[mrec]

I use and generally like Steam, but I'd never call its discovery "good". Users are drowning in the number of titles, the recommendation engine is pretty poor, and there's no support for filtering.

I'll never buy an RTS or a multiplayer-only game, stop showing them to me. I'll never buy an early-access game or a game with 3rd-party DRM, stop showing them to me. I could cull 90% of the stuff in my discovery queues, automatically, with 30 seconds of box-checking.

I'm also surprised that the desktop client has never implemented an "is this machine actually capable of running this game?" check. Non-trivial to implement, sure, but really useful.

[the_af]

Same here. I like and use Steam, but its UI for casually browsing games (aka "casual shopping") isn't very good. For example, when you are browsing a list of games, if you click on one to see additional details and then try going back, you lose your place in the list. This is infuriating for long lists. It kills the browse & shop experience.

The web interface for Steam is better than the app's. Then again, this is what web browsers are there for...

[simcop2387]

Yea that's the exact reason that I end up using the web interface for browsing games. The fact I can browse and open up 30 tabs of games and then look at them one by one without losing my place anywhere is so much nicer.

[robwilliams]

Middle clicking works in the Steam desktop app, at least on Windows. It opens the clicked link in a new window in the foreground.

[mrec]

Yes! Forgot to mention that, but it drives me nuts too.

[Orva]

Funny enough, when looking at Steam processes tree (and if I recall correctly, also licenses), client is done with Chromium Embedded Framework.

[toyg]

Which is probably why it still looks like shit on Retina screens. Oh, how I love HTML apps...

[greggman]

What does that have to do with HTML5 apps? In the browser, when HD-DPI support was added every page started using HD-DPI font rendering, HD-DPI CSS, HD-DPI SVG, only image were still lo-res

Conversely I still have native apps that are low-res only because updating them to HD-DPI meant shipping a new app.

[toyg]

> when HD-DPI support was added

Indeed. Except that, to get it, you-developer actually had to ship an updated browser with your "app" -- basically the same as any native app, except that instead of being "just turn on this setting in a plist and recompile", it means refreshing your little in-house fork of a massive browser project that barely anyone understands. Clearly Valve cannot be arsed to do that, so Steam keeps looking like shit more than 3 years after these screens appeared.

The universal toolkit is not so universal if every app ships its own custom version.

[janfoeh]

This is my pet peeve with Steam, too, especially because it seems to be so easy to fix. If only I had the ability to communicate my dislikes to Steam by blacklisting certain tags, the recommendation system would instantly become an order of magnitude more helpful.

Most tags actually seem to be spot-on and quite helpful, so the data is already there.

> I'm also surprised that the desktop client has never implemented an "is this machine actually capable of running this game?" check. Non-trivial to implement, sure, but really useful.

I've wondered that as well. A bit of data munching together with user feedback, in a similar manner to online fashion stores ("I found these size 11 boots a bit on the small side") would be quite helpful.

[simoncion]

> I'm also surprised that the desktop client has never implemented an "is this machine actually capable of running this game?" check.

The problem with that is that dev's stated minimum system requirements are -frequently enough to cause trouble with such a check- either substantially too high, or quite a bit too low. After all, not every studio has enough cash to discover what is actually the lowest-powered machine that one can play the game on.

[ingenieros]

Are you currently following any curators? This usually helps me find games that I wouldn't have found out about otherwise.

[chc]

Some people, when they have a problem finding good games, think "I know, I'll find a curator." Now they have two problems.

[coldtea]

Besides the chance of a word-play on the famous quote, is there any substance behind this at all?

What would be the problem with finding a curator? It's not unlike how music industry has always worked -- from the era of finding a radio producer you like to the era of following a playlist on Spotify on some genre you like.

[chc]

I have the same trouble finding good curators as I do finding good games. I don't know what curators are good just like I don't know what games are good. I'll look at someone and I'll think, "Is this person really any better than filtering by genre and then randomly buying games?" And I don't feel confident the answer is yes a lot of the time. I feel like I need a meta-curator.

I agree that this is not a particularly novel problem, though.

[coldtea]

>I have the same trouble finding good curators as I do finding good games. I don't know what curators are good just like I don't know what games are good.

You should have even more trouble finding good curators than finding good games.

But the idea is that you need to find a curator ONCE, and then they suggest lots of games for years.

A curator not being 100% perfect it's OK too -- after all even one's self buying games they read about or even try their demos, will pick some dudes. It's unreasonable to expect a curator (or a recommendation) service to not do that too.

[janfoeh]

I think there is, yes. For me, finding a person whose reviews I can trust takes a lot of time, so it's useless for answering the question "anything good here I would like to play right now?"

I have been reading movie reviews for close to 20 years now; in all that time I have found just two people whose reviews I respect and whose taste mirrors mine so much that I can follow them blindly. With games it's just as complicated, if not more so.

[coldtea]

>For me, finding a person whose reviews I can trust takes a lot of time, so it's useless for answering the question "anything good here I would like to play right now?"

Of course. You first need to do some work (find a curator you trust), before you can check a game they suggest.

But what would the alternative be? Either you ALWAYS search for yourself, or you search first for some curators and then follow their suggestions (occassionaly checking out stuff on your own too).

[janfoeh]

I can say with a very high probability whether a game is worth a look, just by looking at its tags. Of course that doesn't tell me whether I will ultimately like it, but then again reviews cannot do that either.

The ability to simply blacklist one or two handful tags would make discovery dramatically easier for me. That's all it would take.

[camperman]

Thank you for the coffee out of the nose :)

[vacri]

> I could cull 90% of the stuff in my discovery queues

The whole point of the discovery queue is to expose you to a variety of options. If you want to search just those things you want, there are already filters you can manually use. These filters don't have all the options (please, no 'early access', I agree), but they give you the filtered lists that it sounds like you want.

[overgard]

A machine requirements check would be useful, but I can't think of many companies that would implement something that results in less sales. What's the upside? Maybe some good will? I'm not saying valve wants people to buy games they can't run or anything like that, but it would be an expensive ongoing effort to make less money.

[greggman]

How would that generate less sales? Isn't the whole point that since the recommendations are irrelevant they're getting less sales not more?

[toyg]

>"is this machine actually capable of running this game?" check. Non-trivial to implement, sure,

Quite trivial, actually. They already do the necessary data-gathering at your end (i.e. their periodic automated survey), the only "hard" bit is data-gathering at the other end (when a game is listed).

It's so trivial to implement, in fact, that one has to wonder whether there are commercial considerations overriding the technical aspect. Maybe they don't want to piss off hardware and game vendors, maybe Valve would rather have you buy their console to bypass these considerations altogether... who knows.

[logfromblammo]

This immediately calls to mind Grim Fandango Remastered which "requires" OpenGL 3.3 .

It doesn't, really. It has the ability to run the game at its original (1998) level of quality. But the developers chose not to enable this as a fallback. You have to rely on some guy on the Internet that figured out how to hack it so that it would work.

It'd be great if I could, for instance, filter out everything that won't run with the Intel HD3000 graphics unit on my cheapo laptop.

[djsumdog]

Not to mention it crashes all the damn time. Your store shouldn't crash when you full-screen a trailer!

I always buy games on Hubble if they're available on both. Sure my save games don't get backed up, but they are drm free. I'll even pay more if the humble price is higher for drm free.