Havok buys Kore

I was just about to post this :)

http://www.siliconrepublic.com/business/item/18511-havok-makes-first/

wow…story seems to suggest that Kore has already been licensed to a good few major companies.

Aphra.

—

Well done to steve and all the guys. Great news for all concerned.

The product, which is shortly to be rebranded as Havok Script, is a Lua compatible VM designed for console game development[/quote:486169a146]

http://www.unknownworlds.com/decoda/consoles/

June 11, 2009 – Decoda now offers on-target, PS3 and Xbox360 Lua debugging.

Connecting via TCP/IP, Decoda links to your running game and provides essential on target debug functionality; Interactive console, Pause/Continue, Step over current line, Step into function call, Step out of the current function, Toggle Breakpoint, Watch window.

Available to registered Sony and Microsoft developers only, the console version of Decoda provides seamless debugging for game professionals.

Please contact Decoda@Kore.net or call +1 415 830 5650 for licensing information.
[/quote:486169a146]

Bits and pieces that helped (me) to fill in the blanks.

Congrats to all involved.

B!

—

Well hooray for Havok. Now, maybe, they can hire more non-indigenous programming talent, seeing as they have zero respect for our 3rd level institutions graduate output. Once again, on one of my very rare visits to this excellent website, I find everyone getting excited about something that’s almost completely irrelevant to Irish GAMES DEVELOPMENT.
Don’t forget that when an internationally prestigious company like Havok badmouth graduates here, it resonates globally and may even influence a decision being made by an actual games development company on whether to relocate here to take advantage of Irelands myriad attributes. Location, language and corporation tax as examples.
Don’t get me wrong, I’m happy to see an Irish company prosper and get acclaim for their really quite amazing work. But when they provide very few jobs and in fact ‘talk down’ the local talent ….. And if anyone thinks that badmouthing our graduates will result in positive change in 3rd level education (especially during a harsh recession that’s actually morphing into a textbook example of a depression) then they are fooling themselves.
:evil:

Well everyones strapped for cash these days, especially the colleges, unis and techs. That’s only gonna get worse. Much worse. So I don’t see any positive change occurring there any time soon. We gotta work with what we got. Anyway, who says things aren’t ‘grand’? Sure, maybe the 3rd level bodies are too focused on quantity rather than quality but that depends on what branch of development your intending to cater for. And all we have here is localisation and porting (for the most part). I’d argue that the majority of our grads can handle that kind of work easily. I’ve had a fair amount of experience with 3rd levels grads here with regards to games development and I’ve rarely come across anyone who can’t do the job. It’s certain companies, in my opinion, who have the medias ear, that are trying to set the bar way too high e.g Havok. I’ve nothing personal against Havok, but I did once see them advertise a 3d artist job where the only candidates considered had to know 3Ds Max, Maya and Softimage to a high level. Those people simply don’t exist. If they do, they are as rare as hens teeth and in my 16 years in game development I’ve never once met one.

I think you need to look at HAVOK’s intentions in criticising the quality of maths and science grads.

Their comments may upset some people but regardless it raises the issue. Ireland needs better math and science from as early as possible.

Not easier. Not assessed with a curve. Not in anyway skirted around.

Teachers need reskilling, courses updated the works.

Everything else is side issues really.

Intel’s chief raised the same concerns, so its not like Havok are alone on the criticism.

As for hiring more locals for the sake of it, or to give them a shot, I’d say thats an ‘ignorant’ statement in the sense that it seriously underestimates what Havok do. They’re world class for a reason. This stuff is hard. Really hard.

Anyway thats my 2 cents, I’m not a Havok fanboy, I believe you need to clearly point out the elephant in the room and make govt take action otherwise it’ll never ever get fixed.

I’m sure Havoks intentions are pure enough, if a little selfish. They want the best coming out of Irish 3rd level institutions. Maybe so they can hire quality grads themselves and save money on relocating people from abroad. Or maybe it’s because they want to see things pick up in Ireland in general. I’m sure Intel have voiced similar views. Either way, they are irrelevant. What, exactly, have Havok and Intel got to do with nurturing an indigenous game DEVELOPMENT industry….in Ireland?? Havok (and Intel) not being developers of games.
For what Havok and Intel do, maybe they do need a higher standard. But how many jobs do they create? And I’m not talking about people working on the production line here. One or two jobs from a couple of companies every six months does not make an industry. And again, neither company develops games.
And with regards to our 3rd level bodies, re-skilling, course updates etc. costs money. People will have to be fired and replaced with more expensive people. Ain’t gonna happen in this current economic environment. You can lobby and fight for it until you’re blue in the face, but it’s just a big old battle that you can’t win, it’s a time waster that will get you nowhere. And why bother? There’s no willingness for this, because, as I’ve said before, the current job market here matches the education system currently.
The UKs massive development industry got going through indie developers making money and growing. The UK should be our model. Small, indie companies become big ones, eventually, if all goes well. There are small companies out there, abroad, who have not planted their roots yet and would consider Ireland as an option. Reading Havoks, blinkered semi-selfish, media promoted views is not going to push them to these shores.
Nobody’s suggesting that people hire locals for the sake of it. And if Irish grads ain’t up to the job for Havok and Intel then, tough. Get what you need from where ever and shut up. Don’t come out in the media bad mouthing the entire education system just because it doesn’t fullfill your particular needs, when the system works fine for everyone else operating here. The 3rd level education system in Ireland fits the market, currently. That’s is localisation, porting and to a much, much lesser extent pure, low level, development.

We need the companies here first, growing, with growing educated staff requirements. Then the education system will follow, and try to provide grads matching those requirements.

And screw Intel by the way. They’ll be outta here as soon as it’s economically viable to go somewhere cheaper (where they can also find the staff).

Well I think fair play to Havok and Kore on this deal, which is what this thread was originally about. This website supports all companies in game development or related in Ireland, and since Havok’s products are aimed at game developers….I think it is fair enough for people to get excited about it.

With regard to talent and recruitment Havok have very particular needs and I think most people realise these are not what many graduates, at least at undergraduate level, have.

The discussion around science and maths education is in my opinion a much larger issue but we can’t focus on these skills solely and expect that is the solution…most studies of innovation in tech companies internationally would suggest that technology skills alone do not a successful business make. And perhaps the discussion is more around creativity and problem solving skills anyway.

Aphra.

Some thoughts:
* Kore are a great company. They’re exactly what the country needs. High tech, high skilled work, successful exit – bought by Intel, brings money into the country. They should be commended. They worked on a lot of AAA games, at a highly skilled level, which you can’t say of many Irish companies.

* Having successful high end middleware companies is surely a good thing for those who want game dev here?

* The CEO of Havoc who ‘bad-mouthed’ Irish 3rd level is no longer with Havok, so its not really relevant to this thread. Its worth noting the lingering perception his comments left as non-constructive criticism though.

* The Irish science and maths education system has a *lot* of issues. It is insufficient to be good enough for localisation and porting. You have to educate ahead of the industry, not behind it. Something I never see in this discussion is an acknowledgement of how much the bar is raising internationally.

* That said, if there’s any success story for the system, Kore has to be it. Skilled developers and experienced management do ninja work, win. ‘Fair play’ doesn’t go nearly far enough.

That’s what the country will need when Intel head off, as you say.

@ Roganski

I’m not sure about why you think its such a big deal that Havok are saying this about Irish grads. All of the UK games companies are saying the exact same thing about the grads in the UK. I’ve even interviewed a few. The skillset of the majority is below par. Universities need to eat some humble pie and stop advertising courses that they are unable to teach. Yes, some onus must go on the graduate themselves to push themselves, but a lot of graduate courses just aren’t preparing the student for the career they are supposedly educating them for. Again, this is not an Irish only problem. Havok are most vocal about it in Ireland as they are probably the largest employer of this talent-pool in the country.

Well, firstly, when a member of staff of a company speaks to the media he represents that companies views. The fact that he’s no longer on that staff is irrelevant. Havok have not withdrawn his statements or distanced themselves from those statements to my knowledge.
Secondly, there is a global dependency on graduates for new staff in the games industry. This is a big change from the 90s when the majority of hires were people with experience. People who were self trained. People who coded on C64s and Spectrums in their bedrooms and worked their way up. People from the demo scene etc. long before the 3rd level educational bodies even got in on the act. Perhaps some companies do expect too much from grads these days. You need experience to do the job. Raw grads arrive with some knowledge but really need years of training, teeth cutting time, to become valuable assets (depending on what kind of development you are talking about). I mean, how many lecturers and tutors actually have experience in top notch games companies? Very few I’d wager. Those who can, do, and those who can’t, teach. So you already have limits to what you can impart to your students from the start. Anyone who expects grads to be able to take on senior roles instantly is going to be disappointed. You gotta put the time in on training. Complaining about it isn’t going to change anything because, where will the new teachers come from? You can’t get a coder with ten years hands on games development experience to jack it all in to become a tutor, for a third of the money and twice the stress. It just won’t happen. And where will all the extra money come from to hire these tutors (or ‘upgrade’ them) if they were willing? And again, what is the point in training people to develop on, for example high end consoles when there’s no jobs in this country?
And I totally disagree that our 3rd level bodies aren’t adequately providing for our current market here, which is mostly localisation and porting. I’ve just spent the last 6 years working in that field and many new hires are straight out of college. Obviously they need onsite training. But that’s always been the way with fresh grads.
So complaining about 3rd level bodies won’t help bring about an industry. It won’t, suddenly, give you grads that you can slot in anywhere on a production pipeline. They are limited in the knowledge they can impart, because they simply don’t possess that knowledge to begin with. They can only help with part of the preparation for the job market. It’s always been that way and it always will be that way. We have to accept it and make do.

The only way, in my opinion, that we can have an industry here on par with the UKs massive industry (which is a good model to copy) is for already experienced staff to set up indie houses with a core, experienced team. Artist, designers, coders, QA, producers/ project managers etc. and bring in grads at junior levels and train them up to the desired level. I think this is already happening. It doesn’t take a lot of money to get this started either. IPhone development, for example, is a great gateway.

If we don’t do this we’ll still be having this debate years from now and there’ll still be no real industry here. There are plenty of ex-pats on this forum who have the required talent and experience who would love to come back here to work. I was in that boat myself for 9 years. To them I would say, start saving money. Try to get a core team together by offering all involved shares and royalties. Start some low level development on iPhone (for example), come back, bring in some grads if required, set up shop and have a go. It’s not impossible and it doesn’t have to cost millions. It’s already happening in fact.
‘Mighty oaks from tiny acorns grow’. That’s how the UK did it.

Random thoughts:

The goal of any company is to make money, grow and survive – not to "support the Irish games industry". Those which do (like Demonware supporting this site) are greatly appreciated, and one would hope they get a return in terms of discovering good people through their involvement with the site. Like all things, it’s far from guaranteed however.

Demonware and Havok were both founded out of Irish universities and managed to get where they are today based heavily on indigenous talent. The fact that they now need very specific skillsets which are in rare supply around the world is hardly surprising, but there is also no suggestion that they are somehow ignoring quality Irish graduates in favour of international ones.

Not everyone who works in games wants to own or start their own games studio. Even if they did, they may be unable to afford it. Of those that do, starting something big in Ireland would require the tax and investment incentives to be competitive with other jurisdictions. In the mean time, doing smaller indie and/or mobile development is a route for people which does not entail the cost base of a massive studio.

Dave

After the discussion at the Dublin Web summit is became clearer that we are not producing enough graduates with appropriate skills. The Havoks and Demonware need very good engineers and computer scientists and unfortunately fewer and fewer students were choosing those courses over the past decade. Those that have and can demonstrate their skills are being snapped up, often before they complete their qualifications.

But it was also emphasised, as Skyclad just mentioned, that there are a lot of opportunities now for small teams and individuals to create indie, social and casual games and to self publish or target some of the online distribution channels. So yes it would be great to have a major studio come here, but we can’t just wait around for that to happen.

Aphra.

I am no expert by any means but the answer seems intrinsic to me. Both companies were started by graduates?

The influx of companies seems to have come afterwards. I don’t think its by chance that games companies locate close to each other which can be seen very quickly from the dev map.
http://www.gamedevmap.com

How to encourage startups… I am not quite sure the best route here.

Perhaps some companies do expect too much from grads these days. You need experience to do the job. Raw grads arrive with some knowledge but really need years of training, teeth cutting time, to become valuable assets (depending on what kind of development you are talking about). [/quote:0b6a627797]
As mentioned before, I have interviewed grads in the UK in the past, and its not a case of oh, they just need some time. Grads are leaving university without a proper understanding of OO programming, without a proper knowledge of how C++ works…that is not the kind stuff we can hope they pickup on the job. That is what I mean when I say they are not being adequately prepared. For a grad, I would expect that a graduate understands how to code, has a pretty decent knowledge of whats going on and has a good demo. Its suprising how rarely this combination comes around. I can only image the situation in Ireland is the same

Was there a switch away from C++ at some point in computer science courses and if so why?

Not being a computer scientist I can’t say much on this topic.

Aphra.

Was there a switch away from C++ at some point in computer science courses and if so why?

Not being a computer scientist I can’t say much on this topic.

Aphra.[/quote:391ee4c557]

I went to DCU and so can’t really comment on any other colleges. Java was the language used to teach the basic concepts of programming in first year and was the main language for the 4 years. There modules that used C and C++ – the majority of which were optional – probably amounting to a semester and a half in total.

It was quite difficult with all the course work to dedicate enough time to learning the essentials of C++ (just as an example it was only through my own research that I learned about things like the copy constructor in C++ as things like that weren’t covered in the module).

Also, Sun Microsystems donated a lot of stuff to the Department so that may have influenced a few faculty decisions.

There are probably a multitude of reasons but IMHO the focus on Java was to produce graduates who could fit right into the business world – almost all of my friends work as Java programmers doing mobile/server side work and went straight into jobs after graduation.

The one friend who works with C++ works abroad for a Telecoms company and in a very niche role.

From my own experience it was a lot faster to knock up a project in Java with a GUI, connect to a database, network it etc in Java than in C++. And faster is better for businesses.

A lot of ‘games’ courses seem to go for Java, Flash development in the early years as (a) flash is fairly visual, and therefore easy to understand and teach, and (b) there was a huge influx of mobile games on phones using java and I assume lots of unis were still thinking dot com and teaching web courses and put 2 and 2 together and thought java + plus is a obvious thing to make a good games course. Later on they get into the nitty gritty I think and concentrate on C++, but many students grads just don’t have the in depth knowledge of the language that a Computer Science graduate would get. If someone asked me the best course to get into games, I would recommend a good Computer Science course over a games course.

The switch to Java was happening as early as 1998 in UCD when I was there. Java makes sense for colleges wanting to place students into the typical programming job. The problem with Java is that for the sort of person who learns concepts by experimentation (me) it doesn’t give you much exposure to direct memory management and this is still very important for performance in games.

More controversially I’ll suggest that Java is over focused on an object oriented model which really a subset of what is possible in programming – and not even a particularly good subset. C++ is too, but it gives you glimpses into a world of coding where you think more in terms of just transforming buffers of data – and that’s usually a better mental model for performance.

On the C++ issue. I think there was a shift away from C/C++ around 2000 or so.

Java came in (as there was, IMO, more java jobs opps than C/C++) and that degraded the core skills for engineers (IMO), low-level memory management. Pointers, allocations, stride and all the rest have (again – IMO and in my experience) become lost skills. Java does all that stuff for you "under the hood."

So, as has been pointed out already, new-fangled "games" courses tend to start off with *pretty / visual / point-and-click* stuff like Flash and then Java and maybe give a cursory discussion on heap/freestore vs. the stack.

Also, C# (proprietary programming language – bad start, IMO) is making its way onto more and more courses, and its having the same effect. It handles all the underlying memory allocations/frees for the programmer.

Again. This is leading to grads shipping with no understanding, or mental models (and experience) of iterating, casting, reinterpreting buffers, or tracking memory leaks.

In short, Software Engineering / Comp Sci should be 4 years of C/C++, with as much Maths, Physics and Graphics labs/papers as possible to reinforce learning the language. IMO.

Now, having said that, I’ll flim-flam and point out (before someone else does) the cases where the above does not apply.

If your not aiming for AAA game development, and want to focus more on casual games / bedroom coding / demos, then Flash, C#, Java may be ok. Just don’t expect Ubisoft, SCEE, Nintendo or anyone else to be interested.

Right. Thats my thoughts.

When I’m screening a graduate CV, here’s what I look for:
A) A strong level of C++ experience
B) A polished game demo that they’ve created on their own time
C) A strong grasp of general computer science & game technology

If I don’t see those skills we can’t hire you, simple as that. It’s frustrating that many games courses tend to focus on B & C and all but ignore A. From talking to lecturers, I think it’s a supply and demand problem. It’s hard & expensive to find experienced C++ lectures and easier to find Java/C# lecturers, so they take the path of least resistance. As a student, you’re missing a critical part of the training you need. Complain loudly and seek out the courses that do have the correct focus – they do exist.

If you are on a games course that lacks a strong C++ component, don’t be passive about it. Start learning it yourself for a demo or join an open source C++ project and learn from others. It could make the difference between getting an interview instead of a rejection letter.

For people interested in what Trinity are teaching in terms of C++ or Java here is what I have seen. There are two options if you want to study Computer Science in Trinity 1. pure Computer Science and 2. Engineering with Computer Science.

1. The Computer Science students start off by learning Java and ARM assemble language in first year, but then have to do some C/C++ in second year.

2. The Engineers start off by learning C++ in first & second year but then in third year have to do a project in Java and also a course in ARM assemble language.

I personally don’t know why the two different approaches are taken, but that is as far as I can see is what is happening.

Thought I might provide the contrarian argument :D

When I were a lad, C/C++ was the high-level language and if you wanted to get real performance you learned assembler. In my course, lecturers taught C/C++ as it was generally considered ‘easier’. Sound familiar?

If you want to write high-performance code you need to have an indepth and uptodate understanding of the platform your code is executing on. Thats true in every language/VM/platform combination.

C/C++ has been the choice for writing high-performance code because it provided a thin veneer over the underlying processing model (single-core, random-access memory) that allowed you to write code very close to the metal and get corresponding high performance.

TBH I don’t think that holds any more, cpu/memory/network architectures have moved on, and the C/C++ mental model is just as abstract as any VM now. I suspect the reason people use C/C++ now is as much to do with familiarity then any inherent advantages over other newer languages.

I think all CompSci / CompGames courses should be theory focussed and language agnostic. Good programmers should be able to write clean / efficient code in any language/vm, and game programmers, more than most, have always had to be master of all trades.

The future for me will be in new languages and DSLs that allow the programmer to focus on the task in hand. There’s several in the mainstream already (HLSL, OpenCL, Erlang, Scala, Lua) and I’d expect that trend to continue. CompSci courses should be educating graduates with that in mind.

-Peter

NB I would take issue with the idea that GC based languages (C#, Java et al) ‘take care’ of memory management for you. The Java Memory / Execution Model is significantly more complicated than perhaps most people realise, but deep understanding of it is critical to writing efficient / correct (esp. thread safe) code.

@petermd

When I were a lad, C/C++ was the high-level language and if you wanted to get real performance you learned assembler. In my course, lecturers taught C/C++ as it was generally considered ‘easier’.
[/quote:3b66ec2581]

-SCEE still push to have assembler taught as much as possible. They had a press release to that effect and still want grads with that skill. So assembler hasnt gone away.

Link: http://develop.scee.net/recruitment

If you want to write high-performance code you need to have an indepth and uptodate understanding of the platform your code is executing on. Thats true in every language/VM/platform combination.

C/C++ has been the choice for writing high-performance code because it provided a thin veneer over the underlying processing model (single-core, random-access memory) that allowed you to write code very close to the metal and get corresponding high performance.
[/quote:3b66ec2581]

Agreed.

TBH I don’t think that holds any more, cpu/memory/network architectures have moved on, and the C/C++ mental model is just as abstract as any VM now.
[/quote:3b66ec2581]

Not sure what you mean here. Threading, multi-core, and everything else that has "moved on" is (in my experience) included in Boost, and/or on its way in 0x.

I suspect the reason people use C/C++ now is as much to do with familiarity then any inherent advantages over other newer languages.
[/quote:3b66ec2581]
Partially agree. I think a major reason why C/C++ is the de facto for games dev is because (1.) all the existing, in-house codebases that have amassed over the past ~20 years and (2.) all the third party libraries (fmod, etc) all use C/C++.

I don’t see either of those changing anytime soon.

I think all CompSci / CompGames courses should be theory focussed and language agnostic. Good programmers should be able to write clean / efficient code in any language/vm, and game programmers, more than most, have always had to be master of all trades.
[/quote:3b66ec2581]

Partially agree – and I guess this really is the crux of the argument. On one hand people go to college so they can learn "how to learn." However, grads also want a career at the other end of the tunnel. And employers want (rightly or wrongly) grads who can be productive from Day 01.

So that (IMO) is the gulf. Shipping grads who on paper know Comp Sci theory verbatim, but don’t know the in’s and out’s of the language, as close to the metal as possible, is doing the students a disservice (IMO).

In the end, that gulf has been there a long time.

Academia will argue that Employers are asking for too much, and that Uni is not a trade school.

Industry will counter-argue that the level of grad coming out is not "industry-fit" or lacking the requisite knowledge.

Academia will argue that they teach to an accepted curriculum and that they can’t pander to the particular needs of Company X.

Industry will counter-argue that they don’t want generalists, that they instead want to take grads who know the field (thesis projects) they want to specialise in (core, graphics, audio, physics, etc.) and continue to groom them for that role.

ad infinitum.

I can see merit on both sides.

What I would say to all involved (student, teacher, employer) is to make your expectations clearly known to all parties.

If your a student, and are signing onto a Game-focused Comp Sci program, I think its fair to have an expectation that you can go and knock on Sony’s door at the end of the course, apply for a grad role, and not be laughed at for having zero C/C++ experience.

More generalists Comp Sci courses – well there’s a range of stakeholders there, (banking / finance software, database admin, network admin, etc.) so on a course like that (IMO) a more language-agnostic approach, I suppose, is fair and equitable.

Clearly defined, informed and reasonable expectations are key.

NB I would take issue with the idea that GC based languages (C#, Java et al) ‘take care’ of memory management for you. The Java Memory / Execution Model is significantly more complicated than perhaps most people realise, but deep understanding of it is critical to writing efficient / correct (esp. thread safe) code.
[/quote:3b66ec2581]

Preventing, detecting and finding memory leaks in C/C++ is a key skill for game development. Shipping a grad off to interview at Sony without this knowledge (from a Games-focused Comp Sci course) is starting off 3-0 down.

Anyway, that’s my contrarian argument to your contrarian argument.

Preemptively agree that not every student will want to work at Sony.

Preemptively agree that there are *games* companies out there who use Java. (Jagex – best example). But if I walk into GameStop, I’d say 99% of titles there are coming from a C/C++ codebase.

Brendan.