data) {
        history_id = data.foo;
        alert(history_id);
    });
});

This is how you recieve JSON on the server (SERVER <- client)

<?php
    $historyId = $_POST['history_id'];
    $x = $x + 1;
    $data = array('next_history_id' => $x, 'html' => "foo");
    echo json_encode($data);
?>

Firebug (preferably the Firefox version; it has the most features) is the best thing ever for debugging this stuff. Right click on the extension icon in status bar and enable all panels. Now you can watch the posts and responses in the console panel.