Sorry about posting this topic here. But I believe this is the most section people read at Gaming forum.
Is there any way to model a game using UML? If it is, can someone give us an example? I can’t see the loopings, actions, events and interactions inside any kind of diagram. (The most closest I think I could use is the Sequence Diagram, but I still can’t visualize a Game Diagram… )
sorry about the off-topic.
In most cases games uses complex class structures in the background to represent game state/AI/etc… so class diagrams are happily useful for this sorta thing.
However, if you’re talking about game loops, state machines and rendering pipelines then you could consider Collaboration Diagrams (really a sequence diagram on its side), Activitiy Diagrams (sorta flow charts) and State Transitions Diagram.
Kev
Normally you can use UML for all types of OO programming (dont know if UML support functional languages like C ??)
But what do you mean with Game Diagram ?
Raven
It would be a kind of Diagram that you could see everything. Visualize the events that users can dispatch, and what they would do with the game state.
How loops would take different ways, and all that stuff.
Maybe a UGML??? UGameML. 
I don’t know, maybe I drunk too much Coca-Cola having fun at Lan House.
loops? … that is already covered in UML as a State Chart
Ok, I know that. My propose is to make something like, put all togheter on the same Diagram. Like WYSIWYG.
And not pumping diagrams for each stuff all over. I don’t know if this is possible, is just a newbie’s idea.
I dream with a diagram that I could see the game itself. Just waiting for been coded. Wouldn’t be a great thing?
you “could” make whole “game” into one big “by-nobody-else-but-you-understood” State UML chart…and it is possible…and I guess it could even be viewed uopn as a noble - though lofty - idea, but in real life and AFA game design and developement goes it is not practical . UML is nice for web design… GUI … and other types of developement where you have MANY programmers to run around. It is meant to be something like map where to go and helps to keeps order, nothing wrong with that. When it comes to game developement - well - tell me how many dozens(or perhaps thousands) of programmers you have?..one?..two? …suprised if more then 5. I tend to believe UML is waste of time for this type of developement (even though I am using it very day at work). You think of what you wanna achieve…make small demo…OK?..if so then go aheed and code bit more…once you get things done…go back refactor what is too many times repeated and move to next task on you list…That is my way of doing games… I have couple of hundred of classes I got done over say 4 year and everything is nice and clear… Reality is if someone from the place I work would read this I would be without paycheck on the spot ;D ;D ;D
UML is always good if you have a complex application and many people are working on. Non-develper also can better understand the internals. Or if new ones comes to the project.
But for a small group of developer, it’s often too time intesive to create UML diagram, because nobody uses it.
Raven
[quote]you “could” make whole “game” into one big “by-nobody-else-but-you-understood” State UML chart…and it is possible…and I guess it could even be viewed uopn as a noble - though lofty - idea, but in real life and AFA game design and developement goes it is not practical . UML is nice for web design… GUI … and other types of developement where you have MANY programmers to run around. It is meant to be something like map where to go and helps to keeps order, nothing wrong with that. When it comes to game developement - well - tell me how many dozens(or perhaps thousands) of programmers you have?..one?..two? …suprised if more then 5. I tend to believe UML is waste of time for this type of developement (even though I am using it very day at work).
[/quote]
Argh! Blind leading the blind!
UML is something like iteration number thirty-something in 4 decades of research into planning software projects. If you think you can improve it, you’re either very experienced at this sort of thing, or your trying to solve a different problem, and shouldn’t look at UML anyway (there’s very likely something else that already solves your problem).
A primiary aim of UML is that you have separate diagrams (six main ones). Trying to put it all on one diagram is the worst possible thing you could do, since you’d break one of the fundamental features of UML. The concept behind this (very briefly) is that the different aspects of a project (and even the different people working on it) require different perspectives on the system. Each of the diagrams approximately maps to a different perspective/view of the system.
UML has nothing to do with the size of the project, or the number of people working on it, etc. It is equally valid and equally useful for a lone programmer as it is for a giant corporation designing an enterprise app for 20,000 users. In fact, to a certain extent, it has problems on really large projects because it can be a bit too brief in some respects (although YMMV depending on how you are using it).
A game developer who decides UML is “not worth it” for game-design probably (note: probably - I can think of exceptions myself, but they are the minority) needs to rethink/relearn their development process…or else accept that they’re not that serious about the game they’re designing. Since this is possibly a controversial statement, I’ll try and explain: lots of people don’t use UML whilst appreciating that they should, and that it would be worth it. For instance, the lack of cheap UML modellers is a major barrier to many people using UML who would otherwise leap at the chance.
And, if you’re a developer who thinks UML isnt’ worth it, you’re in a similar league to those developers who think documentation isn’t worth it, period. No, we don’t need to meticulously document every aspect of everything we do, but research has shown again and again that even lone programmers draw immense benefit from moderate-to-good documenting of their own code.
…Unless you’ve got a similar but alternative methodology to UML, in which case replace “UML” above with your methodology. I’m assuming that the alternative you consider is “no methodology” (apologies if this assumption is wrong)
I suggest the original poster does a google search for UML tutorials and finds one that lists 6 main diagrams and explains the uses of each. There are also entire rainforests of free tutorials, guides, documentation on using UML available from the website of the consortium that manages the language (OMG.org, IIRC). It’ll take you years to read it all and not all of it is worth reading, but it’s a good starting point (if you’re brave enough).
[quote]UML is always good if you have a complex application and many people are working on. Non-develper also can better understand the internals. Or if new ones comes to the project.
But for a small group of developer, it’s often too time intesive to create UML diagram, because nobody uses it.
[/quote]
If you cannot get a UML Modeller, yes. If you have a modeller, that’s just complete rubbish. The aspects of UML that programmers spend most of their time looking at are mainly class diagrams and sequence diagrams, BOTH of which are auto-generated by every UML modeller I’ve ever seen. Time taken to create diagram == not many milliseconds (unless you’re using ArgoUML or ROSE. Ugh)
EDIT: I wasn’t thinking straight. Actually, I’ve seen a lot of things that claim to be UML modellers but don’t do sequence diagrams - they should really call themselves “class diagram modellers” because that tends to be the ONLY thing they can do.
Halleyuuah!
Apart from the major factor of helping you communicate your ideas to peers, UML is immensely useful for organising your own thoughts and refining ideas. Often, once you put your “great” class structure idea and try a few sequence diagrams you can find huge holes that would have otherwise waited until code.
I’m afraid a lot of time I come in the category of knowing its a good idea, but not doing it, mostly cause I enjoy hacking at code (sad I know). However, I always seem to end up going back to UML designs (normally on paper) to figure out what I should have done.
Finally, another plug, http://www.newdawnsoftware.com/jose is a UML editor I worked on that is early stages of development (Class Diagrams, Sequence Diagrams only). Well worth a look if you don’t want to go too heavily into UML but still recognise its usefulness.
All the best,
Kev
"Argh! Blind leading the blind! " just because you are of the different opinion gives you right to mock others … time to go back to your charts and rethink it one more time. - if that brings you back to the initial state you might be hiting never- ending loop …no good …BTW I didnt say UML is bad or useless - I spent 3 hours per day on Together/J to sync team’s changes …I’ve said it is overkill for ‘presumably’ small gaming project. Even on big release game you usaully see couple of names those who actually do dirty coding work … in such a enviroment and in REAL life people talk to each other becouse they CAN…in big team folk needs some other means of “keeping in touch” and that is where UML plays its role. And if you still think this is not the case and have diferrent opinion - so be it - i accept that - but without making any additional remarks poited toward you…
[quote] "Argh! Blind leading the blind! " just because you are of the different opinion gives you right to mock others … time to go back to your charts and rethink it one more time. - if that brings you back to the initial state you might be hiting never- ending loop …no good …BTW I didnt say UML is bad or useless - I spent 3 hours per day on Together/J to sync team’s changes …I’ve said it is overkill for ‘presumably’ small gaming project. Even on big release game you usaully see couple of names those who actually do dirty coding work … in such a enviroment and in REAL life people talk to each other becouse they CAN…in big team folk needs some other means of “keeping in touch” and that is where UML plays its role. And if you still think this is not the case and have diferrent opinion - so be it - i accept that - but without making any additional remarks poited toward you…
[/quote]
Don’t take this personally but you appear not to fully understand what you are talking about. Every time you make statements such as “nd in REAL life people talk to each other becouse they CAN…in big team folk needs some other means of ‘keeping in touch’ and that is where UML plays its role” you show not inconsiderable ignorance of what UML is, what it is intended to be, and how it came about.
I’m not saying you are wrong in highlighting a particular use of UML, nor that you don’t know anything about using it. But your perspective is artificially narrow, and really doesn’t do the tool any justice. Your absolute statements are just plain wrong. Here is a quote about UML and size of projects, from the OMG’s Intro to UML (which is worth reading, it’s fairly light and contains some useful pointers http://www.omg.org/gettingstarted/what_is_uml.htm ):
“Of course a well-designed architecture benefits any program, and not just the largest ones as we’ve singled out here. We mentioned large applications first because structure is a way of dealing with complexity, so the benefits of structure (and of modeling and design, as we’ll demonstrate) compound as application size grows large. Another benefit of structure is that it enables code reuse: Design time is the easiest time to structure an application as a collection of self-contained modules or components.”
The 3 amigos (Booch, Jacboson and Rumbaugh - the inventors of UML) handed over it’s safekeeping to the OMG, so the OMG’s documents are what passes for the official word on the language (and, incidentally, manage the standardization process for new specifications). They clearly do not see it as primarily a communication tool (as you appear to do).
I made the eye-grabbing statement about the blind leading the blind because you were way off base, not just slightly. No offence was intended - just to shock you into re-evaluating quite deeply this tool.
Here are some more salient thoughts from the Intro to UML. Unfortunately, their docs are mainly not for a mass-market, and quickly move into extremely high-level abstract discussions, but I’ve always found it works quite well if you just quickly skim over the parts where they start to use too many vague words (anything that sounds like marketing tripe ;))
“Modeling is the designing of software applications before coding. Modeling is an Essential Part of large software projects, and helpful to medium and even small projects as well. A model plays the analogous role in software development that blueprints and other plans (site maps, elevations, physical models) play in the building of a skyscraper.”
" You can do other useful things with UML too: For example, some tools analyze existing source code (or, some claim, object code!) and reverse-engineer it into a set of UML diagrams. Another example: In spite of UML’s focus on design rather than execution, some tools on the market execute UML models, typically in one of two ways: Some tools execute your model interpretively in a way that lets you confirm that it really does what you want, but without the scalability and speed that you’ll need in your deployed application. Other tools (typically designed to work only within a restricted application domain such as telecommunications or finance) generate program language code from UML, producing most of a bug-free, deployable application that runs quickly if the code generator incorporates best-practice scalable patterns for, e.g., transactional database operations or other common program tasks. Our final entry in this category: A number of tools on the market generate Test and Verification Suites from UML models."
“One characteristic of UML - in fact, the one that enables the widespread industry support that the language enjoys - is that it is methodology-independent. Regardless of the methodology that you use to perform your analysis and design, you can use UML to express the results.”
…that’s also a good place to start when exploring the different diagram types. Although, as I said, I tend to find their documentation a bit too ultra-high-level, and prefer things such as:
http://www.visualcase.com/tutorials/uml-tutorial.htm
where the wording is a bit more natural. A good google search should dig up some good tutorials (but beware the large amount of dross!)
It’s passages such as:
“The UM is specified via a metamodel. The other layers in this pattern are the meta-metamodel layer, the model layer, and the user objects layer. The metamodel layer is derived from the meta-metamodel layer, which for UML is defined by the OMG Meta Objects Facility’s (MOF) meta-metamodel.”
which do it for me
i didn’t take it personally … but I am not plain wrong - I am seeing and commenting on this issue in terms how “real” world is working - and that might be PLAIN wrong I agree. If you are purist with vast knowledge - and I presume you are you shall beat me with ruler over my head as if I was beaten during my undergrat courses :). But I guess issue stated at the start of this thread was should I use UML … I said it might be overkill in most scenarios.- and - then you gave me lesson about what UML is. I can assure you I read some books on it a decade ago and more or less am using it every day so I tend to think I have a bit of understanding of the subject. BUT back to games - almost nobody here - or just a few lucky (unlucky ones) claim making games as their daily job. Do you know many of those few who are using UML as a nessecity for achieving their goals? After reading scores of books on graphics programming and few on game making I am still to see one UML chart in any of these books. And if I say make stuff as you go - stop - go back and clean it up - is not unreasonable approach - that is described in other books written by those who claim that is the only way . HAving said that there has to be structure in what one is doing - cannot agrue with that - and I am not just dreaming and then code I have dozens of charts of project I am on for long time. I was using Java3d and scenegraph representaion at least in my opinion gives you order and that is what I am using. I have charts all branches and sub-branches and it took me minutes to explain it to total newcomers after they looked at it. So I say it is enough for me - but definitively agree that there is some other better more offical way of handling it. But hey its just game 
;D
[quote]i didn’t take it personally … but I am not plain wrong - I am seeing and commenting on this issue in terms how “real” world is working - and that might be PLAIN wrong I agree.
[/quote]
OK, I see what you mean now.
Almost every skilled and serious developer I know well uses UML or an equivalent (e.g. specialist techniques for specialist methodologies, like XP).
Many people I have worked with avoid working anywhere that doesn’t use UML/similar tools, largely because it’s like an architect working on a construction project where no-one uses blueprints / plans - it’s a recipe for disaster, and although you MAY survive, it’s like playing Russian roulette.
I take it you don’t read Game Programming Gems series then? Almost the first gem in GPG2 uses UML diagrams IIRC, and from memory several others do too. And this is in a publication where space is at a huge premium.
“The Mythical Man Month” by Fred Brooks would be a good place to start if you haven’t read it already. For anyone who’s wondering, it’s a small book (very small by today’s standards!) written by the man who ran a 300+ person development team for IBM, producing one of their mainfram OS’s (OS/390 IIRC, or it could have been 360?). It analyses lots of different approaches to development processes in an easy-to-read style, and was written as a kind of extended post-mortem after the project was over.
I’m posting all these comments mainly because it’s basic stuff for any programmer (more improtant than learning that stupid OO thing which people often start with these days…), but I’m well aware that many many people don’t really encounter this stuff.
So, look at my comments as being aimed at the readers in general (and all the lurkers etc). I also get the impression you have had little in the way of deep exposure to things like this - for instance your apparent misunderstanding of what UML is meant for. It’s like someone who uses the side of an axe to hammer in nails, but doesn’t realise the axe can do so much more instead. Nothing wrong with what you’re doing, I just don’t want people who’ve never heard of UML before to read this and get the wrong idea of what it’s all about (which appears to be what had already happened to the original poster - who sounded like he thought UML was superficially similar to XML, and was intended to be extended and subclassed).
[Obviously, UML is a modelling language, XML is a markup language, so in fact there are no meaningful similarities - other than that they both happen to be fairly modern open-standards]
well said and I am going to take it in a similar fashion…BTW I haven’t read that one but I intent to get my hands on it … also I realized what and why you got into this …I was giving bad example to those are still unspoiled by bad practices of old hands ;D … just not sure if you are not the guy you gave me hell in my 4th year OS class - but naaay - that was long time ago he can’t possible be alive now ;D
[quote]Almost every skilled and serious developer I know well uses UML or an equivalent (e.g. specialist techniques for specialist methodologies, like XP).
[/quote]
I’ve been programming for over 15 years, and I have never encountered a place where UML is used or anyone that uses it. I have tried a couple times to use it because I DO like having a design up front. But the tools for UML are all a much greater hinderance than they are a help. That combined with not being able to get any colleagues to use it makes it impractical for me.
The XP view seem to be “What is a better design document than source code?” It is unambiguous and describes EXACTLY how the program will operate. It has important details that effect design decisions - e.g. will this collection be a linked list or an array? It is important to know lest you assume that insertions will be quick and discover months later (when you join your code with the code of the person that did the collection) there is an array copy happening every time. That’s the sort of detail that will get left out of the design document… and if you put that much detail in the design doc you might as well write it with source code.
The only time I programmed with a full design document was in university. And it was great, the programming went very smoothly having that document. unfortunately it has been my experience in the ‘real’ world that nobody thinks they have time to do a proper design document. Which is why I think XP works.
Now I just code. The faster I can get something coded the sooner I know how well it works, how well it fits with the rest of the program, etc. I think this works best for small teams, the time spent on large detailed UML diagrams is better spent coding. IF there were any tools that could keep UML diagrams synchronized to source code without cluttering them with incidental classes that aren’t significant then I would keep a diagram too, and use it to better architect the program. I have yet to see such a tool, and thus UML remains far too time consuming to be practical… You can be through a few iterations of the actual code before you get the diagram done - it just doesn’t make sense. I blame the sorry state of UML tools for this. It would be better if a UML design could be done up front.
I’ve seen all the demos and eval versions of UML tools from Rational, etc. They suck. They are absolutely useless, there is no way that I could get any of the programmers I have worked with to use them. (And I have worked with some really good programmers.) They just slow you down soooo much. I know that having a proper plan & design doc speeds up coding… The problem is that currently it takes so long to make a good UML diagram that all the savings are lost many times over. At least that has been my experience… I wish it wasn’t the case.
Ok, I’m done ranting… now Blah can tear me to pieces
Which has it’s own advantages and disadvantages, and in some scenarios is more appropriate than UML - although XP does do a form of modelling in that you write your test case that tests the code you are about to write before you write the code itself. Hence I included XP when I mentioned that there are other ways beside UML of achieving “similar” aims.
Incidentally there are (of course) other ways of solving that problem, and Eiffel’s Design-by-Contract is a classic example (and gets a lot of evangelism!)…but no solution is going to be 100% perfect :(.
Indeed…although I could find you examples where XP didn’t work too :). XP can “break” if various requirements of the people using it fail (like if some of them can’t cope with ego-less programming).
Typically, if your UML diagrams are taking much time, then there’s something wrong with the way you’re using them. As you point out, these days that can often be because the tool you’re using for them is the problem - and that often you cannot get a better tool :(.
C.f. the threads in the Tools Discussion area about “which UML software should I use?”…and, in all honesty, I still use paper and giant whiteboards as the primary means of doing UML diagrams (and then we use a digital camera to photograph the whiteboard at the end of a session - a very easy, non-time-intensive way of recording things. It works surprisingly well!).
…and every time I do it I feel extremely frustrated that I’m not doing this on a computer, and feel like it’s 1960. MS spends billions of dollars adding bugs and removing features from their flagship word-processor (who needs a goddamn WP these days anyway? DTP apps could do nearly all the work! [tongue in cheek]), but the industry can’t see their way to doing something useful like making a good UML tool.
I’ll do another plug for FUJABA here (from UML to Java and back again) - it does do what it’s supposed to do very well - and it even automatically generates sequence diagrams / call graphs extremely well too. I’ve been using it for 4 years, and my only complaints are that the development pace is GLACIAL and that the feature set is still very small - so all in all it gives you an appetite for more, but takes years to satisfy you!
It’s extremely good for taking a snapshot of someone else’s code when you need to understand it and they haven’t written any docs yet. OTOH, the last version I used still didn’t have automatic layout (which, incidentally, is a VERY difficult problem to solve - algorithms for layout are hard to do…which probably partly explains why none of the cheap/free editors do decent layouts 
), which meant lots of time-wasting by-hand layout (although of course all the connectors moved automatically) every now and then. It’s just the missing features that let it down :(.
My thoughts exactly :(. But I haven’t given up on the process (c.f. the fact I use it on paper etc), just the tools.
Cynical thought: the reason there are no decent free UML tools is that a good UML tool is so valuable everyone charges money for them. The reason ROSE is such crud is that enterprises know its so valuable they’ll pay for it even when it’s almost impossible to use.
[/quote]
Oh, um, er, “you’re wrong!”.
[quote]…in all honesty, I still use paper and giant whiteboards as the primary means of doing UML diagrams (and then we use a digital camera to photograph the whiteboard at the end of a session - a very easy, non-time-intensive way of recording things. It works surprisingly well!).
…and every time I do it I feel extremely frustrated that I’m not doing this on a computer, and feel like it’s 1960.
…Oh, um, er, “you’re wrong!”.
[/quote]
Exactly! An informal team meeting with a white-board gets you so much… it’s a bit hard to do if you can’t get everyone in the same room though. Imagine trying to do the same diagrams with a computer? There was actually a really old maker of pen-based tablet computers (was the name ‘Go’ ?) that had some pretty nice sketching software. That is the stuff you need… it is impossible to sketch with a mouse… so a digitizing white-board or a tablet computer would be needed.
I also had a actual pen and paper notepad with a special digitizing pen. You write and draw in pen on regular paper, but the pen has a radio link to a computer in the clipboard and then you plug in and download your pages – with text recognition and diagrams… sweet, but unfortunately a little too awkward to use.
I also saw a project many years ago at university that had a source code editor very well integrated with flow-chart diagrams and stuff… you could flip between views effortlessly and they were always in sync. We need the UML version of that.
When you can type in a UML ‘diagram’ using source code very quickly it may be just as good if not better than sketching a diagram in the first place. Just write code for basic class structure and methods names… leave the core algorithms to be filled in later.