Dev'ing for Source is tedious

[mrfranswa]

I was a bit hesitate to post with the "apple fags" comment, but your comments on hammer are valid.

As someone who's dev'ed on different engines, source is not exactly the "worst", more like "the oldest" at this point. There are generally two different types of developers: One who doesn't release dev tools, and the other who only releases dev tools, (or spends most of their time doing so). Only unreal and crytek really do both, and both focus more on the licensing aspect since it makes them more money. Source really hasn't changed at all other than add new features. You still develop the games exactly the same way as you did for hl1.

So where does this leave source? I was getting to that! If you notice, there are little/NO proprietary games being released anymore which use the source engine. Many of the most popular mods developed using valve technologies are jumping ship, (age of chivalry, natural selection, etc), and schools are starting to phase out their source classes in favor of unity and unreal. This is due to three main reasons: Unresponsive nature of Valve and the way they do business, Valve's stance on using third party developers to make many of their tools, and the outdated pricing structure of licensing their engine.

Unresponsive nature:

Many game companies are hard to get in contact with. This is a standard of the industry, however Valve has the same stance with their partners. This is mostly due to their flat company structure. Universities who sign contract with valve to teach their engines have absolutely no way of contacting anyone at valve to answer questions. And when we do, we have to use the same channels any person can, (such as sending an email to Mike Durand and "hoping" for a response).

Stance on third party tools:

As all of the loopers know, valve's official way of doing outside hammer-specific things, (such as making texture files, embedding sounds into your maps, etc), is pretty terrible and horribly complicated. This is why we all use third party tools such as nem's awesome vtfedit. Universities interested in teaching classes on source hate relying off of third party publishers so much. There isn't even an official source for information about valve's dev'ing. That's why loopers is so popular. The valve wiki is mostly maintained by community members, and many aspects of it are incomplete. Unity is a perfect counterpoint to this, because they have a huge resource of help out there which is official.

Licensing! Licensing! LICENSING!!!!!

The third, and by far most important argument, is the cost to license the engine. Unity is FREE until 25% of your profits over... I think 5k (could be wrong don't quote me), unless your company makes over 100kyr. In which case you have to buy the 1200$ unity pro. The iphone add on cost 300$ + a 99$ yearly apple subscription to release your content through the apple store. Unreal is 99$ for a royalty-bearing license, (which again I believe is 25%), or 2500$ per seat, and cryengine 3 is experimenting with a similar royalty license structure. Source on the other hand cost 300,000$ to license. What's happening now, is everyone is starting to develop titles with these cheaper engines because they are new, streamlined, and practically free to develop with. With this, universities are starting to jump on board and ditch the more expensive engines. When applying for jobs, I can honestly say about 3/4th's of the jobs asked my experience with UDK or proprietary engines utilizing UDK technology, while two asked my experience with source.

Don't get me wrong, valve tech has been my bread and butter for the past 10 years, and I absolutely love the engine. It's highly optimized and it's incredibly easy to mod. I love hammer's I/O system because it's just so darn easy. However, after sitting down with crytek and testing out cryengine 3's level editor, there is really no appeal to developing a title using source anymore. One of your arguments about source is it's level compiling process. Do you know unity and cryengine 3 have NO COMPILING! You hit play, and you play the game instantly! Do you want to develop your title for PS3, XBOX 360 and PC simultaneously? All you need to do is plug them into your computer with their own displays, and you can test all 3 at the exact same time! Heck, unity doesn't even need it's own texture editor. You just take a jpg and drag and drop it into unity!

Sorry for the very long post, however I thought it was pertinent, and this is what I've come away with at GDC back in May. To be frank, I tried asking people at Valve about this topic, and they were unable to answer my statement. I'm just hoping valve comes up with some streamlined technology soon before they lose their competitive edge. Well, at least they always have steam...

[skoften]

Thanks for your contribution. My initial post was a bit too sarcastic. Your post is very informative.

[Enraged]

I fucking love the Source engine, and consider it THE best engine currently available for modding and serious commercial development projects, and hope that it is not abandoned in the next 10 years. As long as we get the occasional updates like we do now, I do not see it needing a replacement anytime soon. I regularly hear people complain about it's limited graphical capibilities, but I think the engine looks marvellous and it's capibilities have no need of being called into question. Nearly everybody recognises there is a very distinct look for the engine that sets it apart from Unreal and Crysis, and I put it down to the following; The Source engine make a great deal of the small detail. In Unreal, each map usually has so much crap in it you can't really concentrate and appreciate one thing. Which brings me to my second point.

As someone who uses the engine, I feel more in touch with the map I am actually making. In Unreal and Crysis, I couldn't give a crap about a lot of the stuff I am doing, but in Source I care about every detail. This could be put down to the amount of time that it takes to do a single thing, as it does take longer than doing the same thing in any other editor, but even when I've completed collosus scale components in these other engines, I don't really care about them. Source makes me care about every millimetre of my creation, and no other editor has made me feel that way.

Thirdly, I agree that the workflow is exceedingly poor, but after having used the engine for what is coming up to 6 years I can say that what took me half an hour in 2004 can take 2 minutes in 2010. I have simply gotten used to opening notepad, and I think I now know every texture code that the Source engine has.

I love this engine, and am hoping that it does not go anywhere in the forseeable future. That decision isn't up to me, you or anyone on this forum though, so if we must see a replacement, I hope it can rekindle the same joy I get from working with Source.

[Major Banter]

I think with Portal 2 there'll be some rather snazzy new updates to the engine that we can all greatly benefit from.

And also hopefully put Source back into a marketable realm.

[joe_rogers_11155]

So I recently met a guy who used to work in the industry. He happened to have a shit-ton of UDK stuff on him, including over 20 hrs of video tutorials. I also happened to purchase UT3 recently so I am now one step away from getting into UDK - downloading it. Looking forward to seeing what in the world I could do with UDK...

I watched an hour's worth of the tutorial videos. UDK's setup looks similar to Hammer's setup in that there are a possible 4 viewports. It differs in that you can drop lighting into the map at any time and have it show up right away, and you can drag it around in real time and watch the lighting and shadows move in the editor. Very cool. There are a myriad of other things that UDK can do that make me jealous when I compare it to Hammer.

However, with all that said, I think Valve is poised to create a truly awesome new editor with the hypothetical new engine they create. For the most part, Hammer is intuitive and fun to use. I can foresee the next version of Hammer including a lot of the features I spoke of above, as well as including a ton of new things too. We just have to be patient while Valve creates its new engine and wises up when it comes to designing the new editor.

As far as the enhancements to the Source engine that come along with Portal 2 - Will they really be useful when P2 comes out? Or will we all have to wait until Episode 3 comes out? I have to assume that Episode 3 will not only include most of L4D2's and P2's tools and effects, but will also bring its very own unique engine enhancements.

The best thing anyone could ever do IMO is take the Director and the navigation meshes from L4D and take the dynamic lighting solutions of Portal 2...and toss them into a Half-Life 2 setting. It would not be the typical single-player developer-crafted experience, but it would surely be awesome.

[Gary]

Give me proper dynamic lighting and I'll be happy with the engine.

But the editor is still a bitch to work with compared to the CryEngine's Sandbox, I love how it's real time, it makes things so much faster, you can just hit a button and pop right into the game and play your creation. It's pretty user-friendly and runs smoothly.

[Surfa]

Tbh I would much prefer lightmapping on static meshes then full dynamic lighting. The udk is a great example of both combined to reach the best results.

[Portalboat]

I'm not sure if this is just Source or it's with every game engine, but weightmapping is HELL. There HAS to be a better way to do it....

[Gary]

Tbh I would much prefer lightmapping on static meshes then full dynamic lighting. The udk is a great example of both combined to reach the best results.

Why? If it's "radiosity" that you want, dynamic lights can do it too.

[Surfa]

Cheaper and provide better results.

[Smurftyours]

Once my computer returns from repairs I will learn how to use UDK. I still have hope that source will get some major upgrades after the release of portal 2, but until then I am disappointed with its lack of modern technology.

[Gary]

Cheaper and provide better results.

I wouldn't say better... static lightmaps are static.

[Mr. Happy]

The only two things that really annoy me about Hammer are the lack of a visual scripting interface and that as a level gets more and more complicated you start getting things stacked on to each other and you have to spend alot of time positioning the camera just so to select what you want. Look at making a spawn room in TF2 for example, you've got the visualizer brush, the door brush, the respawn room brush and the door model all in the same place and it can be really difficult to select the right one.

[Phott]

The only two things that really annoy me about Hammer are the lack of a visual scripting interface and that as a level gets more and more complicated you start getting things stacked on to each other and you have to spend alot of time positioning the camera just so to select what you want. Look at making a spawn room in TF2 for example, you've got the visualizer brush, the door brush, the respawn room brush and the door model all in the same place and it can be really difficult to select the right one.

Just select the entities/brushes/etc you want to show up.

[Armageddon]

Don't you know Vis groups boi?